![Knowledge Representation and Indexing: Background and Future 27
As this pre-combined approach is rather complicated and would lead to much
overhead in the classification system (i.e. one would certainly end up with classes
that are never used); one might rather identify the basic categories of the domain
and their subclasses (for this purpose, the single facets have to be disjoint; a con-
cept cannot be part of more than one facet). This could look like the following
faceted scheme proposed by Broughton (Broughton, 2004; Broughton, 2006):
Colour Pattern Material Function Length
Black Plain Wool Work Ankle
Grey Striped Polyester Evening Calf
Brown Spotted Cotton Football Knee
Green Hooped Silk Hiking
Blue Checkered Nylon Protective
Red Novelty Latex
By combining these facets we may then represent things like ‘blue striped cotton
socks for football’. The individual complex concepts are only created via combi-
nation at the time they are needed (and not in advance within a complete hierar-
chy). This principle is called post-coordination (Bertram, 2005; Stock Stock,
2008). The citation order defines the order in which the elements of different fac-
ets have to be arranged, which is needed for the construction of notations based on
facet combinations. Specific symbols may be used to separate the faceted elements
from each other within notations.
The first (and still prevalent) example of a faceted classification system is the
Colon Classification (CC) (Ranganathan, 1965; Ranganathan 1987[1933]). Other
examples are the Bliss Bibliographic Classification, 2nd
Edition (BC2) (Mills
Broughton, 1977) and the Classification of Library and Information Science
(Daniel Mills, 1975). The faceted approach to knowledge organization has also
been adopted by other KOS.
Some classifications partly make use of additional facets, without relying en-
tirely on this combinatory approach (Stock Stock, 2008). They contain addi-
tional tables with aspects that often co-occur with other concepts. For example, the
DDC has auxiliary tables for geographical information that can be attached to the
concepts from the main thematic classification.
Thesaurus
The last major type of classical knowledge organization system is the thesaurus
(Aitchison, Gilchrist Bawden, 2000; Foskett, 1981; Lancaster, 1986, Stock
Stock, 2008; Taylor, 1999). While in some contexts as well as in colloquial lan-
guage the term ‘thesaurus’ denotes a lexicon of synonyms, in the field of knowl-
edge representation it describes a complex controlled vocabulary. Thesauri in the
latter sense are standardized by national as well as international norms (BS
5723:1987; ISO 2788:1986; DIN 1463/1:1987). They pay a lot of attention to the](https://image.slidesharecdn.com/25481937-250509124708-c3539fd1/85/Knowledge-Representation-In-The-Social-Semantic-Web-Katrin-Weller-39-320.jpg)
![28 Knowledge Representation and Indexing: Background and Future
collection of synonyms, but also use hierarchical concept structures and associa-
tive relations between related terms to represent domain knowledge. Aitchison,
Gilchrist and Bawden (2000) define a thesaurus as
“[…] a vocabulary of controlled indexing language, formally organized so
that a priori relationships between concepts are made explicit, to be used in
information retrieval systems, ranging from card catalogue to the Internet.”
(Aitchison, Gilchrist Bawden, 2000)
Concepts in thesauri are typically not supported by complete definitions (as in a
dictionary). Yet thesauri may contain (short) definitions for concepts – and more
and more thesauri do make use of this feature. Aitchison, Gilchrist and Bawden
(2000) point out:
“Definitions tend to be necessary most frequently in social science and hu-
manities thesauri, to clarify imprecise terminology, which occurs more of-
ten in these subject areas.” (Aitchison, Gilchrist Bawden, 2000)
Furthermore, scope notes are used in thesauri to indicate restrictions on meaning
or the range of concepts, to give instructions to indexers, to provide examples for
usage or to document the term history (Aitchison, Gilchrist Bawden, 2000).
For the purpose of vocabulary control, most thesauri work with preferred terms
in a set of synonyms. These preferred terms are sometimes called descriptors. De-
scriptors are exclusively used for indexing and retrieval; all their synonyms (the
non-preferred terms) are only used as pointers to the preferred terms. In thesauri
that do not make use of preferred terms, all synonyms are regarded as equal and
may be used for indexing.
The typical way of displaying a thesaurus is in the form of an alphabetical list
of all terms17
, with additional information on concept interrelations and usage: the
term records. A prototypical entry for a preferred term in the alphabetical display
of term records has the following form:
PREFERRED TERM
SN “Scope Note”
D “Definition”
UF “Used For”: References to equivalent non-preferred terms.
BT “Broader Term”: References to broader terms.
NT “Narrower Term”: References to narrower terms.
RT “Related Term”: References to related terms.
An entry for a non-preferred term only contains the pointer to the respective pre-
ferred terms, e.g.:
Non-preferred term
USE PREFERRED TERM
17
Another frequently chosen display option is an additional hierarchical list of terms. For
additional display approaches see Aitchison, Gilchrist Bawden (2000).](https://image.slidesharecdn.com/25481937-250509124708-c3539fd1/85/Knowledge-Representation-In-The-Social-Semantic-Web-Katrin-Weller-40-320.jpg)
![CANTO DECIMOOCTAVO
AY un lugar en el Infierno, llamado Malebolge,[18]
construído todo de piedra y de color ferruginoso, como
la cerca que lo rodea. En el centro mismo de aquella
funesta planicie se abre un pozo bastante ancho y
profundo, de cuya estructura me ocuparé en su lugar.
El espacio que queda entre el pozo y el pie de la dura
cerca es redondo, y está dividido en diez valles, o recintos cerrados,
semejantes a los numerosos fosos que rodean a un castillo para
defensa de las murallas; y así como estos fosos tienen puentes que
van desde el umbral de la puerta a su otro extremo, del mismo
modo aquí avanzaban desde la base de la montaña algunas rocas,
que atravesando las márgenes y los fosos, llegaban hasta el pozo
central, y allí se reunían quedando truncadas. Tal era el sitio donde
nos encontramos cuando descendimos de la grupa de Gerión: el
Poeta echó a andar hacia la izquierda, y yo seguí tras él. A mi
derecha vi nuevas causas de conmiseración, nuevos tormentos y
nuevos burladores, que llenaban la primera fosa. En el fondo](https://image.slidesharecdn.com/25481937-250509124708-c3539fd1/85/Knowledge-Representation-In-The-Social-Semantic-Web-Katrin-Weller-59-320.jpg)
![estaban desnudos los pecadores; los del centro acá venían de frente
a nosotros; y los de esta parte afuera seguían nuestra misma
dirección, pero con paso más veloz. Como en el año del Jubileo, a
causa de la afluencia de gente que atraviesa el puente de San
Angelo, los romanos han determinado que todos los que se dirijan al
castillo y vayan hacia San Pedro pasen por un lado, y por el otro los
que van hacia el monte, así vi, por uno y otro lado de la negra roca,
cornudos demonios con grandes látigos, que azotaban cruelmente
las espaldas de los condenados. ¡Oh! ¡Cómo les hacían mover las
piernas al primer golpe! Ninguno aguardaba el segundo ni el tercero.
Mientras yo andaba, mis ojos se encontraron con los de un pecador,
y dije en seguida: No es la primera vez que veo a ése. Por lo que
me detuve a observarlo mejor: mi dulce Guía se detuvo al mismo
tiempo, y aun me permitió retroceder un tanto. El azotado creyó
ocultarse bajando la cabeza; mas le sirvió de poco, pues le dije:
—Tú, que fijas los ojos en el suelo, si no son falsas las facciones que
llevas, eres Venedico Caccianimico. Pero ¿qué es lo que te ha traído
a tan picantes salsas?
A lo que me contestó:
—Lo digo con repugnancia; pero cedo a tu claro lenguaje, que me
hace recordar el mundo de otro tiempo. Yo fuí aquel que obligó a la
bella Ghisola a satisfacer los deseos del Marqués, cuéntese como se
quiera la tal historia. Y no soy el único boloñés que llora aquí; antes
bien este sitio está tan lleno de ellos, que no hay en el día entre el
Savena y el Reno tantas lenguas que digan sipa,[19] como en esta
fosa; y si quieres una prueba de lo que te digo, recuerda nuestra
codicia notoria.
Mientras así hablaba, un demonio le pegó un latigazo, diciéndole:
Anda, rufián; que aquí no hay mujeres que se vendan.
Me reuní a mi Guía; y a los pocos pasos llegamos a un punto de
donde salía una roca de la montaña. Subimos por ella ligeramente, y
volviendo a la derecha sobre su áspero dorso, salimos de aquel](https://image.slidesharecdn.com/25481937-250509124708-c3539fd1/85/Knowledge-Representation-In-The-Social-Semantic-Web-Katrin-Weller-60-320.jpg)
![CANTO DECIMONONO
H Simón el mago! ¡Oh miserables sectarios suyos,
almas rapaces, que prostituís a cambio de oro y plata
las cosas de Dios, que deben ser las esposas de la
virtud! Ahora resonará la trompa para vosotros, puesto
que os encontráis en la tercera fosa.
Estábamos ya junto a ésta, subidos en aquella parte
del escollo que cae justamente sobre su centro. ¡Oh suma Sabiduría!
¡Cuán grande es el arte que demuestras en el cielo, en la tierra y en
el mundo maldito, y con cuánta equidad se reparte tu virtud! Vi en
los lados y en el fondo la piedra lívida llena de pozuelos, todos
redondos y de igual tamaño, los cuales me parecieron ni más ni
menos anchos que los que hay en mi hermoso San Juan para servir
de pilas bautismales; uno de éstos rompí yo no ha muchos años, por
salvar a un niño que dentro de él se ahogaba; y baste lo que digo,
para desengañar a todos.[20] Fuera de la boca de cada uno de
aquellos pozuelos salían los pies y las piernas de un pecador, hasta
el muslo, quedando dentro el resto del cuerpo. Ambos pies estaban](https://image.slidesharecdn.com/25481937-250509124708-c3539fd1/85/Knowledge-Representation-In-The-Social-Semantic-Web-Katrin-Weller-63-320.jpg)
![encendidos, por cuya razón se agitaban tan fuertemente sus
coyunturas, que hubieran roto sogas y cuerdas. Del mismo modo
que la llama suele recorrer la superficie de los objetos untados de
grasa, así el fuego flameaba desde el talón a la punta en los pies de
los condenados.
—¿Quién es aquél, Maestro, que furioso agita los pies más que sus
otros compañeros—dije entonces—, y a quien corroe y deseca una
llama mucho más roja?
A lo cual me contestó:
—Si quieres que te conduzca por aquella parte de la escarpa que
está más cercana al fondo, él mismo te dirá quién es y cuáles son
sus crímenes.
Le respondí:
—Me parece bien todo lo que a ti te agrada: tú eres el dueño y
sabes que yo no me separo de tu voluntad, así como también
conoces lo que me callo.
Subimos entonces al cuarto margen; después volvimos y bajamos
por la izquierda hacia la estrecha y perforada fosa, sin que el buen
Maestro me hiciera separar de su lado, hasta haberme conducido
junto al hoyo de aquel que daba tantas señales de dolor con los
movimientos de sus piernas.
—¡Oh! Quienquiera que seas, tú, que tienes enterrada la parte
superior de tu cuerpo; alma triste, plantada como una estaca—
empecé a decir—, habla, si puedes.
Yo estaba como el fraile que confiesa al pérfido asesino, que, metido
en la tierra, le llama para que cese su muerte. Y él gritó:
—¿Estás ya aquí derecho, estás ya aquí derecho, Bonifacio?[21] Me
ha engañado en algunos años lo que está escrito. ¿Tan pronto te has
saciado de aquellos bienes, por los cuales no temiste apoderarte con](https://image.slidesharecdn.com/25481937-250509124708-c3539fd1/85/Knowledge-Representation-In-The-Social-Semantic-Web-Katrin-Weller-64-320.jpg)
![embustes de la hermosa Dama,[22] y gobernarla después
indignamente?
Quedéme, al oír esto, como aquellos que, casi avergonzados de no
haber comprendido lo que se les ha dicho, no saben qué contestar.
Entonces Virgilio dijo:
—Respóndele pronto: yo no soy, yo no soy el que tú crees.
Y yo contesté como se me ordenó. Por lo cual el espíritu retorció sus
pies; y luego, suspirando y con llorosa voz, me dijo:
—¿Pues qué es lo que me preguntas? Si te urge conocer quién soy,
hasta el punto de haber descendido para ello por todos estos
peñascos, sabrás que estuve investido del gran manto, y fuí
verdadero hijo de la Osa, tan codicioso, que, por aumentar la
riqueza de los oseznos, embolsé arriba todo el dinero que pude, y
aquí mi alma. Bajo mi cabeza están sepultados los demás papas,
que antes de mí cometieron simonía, y se hallan comprimidos a lo
largo de este angosto agujero. Yo me hundiré también luego que
venga aquel que creí fueses tú, cuando te dirigí mi súbita pregunta.
Pero desde que mis pies se abrasan, y me encuentro colocado al
revés, ha transcurrido más tiempo del que él permanecerá en este
mismo sitio con los pies quemados; porque en pos de él vendrá de
poniente un pastor sin ley, por causa más repugnante, y ése deberá
cubrirnos a entrambos. Será un nuevo Jasón, parecido al de que se
habla en el libro de los Macabeos; y así como el rey de éste fué débil
para con él, así con el otro lo será el que rige la Francia.
No sé si en tal momento fué demasiada audacia la mía; pues le
respondí en estos términos:
—¡Eh!, dime: ¿cuánto dinero exigió Nuestro Señor de San Pedro,
antes de poner las llaves en su poder? En verdad que no le pidió
más sino que le siguiera. Ni Pedro ni los otros pidieron a Matías oro
ni plata cuando por suerte fué elegido en reemplazo del que perdió
su alma traidora. Permanece, pues, ahí, porque has sido castigado
justamente, y guarda bien la mal adquirida riqueza, que tan atrevido](https://image.slidesharecdn.com/25481937-250509124708-c3539fd1/85/Knowledge-Representation-In-The-Social-Semantic-Web-Katrin-Weller-65-320.jpg)
![te hizo contra Carlos. Y si no fuese porque aun me contiene el
respeto a las llaves soberanas, que poseíste en tu alegre vida,
emplearía palabras mucho más severas; porque vuestra avaricia
contrista al mundo, pisoteando a los buenos, y ensalzando a los
malos. Pastores, a vosotros se refería el Evangelista, cuando vió
prostituída ante los reyes a la que se sienta sobre las aguas; a la
que nació con siete cabezas, y obtuvo autoridad por sus diez
cuernos, mientras la virtud agradó a su marido.[23] Os habéis
construído dioses de oro y plata: ¿qué diferencia, pues, existe entre
vosotros y los idólatras, sino la de que ellos adoran a uno y vosotros
adoráis a ciento? ¡Ah, Constantino! ¡A cuántos males dió origen, no
tu conversión al cristianismo, sino la donación que de ti recibió el
primer papa que fué rico!
Mientras yo le hablaba con esta claridad, él, ya fuese a impulsos de
la ira, o porque le remordiese la conciencia, respingaba fuertemente
con ambas piernas. Creo que complací a mi Guía; porque escuchó
siempre con rostro satisfecho el sonido de mis palabras, expresadas
con sinceridad. Entonces me cogió con los dos brazos, y teniéndome
en alto bien afianzado sobre su pecho, volvió a subir por el camino
por donde habíamos descendido, sin dejar de estrecharme contra sí,
hasta llegar a la parte superior del puente que va de la cuarta a la
quinta calzada. Allí, depositó suavemente su querido fardo sobre el
áspero y pelado escollo, que hasta para las cabras sería un difícil
sendero, desde donde descubrí una nueva fosa.](https://image.slidesharecdn.com/25481937-250509124708-c3539fd1/85/Knowledge-Representation-In-The-Social-Semantic-Web-Katrin-Weller-66-320.jpg)
![lloraba en verdad, apoyado contra una de las rocas de la dura
montaña, de suerte que mi Guía me dijo:
—¿Tú también eres de los insensatos? Aquí vive la piedad cuando
está bien muerta. ¿Quién es más criminal que el que se apasiona
contemplando la justicia divina? Levanta la cabeza, levántala y mira
a aquel por quién se abrió la tierra en presencia de los tebanos, que
exclamaban: ¿Adónde caes, Anfiarao? ¿Por qué abandonas la
guerra? Y no cesó de caer en el Infierno hasta llegar a Minos, que
se apodera de cada culpable. Mira cómo ha convertido sus espaldas
en pecho: por haber querido ver demasiado hacia adelante, ahora
mira hacia atrás, y sigue un camino retrógrado. Mira a Tiresias, que
mudó de aspecto cuando de varón se convirtió en hembra,
cambiando también todos su miembros, y hubo de abatir con su
vara las dos serpientes unidas, antes que recobrara su pelo viril. El
que acerca sus espaldas al vientre de aquél es Aronte, que tuvo por
morada una gruta de blancos mármoles en las montañas de Luni,
cultivadas por el carrarés que habita en su falda, y desde allí no
había nada que limitara su vista, cuando contemplaba el mar o las
estrellas. Aquella que, con los destrenzados cabellos, cubre sus
pechos, por lo cual se ocultan a tus miradas, y tiene en ese lado de
su cuerpo todas las partes velludas, fué Manto, que recorrió muchas
comarcas, hasta que se detuvo en el sitio donde yo nací; por lo cual
deseo que me prestes un poco de atención. Luego que su padre
salió de la vida, y fué esclavizada la ciudad de Baco,[24] Manto
anduvo errante por el mundo durante mucho tiempo. Allá arriba, en
la bella Italia, existe un lago al pie de los Alpes que ciñen la
Alemania por la parte superior del Tirol, el cual se llama Benaco. Mil
corrientes, y aun más, según creo, vienen a aumentar, entre Garda,
Val-Camonica y el Apenino, el agua que se estanca en dicho lago. En
medio de éste hay un sitio, donde el Pastor de Trento, y los de
Verona y Brescia, podrían dar su bendición si siguiesen aquel
camino. En el punto donde es más baja la orilla que le circunda, está
situada Peschiera, bello y fuerte castillo, a propósito para hacer
frente a los de Brescia y a los de Bérgamo. Allí afluye
necesariamente toda el agua que no puede estar contenida en el](https://image.slidesharecdn.com/25481937-250509124708-c3539fd1/85/Knowledge-Representation-In-The-Social-Semantic-Web-Katrin-Weller-68-320.jpg)
![Ten cuidado, ten cuidado. Entonces me volví como el hombre que
ansía ver aquello de que le conviene huír, y a quien asalta un temor
tan grande y repentino, que ni para mirar detiene su fuga; y vi
detrás de nosotros un negro diablo, que venía corriendo por el
puente. ¡Oh! ¡Cuán feroz era su aspecto, y qué amenazador me
parecía con sus alas abiertas y sus ligeros pies! Sobre sus hombros,
altos y angulosos, llevaba a cuestas un pecador, a quien tenía
agarrado por ambos jarretes. Desde nuestro puente dijo:
—¡Oh! Malebranche, ved aquí uno de los ancianos de Santa Zita:
ponedle debajo; que yo me vuelvo otra vez a aquella tierra, que está
tan bien provista de ellos. Allí todos son bribones, excepto Bonturo;
y por dinero, de un no hacen un ita.[25]
Le arrojó abajo, y se volvió por la dura roca tan de prisa, que jamás
ha habido mastín suelto que haya perseguido a un ladrón con tanta
ligereza. El pecador se hundió y volvió a subir hecho un arco; pero
los demonios, que estaban resguardados por el puente, gritaban:
—Aquí no está el Santo Rostro; aquí se nada de diferente modo que
en el Serchio. Si no quieres probar nuestros garfios, no salgas de la
pez.
Después le pincharon con más de cien harpones, diciéndole:
—Es forzoso que bailes aquí a cubierto, de modo que, si puedes,
prevariques ocultamente.
No de otra suerte hacen los cocineros que sus marmitones sumerjan
en la caldera las viandas por medio de grandes tenedores, para que
no sobrenaden.
—A fin de que no adviertan que estás aquí—me dijo el buen Maestro
—, ocúltate detrás de una roca, que te sirva de abrigo; y aunque se
me haga alguna ofensa, no temas nada; pues ya conozco estas
cosas por haber estado otra vez entre estas almas venales.](https://image.slidesharecdn.com/25481937-250509124708-c3539fd1/85/Knowledge-Representation-In-The-Social-Semantic-Web-Katrin-Weller-72-320.jpg)
![¿Quiéres que le pinche en la rabadilla?, decía uno de ellos a los
otros. Y respondían: Sí, sí clávale. Pero aquel demonio, que estaba
conversando con mi Guía, se volvió de repente, y gritó: Quieto,
quieto, Scarmiglione. Después nos dijo:
—Por este escollo no podréis ir más lejos, pues el sexto arco yace
destrozado en el fondo. Si os place ir más adelante, seguid esta
costa escarpada: cerca veréis otro escollo por el que podréis pasar.
Ayer, cinco horas más tarde que en este momento, se cumplieron mil
doscientos sesenta y seis años desde que se rompió aquí el camino.
[26] Voy a enviar hacia allá varios de los míos para que observen si
algún condenado procura sacar la cabeza al aire: id con ellos, que
no os harán daño.
—Adelante, Alichino y Calcabrina—empezó a decir—; y tú también,
Cagnazzo; Barbariccia guiará a los diez. Vengan además Libicocco, y
Draghignazzo; Ciriatto, el de los grandes colmillos, y Graffiacane, y
Farfarello, y el loco de Rubicantondad en torno de la pez hirviente:
éstos deben llegar salvos hasta el otro escollo, que atraviesa
enteramente sobre la fosa.[27]
—¡Oh Maestro! ¿Qué es lo que veo?—dije—; si conoces el camino,
vamos sin escolta; yo, por mí, no la solicito. Si eres tan prudente
como de costumbre, ¿no ves que rechinan los dientes, y se hacen
guiños que nos amenazan algún mal?
—No quiero que te espantes—me contestó—; deja que rechinen los
dientes a su gusto. Si lo hacen, es por los desgraciados que están
hirviendo.
Se pusieron en camino por la margen izquierda; pero cada uno de
aquéllos de antemano se habían mordido la lengua en señal de
inteligencia con su jefe, y éste se sirvió de su ano a guisa de
trompeta.](https://image.slidesharecdn.com/25481937-250509124708-c3539fd1/85/Knowledge-Representation-In-The-Social-Semantic-Web-Katrin-Weller-74-320.jpg)
Knowledge Representation In The Social Semantic Web Katrin Weller
- 1. Knowledge Representation In The Social Semantic Web Katrin Weller download https://ebookbell.com/product/knowledge-representation-in-the- social-semantic-web-katrin-weller-50963874 Explore and download more ebooks at ebookbell.com
- 2. Here are some recommended products that we believe you will be interested in. You can click the link to download. Education Professionalization And Social Representations On The Transformation Of Social Knowledge Routledge International Studies In The Philosophy Of Education Mohamed Chaib https://ebookbell.com/product/education-professionalization-and- social-representations-on-the-transformation-of-social-knowledge- routledge-international-studies-in-the-philosophy-of-education- mohamed-chaib-2472498 Advances In Knowledge Representation Logic Programming And Abstract Argumentation Essays Dedicated To Gerhard Brewka On The Occasion Of His 60th Birthday 1st Edition Thomas Eiter https://ebookbell.com/product/advances-in-knowledge-representation- logic-programming-and-abstract-argumentation-essays-dedicated-to- gerhard-brewka-on-the-occasion-of-his-60th-birthday-1st-edition- thomas-eiter-4974024 Intersex Narratives Shifts In The Representation Of Intersex Lives In North American Literature And Popular Culture Viola Amato Verst Knowledge Unlatched Ku Select 2017 Backlist Collection https://ebookbell.com/product/intersex-narratives-shifts-in-the- representation-of-intersex-lives-in-north-american-literature-and- popular-culture-viola-amato-verst-knowledge-unlatched-ku- select-2017-backlist-collection-51807104 Global Perspectives In Modern Italian Culture Knowledge And Representation Of The World In Italy From The Sixteenth To The Early Nineteenth Century 1st Edition Guido Abbattista Editor https://ebookbell.com/product/global-perspectives-in-modern-italian- culture-knowledge-and-representation-of-the-world-in-italy-from-the- sixteenth-to-the-early-nineteenth-century-1st-edition-guido- abbattista-editor-36132388
- 3. The Logic Of Adaptive Behavior Knowledge Representation And Algorithms For Adaptive Sequential Decision Making Under Uncertainty In Firstorder And Relational Domains M Van Otterlo https://ebookbell.com/product/the-logic-of-adaptive-behavior- knowledge-representation-and-algorithms-for-adaptive-sequential- decision-making-under-uncertainty-in-firstorder-and-relational- domains-m-van-otterlo-2541296 The Imposition Of Form Studies In Narrative Representation And Knowledge Course Book Claudia J Brodsky https://ebookbell.com/product/the-imposition-of-form-studies-in- narrative-representation-and-knowledge-course-book-claudia-j- brodsky-51957022 Process Support And Knowledge Representation In Health Care Bpm 2012 Joint Workshop Prohealth 2012kr4hc 2012 Tallinn Estonia September 3 2012 Revised Selected Papers 1st Edition Yuval Shahar Auth https://ebookbell.com/product/process-support-and-knowledge- representation-in-health-care-bpm-2012-joint-workshop- prohealth-2012kr4hc-2012-tallinn-estonia-september-3-2012-revised- selected-papers-1st-edition-yuval-shahar-auth-4241172 Process Support And Knowledge Representation In Health Care Aime 2013 Joint Workshop Kr4hc 2013prohealth 2013 Murcia Spain June 1 2013 Revised Selected Papers 1st Edition Stefan Schulz https://ebookbell.com/product/process-support-and-knowledge- representation-in-health-care-aime-2013-joint-workshop- kr4hc-2013prohealth-2013-murcia-spain-june-1-2013-revised-selected- papers-1st-edition-stefan-schulz-4602022 Advances In Knowledge Representation C Gutierrez https://ebookbell.com/product/advances-in-knowledge-representation-c- gutierrez-2612542
- 6. Knowledge & Information Studies in Information Science Edited by Wolfgang G. Stock (Düsseldorf, Germany) and Ronald E. Day (Bloomington, Indiana, U.S.A.) Sonja Gust von Loh (Düsseldorf, Germany) – Associate Editor Richard J. Hartley (Manchester, U.K.) Robert M. Hayes (Los Angeles, California, U.S.A.) Peter Ingwersen (Copenhagen, Denmark) Michel J. Menou (Les Rosiers sur Loire, France, and London, U.K.) Stefano Mizzaro (Udine, Italy) Christian Schlögl (Graz, Austria) Sirje Virkus (Tallinn, Estonia) Knowledge and Information (K&I) is a peer-reviewed information science book series. The scope of information science comprehends representing, providing, searching and finding of relevant knowledge including all activities of information professionals (e.g., indexing and abstracting) and users (e.g., their information behavior). An important research area is information retrieval, the science of search engines and their users. Topics of knowledge representation include metadata as well as methods and tools of knowledge organization systems (folksonomies, nomenclatures, classification systems, thesauri, and ontologies). Informetrics is empirical information science and consists, among others, of the domain-specific metrics (e.g., scientometrics, webometrics), user and usage research, and evalua- tion of information systems. The sharing and the distribution of internal and exter- nal information in organizations are research topics of knowledge management. The information market can be defined as the exchange of digital information on networks, especially the World Wide Web. Further important research areas of information science are information ethics, information law, and information soci- ology. De Gruyter Saur
- 7. Katrin Weller Knowledge Representation in the Social Semantic Web De Gruyter Saur
- 8. D 61 ISBN 978-3-598-25180-1 e-ISBN 978-3-598-44158-5 ISSN 1868-842X Library of Congress Cataloging-in-Publication Data Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de. © 2010 Walter de Gruyter GmbH & Co. KG, Berlin/New York Printing: Hubert & Co. GmbH & Co. KG, Göttingen ∞ rinted on acid-free paper Printed in Germany www.degruyter.com Weller, Katrin. Knowledge representation in the social semantic Web / Katrin Weller. p. cm. -- (Knowledge & information, ISSN 1868-8144) Includes bibliographical references and index. ISBN 978-3-598-25180-1 -- ISBN 978-3-598-44158-5 (ebook) 1. Semantic Web. 2. Knowledge representation (Information theory) I. Title. TK5105.88815.W45 2010 025.042'7--dc22 2010030140
- 9. Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 State of Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Objectives and Main Research Questions of this Book. . . . . . . . . . . . . . . 9 Outline of this Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Formal Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Chapter 1 Knowledge Representation and Indexing: Background and Future . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 1.1 Knowledge Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 1.1.1 Metadata, Indexing and Retrieval . . . . . . . . . . . . . . . . . . 17 1.1.2 Classical Knowledge Organization Systems . . . . . . . . . . . 21 1.1.3 Knowledge Organization Systems in Practice . . . . . . . . . . 32 1.2 The Semantic Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 1.2.1 The Vision of a Semantic Web . . . . . . . . . . . . . . . . . . . . 53 1.2.2 Indexing the Semantic Web: Ontologies. . . . . . . . . . . . . . 55 1.3 The Social Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 1.3.1 Web 2.0 and the Role of the User. . . . . . . . . . . . . . . . . . . 67 1.3.2 Indexing the Social Web: Folksonomies. . . . . . . . . . . . . . 70 1.3.3 Collaboration vs. Collection . . . . . . . . . . . . . . . . . . . . . . 73 1.4 The Social Semantic Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 1.4.1 A Networked World: Web 2.0, Semantic Web & e- Science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 1.4.2 Approaches for Mashing-Up the Social and Semantic Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 1.4.3 Knowledge Representation in the Social Semantic Web. . . 93 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Chapter 2 Ontologies: Semantics for the Web. . . . . . . . . . . . . . . . . . . . 115 2.1 Ontologies: Definition and Background . . . . . . . . . . . . . . . . . . . . 115 2.1.1 From Philosophy to Computer Science. . . . . . . . . . . . . . . 115 2.1.2 Ontology Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 2.2 Structure and Elements of Ontologies. . . . . . . . . . . . . . . . . . . . . . 123 2.2.1 Basic Elements of Ontologies . . . . . . . . . . . . . . . . . . . . . 123
- 10. vi Contents 2.2.2 Concepts and Instances. . . . . . . . . . . . . . . . . . . . . . . . . . 129 2.2.3 Relations and Properties . . . . . . . . . . . . . . . . . . . . . . . . . 144 2.2.4 Selected Types of Specified Semantic Relations . . . . . . . . 156 2.2.5 Attributes of Semantic Relations . . . . . . . . . . . . . . . . . . . 173 2.2.6 Formal and Informal Semantics. . . . . . . . . . . . . . . . . . . . 182 2.3 Ontologies as Knowledge Organization Systems . . . . . . . . . . . . . . 190 2.3.1 Selected Popular and Important Ontologies . . . . . . . . . . . 190 2.3.2 From Vocabulary to Ontology: Towards and Analytical Description of Knowledge Organization Systems . . . . . . . 204 2.3.3 Requirements for Ontologies in the Social Semantic Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 Chapter 3 Ontology Engineering in the Era of the Social Semantic Web 241 3.1 Ontology Engineering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 3.1.1 Methodologies, Dimensions and Principles of Ontology Engineeering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 3.1.2 Tools for Ontology Engineering . . . . . . . . . . . . . . . . . . . 256 3.1.3 Recent Trends and Open Challenges in Ontology Engineeering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264 3.2 Community-based Ontology Engineering . . . . . . . . . . . . . . . . . . . 272 3.2.1 Basic Principles of Community-based Ontology Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 3.2.2 New Approaches and New Tools for Community-based Ontology Engineering. . . . . . . . . . . . . . . . . . . . . . . . . . . 285 3.3 Semantic Upgrades & Tag Gardening. . . . . . . . . . . . . . . . . . . . . . 317 3.3.1 From Tagging to Ontologies . . . . . . . . . . . . . . . . . . . . . . 318 3.3.2 Collaborative, Personal and Cross-Platform Tag Gardening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329 3.3.3 Ontology Gardening: Engineering Based on Semantic Enrichment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336 3.3.4 Additional Knowledge Resources for Ontology Gardening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342 3.4 Co-Existence of Knowledge Organization Systems . . . . . . . . . . . . 351 3.4.1 Multiple Knowledge Representations on the Social Semantic Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352 3.4.2 KOSO: A Meta-KOS for Knowledge Organization Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361 3.4.3 Mapping the Social Semantic Web. . . . . . . . . . . . . . . . . . 370 Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377 Conclusion & Outlook. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403 Lessons Learned & Contribution of this Book . . . . . . . . . . . . . . . . . . . . . 403
- 11. Contents vii Ongoing Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406 Future Perspectives & Related Topics. . . . . . . . . . . . . . . . . . . . . . . . . . . 408 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409 Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411 Index of Names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411 Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
- 13. Introduction In 2009, the publicly accessible World Wide Web celebrated its 18th Birthday1 . During its childhood and adolescence it has grown to become an enormous accu- mulation of information, entertainment, and spam – in a variety of forms and for- mats and with differing quality. All this has become part of our daily life. And never before have such an amount and such a variety of information been as easily accessible. On the other hand, navigating through a collection of documents and finding just the right piece of information has probably never been a greater chal- lenge. We are experiencing an enormous “information overload” (e.g., Neill, 1992) on the WWW, a phenomenon also referred to as “infosmog” (e.g., Shadbolt, Burke & Friedland, 2003) or the “data deluge” (e.g., Gershon & Miller, 1993). Web search engines are doing a fine job, but difficulties remain in finding ap- propriate information in suitable formats and within a moderate time period. One challenge is the lack of either recall or precision in search results – achieving a simple and exact overview on retrieved documents is another one. Thus in the time of Google the precise and effective retrieval of information is still one of the biggest challenges for information societies. But although we can hardly imagine life without the WWW, its development is still far from being finished. Optimiza- tion of Web technologies and Web contents are a constant issue. In very recent years, however, a huge development has happened which has revolutionized the WWW: the users have conquered the Web. Social software such as wikis, blogs, podcasts and vodcasts, social networking services and social media sharing platforms have enabled common Web users without any back- ground in computer science or skills in programming languages to publish content on the Web and to share it with others. This revolutionary development has been called Web 2.0 (O’Reilly, 2005). Due to the significant new dimensions that were added to the ‘former’ Web it was acclaimed as a new release version. Sometimes it is also referred to as the Social Web, because its figureheads are broad commu- nities of users. The power of Web 2.0 applications lies in these social communi- ties rather than in any technical innovations. Huge groups of people are able to contribute Web content on a variety of topics; they may ensure the coverage of minority issues (the so-called “long tail”; see Anderson, 2006) as well as mutual quality control. 1 Whereas several technical preconditions and the underlying Internet have been avail- able before the “birth” of the public WWW in August 1991.
- 14. 2 Introduction By now, Web 2.0 principles and tools are no longer an innovation for many users, but have become a matter of fact. Services like Wikipedia2 , Facebook3 and Flickr4 are already part of many people’s daily life. In scientific discourse on the topic, the initial enthusiasm has been dampened. What promised to be the manifestation of the “wisdom of the crowds” (Su- rowiecki, 2004), or “collective intelligence” (Weiss, 2005), has by now also re- vealed its weaknesses and posed a variety of new challenges for Web developers and researchers in this field, for example concerning antisocial behavior like spamming, cyber stalking or misuse of personal data. Furthermore, with every- body contributing, the Web is growing even faster than it already had done in its previous form – and so is the information overload and the challenge to find the right information at the right time. But the Social Web has also introduced its own approach to this problem of finding and retrieving information: folksonomies. Folksonomies are collections of keywords (called tags), which users of certain Web 2.0 services may freely add to documents such as photos, videos or book- marks to describe their contents and make them manageable and retrievable. Users are tagging the documents. With this, folksonomies have transferred a fundamen- tal idea of library and information science to the Web: indexing documents with content-descriptive metadata. For decades, librarians, information specialists and professional indexers have assigned keywords to library holdings, digital docu- ment collections or archives in order to apply some structured access to the re- spective information. But there has never been enough capacity to do so for the huge amounts of documents on the Web. Now the Web users have taken over this challenge themselves, in a rather intuitive manner and without any rules. Folksonomies enable the searching of and browsing through large social docu- ment collections in a novel way, but they also show different problems and short- comings – and thus are not the ultimate solution for accessing information on the Web. One major shortcoming of folksonomies is the complete lack of vocabulary control. This means, for one, that synonyms are not bound together and a user searching for New York in a folksonomy-based system will have to think of alter- native denotations, such as NY or Big Apple, and add them to his search query. This problem and other effects of free tagging have been heavily reduced in con- trolled vocabularies, which are used for indexing in library and information sci- ence. Such methods of vocabulary control for indexing and retrieval are essentials in the field of knowledge representation as it is perceived in information science. Realized in the form of classifications, thesauri, nomenclatures or other knowl- edge organization systems (KOS), they have long been applied in different practi- cal settings. Thus it seems only natural that years of expertise in this discipline should now be combined with novel indexing approaches in social tagging appli- 2 Wikipedia: http://www.wikipedia.org. 3 Facebook: http://www.facebook.com. 4 Flickr: http://www.flickr.com.
- 15. Introduction 3 cations: The classical knowledge representation techniques are among the first as- pect that should be considered for use within the context of the (Social) Web. There is another ongoing research area which is of high interest for the problem of handling the enormous growth of information on the Web: the Semantic Web (Berners-Lee, Hendler & Lassila, 2001). The principles of Semantic Web and of Web 2.0 differ not only in their particular goals and features, but also fundamen- tally in their origins: The term ‘Web 2.0’ describes a kind of movement on the Web which emerged independently over a period of time and was first and most prominently noticed and reported by Tim O’Reilly. The Semantic Web, on the other hand, has not really become a reality yet. It is based on a vision (most popu- larly expressed by Tim Berners-Lee, James Hendler and Ora Lassila) of how the current WWW could be improved by adding a supportive layer of machine- readable, semantic metadata and by enhancing the topical linkage of Websites, in order to make navigation, information retrieval and information integration easier and more effective. The aim is to close the semantic gap (e.g. Ehrig, 2007) be- tween the words as symbols and their meanings; between computers that match word strings and humans who interpret meanings of words. An enormous research community has developed around this basic idea in order to create whatever tech- nical background would be needed to achieve this goal. Presently, the Semantic Web is still rather an enormous research goal than a real experience on the Web. But, as we will see throughout this book, these efforts have yielded first results in the form of several semantic applications and, what is more, have already resulted in a sophisticated technical infrastructure. In this research field, one major point of emphasis is the development of ontologies. Ontologies are complex forms of knowledge organization systems represented in a machine-readable formal language. They are needed to provide the semantic layer for the Web. Principally, ontologies can represent a vocabulary for a domain of interest in form of general concepts in this domain, particular instances that embody these concepts, relations which describe the properties of concepts, and axioms that capture general facts about them. If an ontology makes appropriate use of all these elements, it can become an elaborate form of controlled vocabu- lary for indexing Web documents. However, ontologies may further include gen- eral facts and statements (which is not the case in traditional KOS). And, as they make use of formal representation languages, they can include explicit concept definitions that may be interpreted by machines and can thus even be used to infer implicit knowledge. Apart from a specific technical infrastructure, the Semantic Web thus urgently requires a set of high-quality ontologies for different domains of interest, which provide a shared vocabulary and place concepts from a domain of interest into semantic contexts. The practical realization of the Semantic Web will heavily de- pend on future ideas for providing such semantic knowledge models. Due to their complexity, the development of ontologies is rather costly and laborious. This is one of the reasons why their distribution over the Web is only proceeding slowly. But if we succeed in spreading ontologies comprehensively thoughout the Web, the idea of the Semantic Web can become reality. Previous approaches have not
- 16. 4 Introduction succeeded, yet. But recent developments in the Social Web represent a highly promising opportunity for the Semantic Web: with the help of broad user commu- nities, lots of manpower could be gathered for the creation and application of on- tologies on the Web. This leads us right to the main objective of this book: we want to explore the opportunities and challenges of combining Social Web activities with the efforts to establish a semantic layer for the Web. We focus particularly on the develop- ment of rich semantic knowledge representations and on ways in which user communities may (directly or indirectly) contribute to this process. Ontologies on the Semantic Web, folksonomies on the Social Web, and classi- cal methods of knowledge representation in information science principally all share the same aim: to index documents with content-descriptive metadata in or- der to reduce semantic ambiguity and to improve retrievability. They are all mani- festations of knowledge representation; yet they all differ in their particular char- acteristics, features and problems. Classical approaches mainly rely on vocabulary control, ontologies add aspects of machine-readability and automatic reasoning, while folksonomies do the complete opposite and make use of uncontrolled, but socially collected keywords. All these different approaches should not be used in isolation, but must become interrelated on the Web, if a benefit is to be drawn from their particular advantages. Similarly, the research communities of the Social and the Semantic Web are in- creasingly collaborating and converting, in order to discuss what has been named the Social Semantic Web. The term ‘Social Semantic Web’ refers to the effect that the boundary between social software on the one hand, and the development of semantic applications on the other, is becoming increasingly blurred. Web tools of the future should consider focusing both on communities, interaction and commu- nication as well as on navigation based on meanings and the semantic interlinking of data collections. The aim is to profit both from broad user communities with their collected knowledge and manpower and from semantic technologies, along with navigation based on meanings, in order to make navigation and retrieval on the WWW more effective. For example, semantic layers can be added to large community-based document collections and help to structure their contents and es- tablish meaningful links between single pieces of information. On the other hand, user communities can help to construct novel ontologies and combine their skills and expertise in order to put these knowledge models to practice. Since 2008, the number of (announced or released) tools that focus particularly on combining so- cial and semantic technologies has been rising; the most popular examples are se- mantic wikis and semantic blogs which interlink community-created contents with simplified (lightweight) ontologies. These tools can be viewed as the herald of the Social Semantic Web. Although the combination of social and semantic applica- tions now seems so fundamental, it has only started to be addressed quite recently and still has a long way to go. The development of the Social Semantic Web up to now will be recapitulated in this book, and application scenarios of social semantic tools will be discussed. We will not, however, describe the technical backgrounds and computational as-
- 17. Introduction 5 pects of the Social Semantic Web in more detail. Instead, the general focus of this book is directed at a more specific sub-topic: it will address the problem of knowl- edge representation as one particular challenge in the emerging Social Semantic Web. We will comprehensively address the question of what knowledge represen- tation in a Social Semantic Web might look like, which includes both rich seman- tic structures and community contributions. This book starts mainly from the Se- mantic Web point of view and explores in detail, how Social Web activities con- tribute to ontology development – rather than a Social Web perspective, which might, for example, focus on the question of how semantic models can be applied to enhance navigation in social communities. The latter aspect will only be ad- dressed briefly. State of Research This book touches on a variety of different research areas from the field of infor- mation science and related disciplines (e.g. knowledge representation models and their development, social interaction on the Web, collaborative ontology engineer- ing and semantic upgrades) – and, unfortunately, we cannot provide a comprehen- sive overview on the state of research for every single sub-topic. Some of these relevant topics are already very well investigated; others are only just beginning to attract broad interest. We have sought to provide useful references for additional readings throughout the single chapters and sections. Additionally, we will now present a short overview on the state of research for the most important related topics and highlight some of the most pertinent related publications. Knowledge Representation and Knowledge Organization Systems in Information Science The topic of knowledge representation has been a priority in information science for a long time. Consequently, various publications on this topics exist, both in the form of both scientific articles (in journals, proceedings or edited volumes) and monographic books. Of the latter, some provide a broad overview on different knowledge representation and indexing methods, e.g. Cleveland and Cleveland (2001), Lancaster (2003), Stock and Stock (2008). Others focus on one particular method, e.g. Aitchison, Gilchrist and Bawden (2000) on thesauri and Batley (2005) on classifications. And some consider methods of knowledge representa- tion in a certain application setting, like Caplan (2003) and Taylor (1999) in the context of libraries and library science, Foulonneau and Riley (2008) in digital document collections and Lambe (2007) in corporate knowledge management. More detailed publications on single knowledge organization systems and their application will be cited in the respective sections throughout the book. Knowledge representation is still an important research topic in information science, library science and documentation today and has also made its way into scientific fields related to the Web. Among the recent research trends in this field are: the interoperability of different representation systems, structural enrichments of KOS and the involvement of user communities into knowledge representation
- 18. 6 Introduction activities. This book refers to these aspects and considers knowledge organization systems in the broader setting of Social Web and Semantic Web technologies. Web 2.0 Technologies and Knowledge Representation in Web 2.0 The developments in the WWW which are referred to as ‘Web 2.0’ or ‘Social Web’ have recently started to become a sizeable new research area with a very broad spectrum of sub-topics. Literature in this domain of interest is highly scat- tered among a variety of conference proceedings and journals – in disciplines ranging from computer science (technical and computational aspects of social software services) to economics (application of social software in organizational settings like knowledge management or marketing), also including social sciences (e.g. user behavior, media reception) and humanities (e.g. considerations on new text genres like blogs and microblogging). Monographic scientific considerations and broad topical collections on Social Web technologies are still rare. One rela- tively broad overview can be found in Dumova and Fiordo (2009). One example for a quite different perspective on the Web 2.0 phenomenon is presented by Gil- trow and Stein (2009), who have collected different discussions on new (literary) genres in the Social Web environment. In analogy to the phrase ‘Web 2.0’, several application scenarios for social software services have been announced by adding the ‘2.0’; such as ‘enterprise 2.0’, ‘library 2.0’ or ‘e-learning 2.0’. Literature exists for several application areas (e.g., Koch & Richter, 2007), focusing also on individual social software services like wikis (e.g., Klobas, 2006) or blogs (e.g., Blood, 2002). Due to the great public interest, a variety of popular science books are available as well. And apart from scientific conferences and journals, a lot of innovation takes place directly on the Web: online publications, in the form of blog posts or comments, also play an important role for scientific discussions in this field. The same goes for the particular aspect of indexing and retrieval in Web 2.0 using folksonomies. The most comprehensive overview on this topic is found in Peters (2009). A less scientific point of view on social tagging systems is presented by Smith (2008). Many of the latest research questions on the Social Web concern aspects of how different types of social software can be interrelated (or mashed-up). With re- spect to the overall topic of this book, we will consider some aspects of how social applications (e.g. wikis and blogs) can be enhanced with semantic technologies. Then, we will have a closer look at the role of social tagging in a Social Semantic Web scenario. To wit, we will investigate how social tagging applications can in- teract with ontologies or other semantic approaches, and show that folksonomies may profit from additional semantic structures, providing a useful starting point for the development of ontologies. Ontologies and Semantic Web Technologies Ontologies and ontology engineering have established themselves as prominent research topics, mainly in the field of computer science and related disciplines where particular emphasis is placed on ontologies as a new means for finding and
- 19. Introduction 7 managing information on the Web. And yet there also exist other points of view on this topic, e.g. basic questions of knowledge modeling with philosophical tradi- tions and considerations on formal representation languages and their complexity. One of the most important publications on ontologies in the context of Seman- tic Web research is surely the “Handbook on Ontologies”, a collection of articles covering a wide range of topics (including ontology languages, engineering meth- ods, exemplary ontologies and applications), edited by Staab and Studer (2004). It is now also available in a revised version (Staab & Studer, 2009). Another true classic amongst publications on ontologies is the book on “Ontology Engineering” by Gómez-Pérez, Fernández-López and Corcho (2004). It provides a useful over- view on the theoretical foundations of ontologies as well as technical backgrounds in ontology engineering, from representation languages to engineering tools. Sev- eral other important books with slightly different foci are available. Stucken- schmidt (2009) focuses on the expressiveness of ontologies and discusses concep- tual definitions, methods for representing meaning and the capacities of different ontology languages. Allemang and Hendler (2008) present a broad overview on issues of modeling ontologies in the representation languages OWL, RDF and RDF(S). Davies, Studer and Warren (2006) cover a broad spectrum of technical background on ontology engineering and usage. Additional general literature on ontologies includes books by Breitmann, Casanova and Truszkowski (2007) as well as Sharman, Kishore and Ramesh (2007). The broad variety of scientific articles also deals with very specific sub- disciplines of ontologies and ontology engineering, which will not be listed here. Similarly, a sheer endless number of journal articles and conference proceedings deal with the different aspects of developing a technological infrastructure for the Semantic Web, with the International Semantic Web Conference (ISWC) as well as the European (ESWC) and Asian Semantic Web Conferences (ASWC) as the leading discussion forums for recent developments. Important overviews on tech- nological achievements in this field are (amongst others) given by Antoniou and van Harmelen (2004), Hitzler, Krötzsch et al. (2008), and Stuckenschmidt and van Harmelen (2005). Kashyap, Bussler and Moran (2008) and Yu (2007) focus on the development of Semantic Web Services. Pellegrini and Blumauer (2006) have col- lected a variety of research papers on personal and organizational usage of seman- tic approaches. Other collections were published by Fensel (2004), Hepp, de Leenheer et al. (2008), and Taniar and Rahayu (2006). All these renowned publications indicate that the research on ontologies has reached a remarkable level. The same holds for the more general topic of Seman- tic Web research. Still, there are lots of open research challenges – the Semantic Web has not yet become a reality. Over the last years, the fundamental techno- logical infrastructure as well as several basic technologies (like representation lan- guages, inference engines, ontology editors, methodologies for creating, merging and mapping ontologies) have been developed, and future work will be dedicated to the practical realization of the Semantic Web and the establishing of ontology- based applications in various contexts.
- 20. 8 Introduction Important research interests in the area of the Semantic Web are, for instance, practical applications for certain contexts (e.g., semantic search and semantic desktop tools), the further development of standards and supportive technologies, data integration and exchange as well as the interlinking of different tools and data. This book focuses intensively on ontologies as the most complex form of KOS that we know so far. But it addresses the topic of ontologies and ontology engi- neering from an information scientist’s perspective and with respect to the general topic of knowledge representation in the Social Semantic Web. This means that little to no attention is paid to some of the topics that are of great importance in computer science, such as the technical realization of tools for ontology engineer- ing, reasoning or information extraction or the development of formal ontology languages. Instead, we will elaborate on the conceptual challenges in ontology en- gineering (including questions such as how to properly model concepts, instances and relations in ontologies) in order to outline their semantic capabilities. We will discuss the spectrum of traditional and novel knowledge organization systems and place ontologies into this context. And we will consider ontology engineering in the social context of Web communities, in order to demonstrate the potential of the Social Web for the creation of the Semantic Web. The Social Semantic Web After the individual topics of Web 2.0 and Semantic Web had started to be inten- sively discussed, research on the potential of combining both principles began to follow. Conferences and workshops have started to address this intersection (e.g., Auer, Bizer et al., 2007); single applications have been developed and published. The Journal of Web Semantics had a special issue on “Semantic Web and Web 2.0” (Mika & Greaves, 2008). And, very recently, books have been published for the first time presenting a general introduction to the Social Semantic Web, most impotantly Breslin, Passant and Decker (2009), which describes how social soft- ware tools make use of semantics and how different applications are increasingly interlinked on the Web. Furthermore, Blumauer and Pellegrini (2008), Cardoso and Lytras (2009), Cunha (2009), and Lytras and Ordonez de Pablos (2009) have collected a variety of practical examples of enhanced Social Web applications. Kinsella, Passant et al. (2009) discuss semantic extensions for wikis and blogs and describe how semantic data can be added to software projects and may be aggre- gated over the Web. All in all, this combined research area is still in its infancy and the full potential of social semantic technologies has not yet been fully unleashed; more publica- tions are to follow. The literature available provides a useful overview on current initiatives and approaches. Our book focuses on one particular aspect within this emerging Social Semantic Web, namely knowledge representation.
- 21. Introduction 9 Objectives and Main Research Questions of this Book The book intends to help closing the gap between knowledge representation meth- ods in the Social and the Semantic Web. This is, currently, the gap between user- created uncontrolled tags in folksonomies on the one side, and highly formalized ontologies created by trained experts on the other. To bridge this particular gap, we will incorporate background knowledge from classical knowledge representa- tion initiatives. At first sight, the differences between folksonomies and ontologies may seem enormous. But a comprehensive consideration of folksonomies, no- menclatures, classifications, thesauri and ontologies already reveals a scale of step-by-step semantic enrichments. So we already have the means to bridge folk- sonomies and ontologies. Yet in many current studies, knowledge representation methods are not considered in their full spectrum. This book provides an overview on classical and novel approaches in knowledge representation, with a detailed view on ontologies as the most elaborate of these methods. It highlights the differ- ences of several methods as well as their similarities, points of contact and interac- tions. By combining several approaches, the Social Semantic Web will be able to profit from long traditions in knowledge representation for document indexing: from user collaboration in folksonomies and from rich semantics in ontologies. Nevertheless, this book will not solve the information overload problem or close the semantic gap. But it might act as another little step on the long path to- wards this ultimate goal. The main purpose of this book is to sum up the vital and highly topical research issue of knowledge representation on the Web and to dis- cuss novel solutions by combining some advantages of folksonomies and Web 2.0 approaches with ontologies and semantic technologies. We discuss the structure of ontologies in order to explain the complexity of semantic models for the Web and to investigate ways in which Web users may contribute to respectable rich KOS. A broad range of questions and challenges are centered on this topic, the most important ones for our perspective being the following: x What knowledge organization systems are already available on the Web or in other application settings? What are their characteristics and features? x What kind of knowledge can be represented in knowledge organization sys- tems? Which degrees of structural complexity have to be distinguished to characterize different KOS types? x How can complex KOS (ontologies) be built? Which tools are available? Which functionalities will be needed in the Social Semantic Web? x How can users participate in ontology engineering processes? How can they be motivated and taught to contribute to structured KOS? x Can pieces of KOS be created through little effort? Can particular knowledge bases be reused? Can users even help to create ontologies without noticing it? x Which resources on the Web can be reused for ontology engineering? x How can different approaches learn from each other? x How and in which context can different KOS be combined? What is needed to let them interact? In this book, we have addressed these questions as detailed as possible.
- 22. 10 Introduction Outline of this Book This book is organized into three main chapters, which capture the topic of knowl- edge representation in the Social Semantic Web as follows: Chapter 1 In this chapter, we will provide an introduction to the importance of knowledge representation in information science. We will explain the correlations between knowledge representation, knowledge organization systems, document indexing, metadata and information retrieval. The different types of classical knowledge or- ganization systems are introduced (mainly: nomenclature, classification and the- saurus), and we will present several examples of how these KOS are put into prac- tice and applied to different domains and usage scenarios, e.g. for intellectual property or medical information. With this background in knowledge representation traditions we will then pro- ceed to the Semantic Web. We will explain the underlying ideas of the Semantic Web and demonstrate how they are related to the aims of classical KOS – and how they go beyond classical approaches. We will give an initial short introduction to ontologies as the key elements of knowledge representation in the Semantic Web (a more detailed discussion of ontologies will follow in Chapter 2) and discuss the Semantic Web’s state of development. Next, the Social Web or Web 2.0 will be introduced as another recent develop- ment in the WWW that has also brought forward a novel approach to knowledge representation and indexing: folksonomies. This section comprises considerations of the advantages and shortcomings of folksonomies, and of general aspects of user communities and Web collaboration as propagated under the Web 2.0 para- digm. Finally, this first chapter concludes with the explanation of the novel term ‘So- cial Semantic Web’. We will provide a definition for the Social Semantic Web and explain how aspects of semantic indexing and social Web collaboration are al- ready converging and what new developments can be expected in the near future. Chapter 2 After this comprehensive overview on the past, present and future of knowledge representation, the second chapter focuses primarily on the aspect of semantics of the Social Semantic Web. We will discuss the elements and structure of ontologies in detail, as they provide the richest semantic structures of all current knowledge organization systems. This is mainly done on an abstract level, i.e. without ex- plaining the capabilities of single ontology languages. For some aspects, illustra- tive examples are provided, based on the ontology language OWL or the ontology editor Protégé. Yet it is not the aim of the book to compare characteristics of indi- vidual ontology languages. Instead, we want to highlight the novel features of on- tologies in comparison with other KOS and hint at the complexity of possibilities
- 23. Introduction 11 for knowledge representation – in order to sensitize the growing community within the Social Semantic Web to these difficulties and challenges. We thus in- troduce the most important elements of ontologies: concepts, instances, proper- ties/relations and axioms, and see how they make up the structure of an ontology and how they can be used for representing knowledge in general. A particular fo- cus is placed on semantic relations as KOS elements. While the core of this chapter is the detailed analysis and introduction of ontol- ogy elements and their role in semantic knowledge representation, it also features a brief introduction to the history of ontology in computer science, some discus- sions on ontology definitions, and a comparison of structural differences across several KOS and other knowledge resources. Chapter 3 In the last chapter, the aspect of semantics for the Web is now combined with so- cial dimensions. We will discuss the principles of ontology engineering, i.e. the construction of semantic models, which have to be placed into the context of the Social Semantic Web. Two different points of view will be considered: a) how broad user communities may participate in processes of developing ontologies (or other KOS) and b) how general activities of Web users may be exploited to im- prove the state of knowledge representation on the Web. To provide an initial overview on the current state, different tools that support ontology engineering activities are collected and the general dimensions of ontol- ogy engineering outlined. We will then address the specific question of how social communities may be involved in the engineering process. Collaborative vs. com- munity-based approaches in ontology engineering will be distinguished, novel tools and utilities described. Furthermore, a variety of current approaches will be discussed that have the potential to reuse the activities of broad Web communities and to transfer them to the development of semantic knowledge representations; among them are community knowledge bases, controlled natural languages and games with a purpose. A particular focus is placed on the potential of folksono- mies as user created index terms. We will discuss how tag gardening can be ap- plied to improve the performance of folksonomies and to enrich their semantic structure. On the other hand, folksonomies may also be used as a source for creat- ing richer KOS. Finally, the chapter addresses a topic of high importance for the Social Seman- tic Web: the co-existence of several knowledge organization systems. Different types of interactions between KOS as well as single application scenarios are out- lined. Attempts to enable easy access at available KOS are discussed and the Knowledge Organization Systems Ontology (KOSO) is introduced as a suggested meta model for classifying and describing existing KOS. We will conclude this chapter with a visionary view on KOS interactions in the Social Semantic Web.
- 24. 12 Introduction Acknowledgements This book could not have been written without the inspiring collaboration with my colleagues and the interesting and motivating discussions with fellow researchers. I particularly thank all my colleagues at the Department of Informaton Science, Heinrich-Heine-University Düsseldorf: many thanks to Wolfgang G. Stock for his constant support and advice. Thanks to Isabella Peters, Jasmin Schmitz, Violeta Trkulja and Sonja Gust von Loh – and also to all other members of the department and those students who worked with me in several projects. Further thanks go to all my co-workers in the Ontoverse project (funded by the BMBF) for insights into their disciplines and interdisciplinary discussions; thanks to Indra Mainz, Dominic Mainz, Ingo Paulsen. Thanks to James Kilbury. Thanks to Paul Becker for proofreading this book. Of course, there are many more people I met during the writing of this book and who helped me with advice and inspiration. Hope- fully, I will manage to thank all of them in person one day. This book is dedicated to my family with many thanks for all the years of lov- ing support. Formal Remarks The three main parts of this book, as well as both introduction and conclusion, all have their own list of references. If not indicated otherwise, all Web sources cited in the reference sections and in other parts of this book (e.g. in footnotes and cap- tions) have last been accessed on November 20, 2009. References Aitchison, J., Gilchrist, A., & Bawden, D. (2000). Thesaurus Construction and Use (4th Edition). London: Aslib. Allemang, D., & Hendler, J. A. (2008). Semantic Web for the Working Ontologist: Modeling in RDF, RDFS and OWL. Amsterdam: Morgan Kaufmann/Elsevier. Anderson, C. (2006). The Long Tail: Why the Future of Business is Selling Less of More. New York: ACM. Antoniou, G., & van Harmelen, F. (2004). A Semantic Web Primer. Cambridge: MIT Press. Auer, S., Bizer, C., Müller, C., Zhdanova, A.V. (Eds.) (2007). The Social Semantic Web: Proceedings of the 1st Conference on Social Semantic Web (CSSW), Leipzig, Germany. GI-Edition Proceedings, 113. Bonn: Gesellschaft für Informatik. Batley, S. (2005). Classification in Theory and Practice. Oxford: Chandos. Berners-Lee, T., Hendler, J., Lassila, O. (2001). The Semantic Web. Scientific American, 284(5), 34–43.
- 25. Introduction 13 Blood, R. (2002). The Weblog Handbook: Practical Advice on Creating and Maintaining your Blog. Cambridge: Perseus Pub. Blumauer, A., Pellegrini, T. (Eds.) (2008). Social Semantic Web: Web 2.0 – Was nun? Berlin: Springer. Breitman, K., Casanova, M. A., Truszkowski, W. (2007). Semantic Web: Concepts, Technologies and Applications. London: Springer. Breslin, J. G., Passant, A., Decker, S. (2009). The Social Semantic Web. Berlin: Springer. Caplan, P. (2003). Metadata Fundamentals for All Librarians. Chicago: ALA Editions. Cardoso, J., Lytras, M. D. (Eds.) (2009). Semantic Web Engineering in the Knowledge Society. Hershey: Information Science Reference. Cleveland, D. B., Cleveland, A. D. (2001). Introduction to Indexing and Abstracting (3rd Edition). Englewood: Libraries Unlimited. Cunha, M. M. (Ed.) (2009). Handbook of Research on Social Dimensions of Semantic Technologies and Web Services. Hershey: Information Science Reference. Davies, J., Studer, R., Warren, P. (Eds.) (2006). Semantic Web Technologies: Trends and Research in Ontology-Based Systems. Chichester: Wiley Sons. Dumova, T., Fiordo, R. (Eds.) (2009). Handbook of Research on Social Interaction Technologies and Collaboration Software: Concepts and Trends. Hershey: IGI Publishing. Ehrig, M. (2007). Ontology Alignment: Bridging the Semantic Gap. New York: Springer. Fensel, D. (2004). Ontologies: A Silver Bullet for Knowledge Management and Electronic Commerce (2nd Edition). Chichester: Springer. Foulonneau, M., Riley, J. (2008). Metadata for Digital Resources: Implementation, Systems Design and Interoperability. Oxford: Chandos Publishing. Gershon, N. D., Miller, C. G. (1993). Dealing with the Data Deluge. IEEE Spectrum, 30(7), 28–32. Giltrow, J., Stein, D. (Eds.) (2009). Genres in the Internet: Issues in the Theory of Genre. Amsterdam: John Benjamins Publishing. Gómez-Pérez, A., Fernández-López, M., Corcho, O. (2004). Ontological Engineering: Advanced Information and Knowledge Processing (3rd Print). London: Springer. Hepp, M., de Leenheer, P., de Moor, A., Sure, Y. (Eds.) (2008). Ontology Management: Semantic Web, Semantic Web Services, and Business Applications. Boston: Springer. Hitzler, P., Krötzsch, M., Rudolph, S., Sure, Y. (2008). Semantic Web. Berlin, Heidelberg: Springer. Kashyap, V., Bussler, C., Moran, M. (2008). The Semantic Web: Semantics for Data and Services on the Web. Berlin, Heidelberg: Springer.
- 26. 14 Introduction Kinsella, S., Passant, A., Breslin, J. G., Decker, S., Jaokar, A. (2009). The Future of Social Web Sites: Sharing Data and Trusted Applications with Semantics. Advances in Computers, 76, 121–175. Klobas, J. (2006). Wikis: Tools for Information Work and Collaboration. Oxford: Chandos Publishing. Koch, M., Richter, A. (2007). Enterprise 2.0: Planung, Einführung und erfolgreicher Einsatz von Social Software in Unternehmen. München: Oldenbourg. Lambe, P. (2007). Organising Knowledge: Taxonomies, Knowledge and Organisational Effectiveness. Oxford: Chandos. Lancaster, F. W. (2003). Indexing and Abstracting in Theory and Practice. Champaign, Ill.: University of Illinois, Graduate School of Library and Information Science. Lytras, M. D., Ordonez de Pablos, P. (Eds.) (2009). Social Web Evolution: Integrating Semantic Applications and Web 2.0 Technologies. Hershey: Information Science Reference. Mika, P., Greaves, M. (2008). Editorial: Semantic Web Web 2.0. Journal of Web Semantics, 6(1), 1–3. Neill, S. D. (1992). Dilemmas in the Study of Information: Exploring the Boundaries of Information Science. New York: Greenwood Press. O'Reilly, T. (2005). What is Web 2.0: Design Patterns and Business Models for the Next Generation of Software. Retrieved from http://www.oreilly net.com/pub/a/oreilly/tim/news/2005/09/30/what-is-web-20.html. Pellegrini, T., Blumauer, A. (Eds.) (2006). Semantic Web: Wege zur vernetzten Wissensgesellschaft. Berlin, Heidelberg: Springer. Peters, I. (2009). Folksonomies: Indexing and Retrieval in Web 2.0. Berlin: De Gruyter Saur. Shadbolt, N., Burke, M. A., Friedland, N. S. (2003). Panel Discussion on Knowledge Acquisition Projects (Halo, RKF, and AKT). In J. H. Gennari; B. Porter, Y. Gil (Eds.), Proceedings of the KCAP'03 Workshop on Knowledge Capture and Semantic Annotation, Florida, USA (pp. 2–3). New York: ACM. Sharman, R., Kishore, R., Ramesh, R. (Eds.) (2007). Ontologies: A Handbook of Principles, Concepts and Applications in Information Systems. Boston: Springer. Smith, G. (2008). Tagging: People-powered Metadata for the Social Web. Berkeley: New Riders. Staab, S., Studer, R. (Eds.) (2004). Handbook on Ontologies. Berlin, Heidelberg, New York: Springer. Staab, S., Studer, R. (Eds.) (2009). Handbook on Ontologies (Second Edition). Dordrecht et al.: Springer. Stock, W. G., Stock, M. (2008). Wissensrepräsentation: Informationen auswerten und bereitstellen. München, Wien: Oldenbourg. Stuckenschmidt, H. (2009). Ontologien: Konzepte, Technologien und Anwendungen. Berlin, Heidelberg: Springer.
- 27. Introduction 15 Stuckenschmidt, H., van Harmelen, F. (2005). Information Sharing on the Semantic Web. Berlin, Heidelberg: Springer. Surowiecki, J. (2004). The Wisdom of Crowds: Why the Many are Smarter than the Few and How Collective Wisdom Shapes Business, Economies, Societies and Nations. New York: Anchor Books. Taniar, D., Rahayu, J. W. (Eds.) (2006). Web Semantics and Ontology. Hershey: Idea Group Publishing. Taylor, A. G. (1999). The Organization of Information. Englewood, CO: Libraries Unlimited. Weiss, A. (2005). The Power of Collective Intelligence. netWorker, 9(3), 16–23. Yu, L. (2007). Introduction to the Semantic Web and Semantic Web Services. Boca Raton: Chapman Hall/CRC.
- 29. Chapter 1 Knowledge Representation and Indexing: Background and Future 1.1 Knowledge Representation The term knowledge representation is defined differently in various scientific fields. A sophisticated overview with largely theoretical background (philosophi- cal and logical) is given by Sowa (2000). Research in Artificial Intelligence (AI) discusses methods of knowledge modeling or knowledge engineering (e.g. Studer, Benjamins Fensel, 1998). Another different point of view is obtained through linguistic studies and their elucidation of knowledge representation (e.g. Löbner, 2002). The considerations within this book are placed in the specific and practical con- text of information science and its particular definition of knowledge representa- tion (Cleveland Cleveland, 2001; Lancaster, 2003; Stock Stock, 2008). This discipline has its origins in philosophy and, to a greater extent, in library-related developments. The latter can be traced back to the libraries in Egypt and Mesopo- tamia in the pre-Christian era. A comprehensive historical review is provided by Stock Stock (2008). While philosophical approaches have always aimed at rep- resenting or organizing world knowledge on a theoretical level, the practical aim of providing access to knowledge was central within the context of libraries. Thus, systematic catalogues for organizing library holdings can be regarded as the first practical methods of knowledge representation; they formed the basis for a long tradition of classification schemes5 . 1.1.1 Metadata, Indexing and Retrieval Today, knowledge representation has to be viewed as an approach to handling the key problems of the information society: how to structure and store information and, how to find and retrieve it precisely and effectively. Huge research efforts are 5 In this book we will also use the term ‘classification’ as a synonym for ‘classification scheme’ and ‘classification system’. Yet we must point out that this is not unambigu- ous; classification may also refer to the acts of classifying and classing (Wellisch, 2000).
- 30. 18 Knowledge Representation and Indexing: Background and Future concentrating on these challenges, many of them in the field of information re- trieval (Chu, 2007; Frakes Baeza-Yates, 1992; Lewandowski, 2005; Stock, 2007a). These studies focus on methods and algorithms to enable the precise and comprehensive searching of document collections. Methods of knowledge representation have been established as a complemen- tary approach to tackling the problem (Belkin Croft, 1997). They are applied to provide a better basis for information retrieval tools and are based on the represen- tation of document contents, and thus offer an alternative or additional approach to text statistics and content-based retrieval6 . Figure 1.1 shows how information re- trieval is based on document indexing and on the development of methods and models of knowledge representation. Figure 1.1. Value chain of knowledge representation, document indexing and information retrieval. Providing information about a document’s content may be done in two ways: by abstracting the topics of a document, i.e. writing an abstract which captures the main contents of a document (Stock Stock, 2008, Chapter 21), and by document indexing, i.e. assigning content-descriptive keywords to a document or placing it into a classification scheme (Cleveland Cleveland, 2001; Lancaster, 2003). Both approaches enhance retrieval techniques and aid users in deciding on a document’s relevance in different ways. Indexing is used to represent certain ob- jects from a document’s content by means of keywords; abstracting is used to rep- resent facts and statements from a document’s content in the form of complete sentences (Stock Stock, 2008). Both indexing and abstracting must be distinguished from formal metadata7 , i.e. information about the characteristics of a document which do not involve its content (or its aboutness) (Stock Stock, 2008). This might be the document’s author, its title, publication year or publishing company. Both formal and content 6 The term ‘content-based retrieval’ can easily be misunderstood: it is particularly used for image retrieval and denotes retrieval based on characteristics of the document file itself, e.g. colors, shapes, textures (Lew, Sebe et al., 2006; Rasmussen, 1997). This approach does not make use of metadata and knowledge representation. These studies speak of concept-based approaches when index terms are assigned to documents. 7 Also known as bibliographic metadata.
- 31. Knowledge Representation and Indexing: Background and Future 19 descriptive metadata should ideally be used in combination. Professionally gener- ated metadata are usually segmented into different fields, separating content- descriptive metadata from different types of formal metadata. In general, metadata can very broadly be defined as ‘data about data’, or, more precisely, as quoted be- low: “Metadata is here used to mean structured information about an information resource of any media type or format.” (Caplan, 2003) “Traditionally, librarians and archivists have used the term metadata for de- scriptive information used to index, arrange, file, and improve access to a library’s or museum’s resources.” (Gilliland-Swetland, 1998) A pragmatic and comprehensive discussion of metadata for document collections is found in Foulonneau and Riley (2008). In this book, we will mainly focus on the use of knowledge representation for the purpose of indexing. Indexing techniques typically rely on the use of con- trolled vocabularies, i.e. they provide a predefined set of concepts that are used to describe documents as well as for query formulation in information retrieval. Ap- plying methods of vocabulary control for indexing and retrieval helps to enhance consistency in the choice of index terms and to unify the user’s and the indexer’s vocabularies. As controlled vocabularies typically make use of explicit concept in- terrelations (hierarchical structures, synonymy and associations, which will be discussed intensively in Chapter 2.2), they can also provide suggestions for query expansions and modifications, and reduce semantic ambiguity (Lancaster, 1986). Indexing is a process made up of different phases. A general precondition is the availability of an appropriate knowledge representation model, e.g. a domain- specific classification or a thesaurus. Furthermore, indexing is always done for a certain collection of documents. Documents may be textual documents (e.g. books, articles, patents), but also pretty much anything that can be collected and stored, that carries a message or information and should thus be made searchable (e.g. videos, images, artistic works). The ISO definition (ISO 5963:1985) for ‘document’ is: “any item, printed or otherwise, which is amenable to cataloguing or indexing”. Wellisch (2000) provides a more detailed definition: “Document. A medium on or in which a message is encoded; thus, the combination of medium and message. The term applies not only to objects written or printed on paper or on microforms (for example, books, periodi- cals, maps, diagrams, tables, and illustrations) but also to non-print media (for example, artistic works, audio and video recordings, films, machine- readable records, and multimedia) and, by extension, to naturally occurring or humanly made objects intended to convey information (for example, zoo animals, plants in botanical gardens, museum collections of hand tools, etc.).” (Wellisch, 2000) Another extended definition is given by Buckland (1997). Furthermore, one often distinguishes documents and documentary units. One document may consist of several documentary units that should be indexed independently, e.g. one journal
- 32. 20 Knowledge Representation and Indexing: Background and Future volume consisting of several articles. Also, several documents may form one documentary unit that should be indexed as a whole, e.g. a complete series of books. In the context of the WWW the definition of ‘document’ might have to be refined in future, particularly in Social Web environments with their variety of user-generated content. Mai (2000) aptly sums up the elements of the indexing process: “The first step, the document analysis process, is the analysis of the docu- ment for its subjects. The second step, the subject description process, is the formulation of an indexing phrase or subject description. The third step, the subject analysis process, is the translation of the subject description into an indexing language. The three steps link four elements of the process. The first element is the document under examination. The second element is the subject of the document. This element is only present in the mind of the in- dexer in a rather informal way. The third element is a formal written de- scription of the subject. The fourth is the subject entry, which has been con- structed in the indexing language and represents the formal description of the subject.” (Mai, 2000) The indexing of professional and commercial services is typically (and ideally) done by professional indexers, i.e. people trained in the fields of knowledge repre- sentation and documentation and familiar with the use of controlled vocabularies. Intellectual indexing is a time-consuming and thus cost-intensive task. This way of professional, intellectual indexing provides high-quality indexing results. Yet it cannot guarantee perfection. Inconsistent indexing behavior may still occur on the inter-indexer level, if different indexers assign different index terms, as well as on an intra-indexer level, if one indexer uses an index vocabulary inconsistently over time (Lancaster, 2003; Stock Stock, 2008). Failures in indexing lead to deteriorating quality in information retrieval. If the quality and consistency of indexing fail, shortcomings in recall and precision will follow: “If an indexer fails to assign X when it should be assigned, it is obvious that recall failures will occur. If, on the other hand, Y is assigned when X should be, both recall and precision failures can occur. That is, the item will not be retrieved in searches for X, although it should be, and will be retrieved for Y, when it should not be.” (Lancaster, 2003) To reduce the costs of intellectual indexing, the development of automatic tech- niques to assist the classification and indexing of digital resources has been a re- search field for several years (Lancaster, 2003; Mani, 1999; Moens, 2000). Sev- eral approaches can be distinguished, particularly those that work with a form of knowledge representation and those that try to ‘classify’ documents based on simi- larities (Stock Stock, 2008, Chapter 20). In practical application, automatic ap- proaches have not yet displaced intellectual indexing.
- 33. Knowledge Representation and Indexing: Background and Future 21 1.1.2 Classical Knowledge Organization Systems Different types of knowledge organization systems (KOS) have been developed for the indexing of documents with content-descriptive keywords. In this chapter, we will introduce three of them: nomenclatures, classifications and thesauri. These traditional KOS are well established in practical contexts of library, archive and documentation sciences. For classifications and thesauri, national or interna- tional standards exist. Figure 1.1. The three most prominent classical KOS arranged according to complexity and broadness – with ontologies and folksonomies extending the spectrum at both sides. Source: Modified from Weller, 2007. All three are structured – controlled – vocabularies. They comprise the concepts relevant for a certain domain of interest (the vocabulary) and impose a structure by interrelating the concepts with different semantic relations. The foundation of a KOS is usually in the form of a concept hierarchy that may be enriched with fur- ther semantic relations, e.g. relations of equivalence and concept associations (we will discuss different types of relations and their importance for distinguishing dif- ferent types of KOS in Sections 2.2 and 2.3.2). The more a knowledge organiza- tion system makes use of semantic relations, the more complex it is in semantic structure. But the more complex the structure, the smaller the captured knowledge domain will have to be8 , due to reasons of feasibility in knowledge engineering (Figure 1.2). Recently, two new approaches, which will be discussed in the following sec- tions, have joined classical KOS: ontologies and folksonomies. They also address issues of document indexing and knowledge representation and have thus contrib- uted a revival in discussions about metadata on the Web (Madhavan, Halevy et al., 8 Either in terms of the broad coverage of domains or in terms of the in-depth representa- tion of a domain.
- 34. 22 Knowledge Representation and Indexing: Background and Future 2006; Safari, 2004). Their use has led to an increasing awareness of knowledge representation issues in scientific areas and even within the common Web-user community. Both embody different principles and complement traditional techniques from opposite sides (Figure 1.2). Folksonomies are based on the social approaches of Web 2.0 and include novel, user-centered (collaborative) dimensions while re- nouncing the use of vocabulary control9 . Ontologies extend the possibilities of formal vocabulary structuring in the sense of a Semantic Web. Figure 1.2. The keyword ‘World Wide Web 2.0’ in the SWD, with synonyms (BF) and associated keywords (VB). Source: http://z3950gw.dbf.ddb.de/z3950/zfo_ get_file.cgi?file Name=DDB/searchForm.html. Nomenclature A nomenclature (Stock Stock, 2008) is a comparatively simple form of knowl- edge organization system that works with controlled keywords extracted from natural language. These keywords are structured via the precise identification and labeling of synonyms, sometimes enriched by associative cross-references within 9 It may thus be argued about whether they should be regarded as knowledge organiza- tion systems or not. If one considers the existence of a concept structure as prerequisite for KOS, folksonomies have to be regarded as a different kind of knowledge represen- tation technique (like for example citation indexing which also cannot be regarded as KOS but is used for indexing). However, in this book we will regard folksonomies as a very lightweight form of KOS. As we will see below folksonomies at least contain some forms of implicit (statistical) connections between tags, due to their social dimen- sions.
- 35. Knowledge Representation and Indexing: Background and Future 23 the vocabulary. Homonyms are distinguished for unambiguous term definitions. The tradition of nomenclatures goes back to C.A. Cutter’s “dictionary catalog” from 1876 (Foskett, 1982). A nomenclature is often accompanied by a set of rules, which directs the index- ing process. One example is the Schlagwortnormdatei (SWD) with its rule system Regeln für den Schlagwortkatalog (RSWK)10 (RSWK, 1998; Umlauf, 1999). They were developed in the 1980s, and are still used in scientific and public libraries in Germany and German-speaking countries. The SWD lists preferred keywords to be used for document indexing, together with their synonyms which must not be used as index terms. Figure 1.3 shows the SWD account for the keyword ‘World Wide Web 2.0’: the abbreviation BF (which stands for ‘benutzt für’ = ‘used for’) points to synonyms, the abbreviation VB (‘Verwandter Begriff’ = ‘related term’) can be used for navigating to themati- cally related terms in the vocabulary. Furthermore, we find a definition (D) and indication of source (Q). Figure 1.4. CAS Registry file entry example. Source: Weisgerber, 1997. 10 Regeln für den Schlagwortkatalog (RSWK), 3rd Edition: http://deposit.ddb.de/ep/netpub/89/96/96/967969689/_data_stat/www.dbiberlin.de/dbi_ pub/einzelpu/regelw/rswk/rswk_00.htm.
- 36. 24 Knowledge Representation and Indexing: Background and Future The RSWK provide guidance for proper indexing based on the vocabulary. They contain, for instance, the instruction to use the most specific index terms and to avoid pleonasms in indexing (RSWK, 1998). The system envisions syntactic in- dexing in the form of keyword chains. Several chains are used to separate different topics appearing in a document. For this purpose, different categories of keywords are distinguished from general content-descriptive keywords: persons, geographic specifications, time specifications and formal keywords (regarding document types) (Umlauf, 2007). The SWD is constructed for a universal domain, not for domain-specific library collections. As an example of a nomenclature with a specialized, delimited do- main, we will look at the leading database for chemical literature, Chemical Ab- stracts Services (CAS)11 , and its registry file (CAS Registry) (Weisgerber, 1997). In the domain of chemistry, synonyms may not only be natural language words but also the molecular or structural formula of a chemical compound. The CAS Registry provides a unique identifier, the CAS Registry Number12 , to explicitly identify every chemical substance. These identifiers are linked with the com- pound’s name and synonyms, such as the chemical formula, generic names or col- loquial expressions, as depicted in Figure 1.4 (Stock Stock, 2008). Classification Classification systems (Batley, 2005; DIN 32705:1987; Foskett, 1982; Stock Stock, 2008, Chapter 12) have a long tradition in knowledge representation, both for indexing library collections and for use in online databases. They focus par- ticularly on the structuring of knowledge in the form of concept hierarchies. Thus the main effort in creating a classification lies in choosing the appropriate classes to capture a domain of interest, and establishing a suitable hierarchical structure. The most important peculiarity of classifications is that they work with nota- tions. These are sets of characters that represent concepts in a language- independent way. For example, in the International Patent Classification (IPC, which is used for world-wide patent indexing)13 the class ‘shoe lacing fastenings’ is represented by the notation A43C 1/00 (Figure 1.5). Some notations reflect the concepts’ position in the classification hierarchy and thus enable hierarchical query expansions with truncations (Stock Stock, 2008). With notations as unique identifiers, the classes’ names may be more easily translated into different languages for international usage, while consistent indexing is still possible across languages. For the long-term usage of a classification, it is important that the nota- tion system is able to incorporate new classes. This capability is called hospitality. 11 Chemical Abstracts Services (CAS): www.cas.org. 12 A unique identifier is needed in addition to the molecular formula, because one molecu- lar formula may denote different substances. The structural formula is unique but diffi- cult to handle. 13 International Patent Classification (IPC), English version: http://www.wipo.int/classifications/ipc/en/.
- 37. Knowledge Representation and Indexing: Background and Future 25 Figure 1.5. An excerpt from the International Patent Classification. Source: http://www.wipo.int/classifications/ipc/ipc8/?lang=en. We must distinguish between several specific forms of classification systems: A decimal classification is based on the principle that every class is divided into ten subclasses. The classical example for this type of classification is the Dewey Decimal Classification (DDC)14 (Chan Mitchell, 2006; DDC, 2003). The high- est hierarchical level of the DDC consists of ten classes; each of them has ten sub- classes, which are again subdivided into ten subclasses and so on. The ten top level classes of the DDC are: 000 Computers, Information General Reference 100 Philosophy Psychology 200 Religion 300 Social Sciences 400 Language 500 Science 600 Technology 700 Arts Recreation 800 Literature 900 History Geography The DDC (Figure 1.6) is also an example for another type of classification: the universal classification. The approach of universal classifications is to capture the entire domain of human knowledge. They are primarily used for the systematic shelf arrangement of library holdings, but are also of interest for classifying web- 14 Dewey Decimal Classification (DDC): http://www.oclc.org/dewey/.
- 38. 26 Knowledge Representation and Indexing: Background and Future sites, e.g. in the Yahoo! Directory15 and the Open Directory Project16 (see below). The DDC is the most important and most widely used universal classification, several more specific and refined versions having been derived from it for the use in some (European) countries, e.g. the Universal Decimal Classification (UDC) (McIlwaine, 2000; UDC, 2005). Main Classes 500 Science 579-590 Natural history of specific kinds of organisms 580-590 Plants and animals 590 Animals (Zoology) 592-599 Specific taxonomic groups of animals 597 *Cold-blooded vertebrates Pisces (Fishes) 597.9 *Reptilia (Reptiles) 597.948-597.96 Squamata (Scaly reptiles) 597.95 *Sauria (Lizards) 597.955 *Agamidae Figure 1.6. Excerpt from the Dewey Decimal Classification. Source: http://www.oclc.org/dewey/resources/tutorial/. Faceted classifications are another interesting type of classification systems (Broughton, 2002; Stock Stock, 2008). A faceted classification is not one con- tinuous classification system but comprises several sub-classifications: the facets. Facets should represent the fundamental categories of a domain of interest. The idea is to enable the combination of a fixed set of simple basic concepts to produce a multitude of complex concepts. In this way, it is not necessary to decide on one hierarchical system for the entire domain, but only to structure the concepts within one category or facet. Consider the following example, provided by Broughton (2006; see also Stock Stock, 2008): an excerpt from a hierarchical classification of the simple domain of “socks”. Grey socks Grey wool socks Grey wool work socks Grey wool hiking socks Grey wool ankle socks for hiking Grey wool knee socks for hiking Grey spotted wool knee socks for hiking 15 Yahoo! Directory: http://search.yahoo.com/dir. 16 Open Directory Project: http://www.dmoz.org.
- 39. Knowledge Representation and Indexing: Background and Future 27 As this pre-combined approach is rather complicated and would lead to much overhead in the classification system (i.e. one would certainly end up with classes that are never used); one might rather identify the basic categories of the domain and their subclasses (for this purpose, the single facets have to be disjoint; a con- cept cannot be part of more than one facet). This could look like the following faceted scheme proposed by Broughton (Broughton, 2004; Broughton, 2006): Colour Pattern Material Function Length Black Plain Wool Work Ankle Grey Striped Polyester Evening Calf Brown Spotted Cotton Football Knee Green Hooped Silk Hiking Blue Checkered Nylon Protective Red Novelty Latex By combining these facets we may then represent things like ‘blue striped cotton socks for football’. The individual complex concepts are only created via combi- nation at the time they are needed (and not in advance within a complete hierar- chy). This principle is called post-coordination (Bertram, 2005; Stock Stock, 2008). The citation order defines the order in which the elements of different fac- ets have to be arranged, which is needed for the construction of notations based on facet combinations. Specific symbols may be used to separate the faceted elements from each other within notations. The first (and still prevalent) example of a faceted classification system is the Colon Classification (CC) (Ranganathan, 1965; Ranganathan 1987[1933]). Other examples are the Bliss Bibliographic Classification, 2nd Edition (BC2) (Mills Broughton, 1977) and the Classification of Library and Information Science (Daniel Mills, 1975). The faceted approach to knowledge organization has also been adopted by other KOS. Some classifications partly make use of additional facets, without relying en- tirely on this combinatory approach (Stock Stock, 2008). They contain addi- tional tables with aspects that often co-occur with other concepts. For example, the DDC has auxiliary tables for geographical information that can be attached to the concepts from the main thematic classification. Thesaurus The last major type of classical knowledge organization system is the thesaurus (Aitchison, Gilchrist Bawden, 2000; Foskett, 1981; Lancaster, 1986, Stock Stock, 2008; Taylor, 1999). While in some contexts as well as in colloquial lan- guage the term ‘thesaurus’ denotes a lexicon of synonyms, in the field of knowl- edge representation it describes a complex controlled vocabulary. Thesauri in the latter sense are standardized by national as well as international norms (BS 5723:1987; ISO 2788:1986; DIN 1463/1:1987). They pay a lot of attention to the
- 40. 28 Knowledge Representation and Indexing: Background and Future collection of synonyms, but also use hierarchical concept structures and associa- tive relations between related terms to represent domain knowledge. Aitchison, Gilchrist and Bawden (2000) define a thesaurus as “[…] a vocabulary of controlled indexing language, formally organized so that a priori relationships between concepts are made explicit, to be used in information retrieval systems, ranging from card catalogue to the Internet.” (Aitchison, Gilchrist Bawden, 2000) Concepts in thesauri are typically not supported by complete definitions (as in a dictionary). Yet thesauri may contain (short) definitions for concepts – and more and more thesauri do make use of this feature. Aitchison, Gilchrist and Bawden (2000) point out: “Definitions tend to be necessary most frequently in social science and hu- manities thesauri, to clarify imprecise terminology, which occurs more of- ten in these subject areas.” (Aitchison, Gilchrist Bawden, 2000) Furthermore, scope notes are used in thesauri to indicate restrictions on meaning or the range of concepts, to give instructions to indexers, to provide examples for usage or to document the term history (Aitchison, Gilchrist Bawden, 2000). For the purpose of vocabulary control, most thesauri work with preferred terms in a set of synonyms. These preferred terms are sometimes called descriptors. De- scriptors are exclusively used for indexing and retrieval; all their synonyms (the non-preferred terms) are only used as pointers to the preferred terms. In thesauri that do not make use of preferred terms, all synonyms are regarded as equal and may be used for indexing. The typical way of displaying a thesaurus is in the form of an alphabetical list of all terms17 , with additional information on concept interrelations and usage: the term records. A prototypical entry for a preferred term in the alphabetical display of term records has the following form: PREFERRED TERM SN “Scope Note” D “Definition” UF “Used For”: References to equivalent non-preferred terms. BT “Broader Term”: References to broader terms. NT “Narrower Term”: References to narrower terms. RT “Related Term”: References to related terms. An entry for a non-preferred term only contains the pointer to the respective pre- ferred terms, e.g.: Non-preferred term USE PREFERRED TERM 17 Another frequently chosen display option is an additional hierarchical list of terms. For additional display approaches see Aitchison, Gilchrist Bawden (2000).
- 41. Knowledge Representation and Indexing: Background and Future 29 An example for a preferred term (‘deprived families’) and non-preferred term (‘underprivileged families’) may look like this (Aitchison, Gilchrist Bawden, 2000): DEPRIVED FAMILIES SN Socially disadvantaged and underprivileged UF Underprivileged families BT Families NT Homeless families One parent families RT Deprivation Underprivileged families USE DEPRIVED FAMILIES Figure 1.7 shows some more examples taken from the alphabetical term records in the ASIST Thesaurus of Information Science, Technology and Librarianship (Redmond-Neal Hlava, 2005). The term records may also contain more detailed information. For example, the quality of broader and narrower terms may be specified according to partitive and generic hierarchical structures (e.g. BTP = broader term partitive, BTG = broader term generic). A list of these abbreviations used in term records in English and German thesauri is provided by Stock and Stock (2008). Types of thesauri are mainly distinguished based on the depicted domain. Fac- eted approaches to thesauri are also available but rather rare (Aitchison, Gomersall Ireland, 1969; Spiteri, 1999; Stock Stock, 2008). Furthermore, we may con- sider multilingual thesauri, for which specific guidelines have been developed (e.g. ISO 5964:1985). The main problems in providing multilingual thesauri are due to the fact that concepts often cannot be translated exactly. Thus terms in two languages may not have one hundred per cent overlapping meanings. A prominent example is the German term ‘Wissenschaft’ which has a broader meaning than the English ‘sci- ence’. A common method for establishing multilingual thesauri is to choose one source language that serves as a starting point and provides the basis for the trans- lations, and one or more target languages that are mapped to the first (Aitchison, Gilchrist Bawden, 2000). For the connection of several languages, one source thesaurus is mostly used as standard reference: “In practice, and more happily in some systems than in others, the target thesauri are not directly related to each other, leaving the source thesaurus in the hub position of a radial network.” (Aitchison, Gilchrist Bawden, 2000)
- 42. 30 Knowledge Representation and Indexing: Background and Future Figure 1.7. Excerpt from the ASIST Thesaurus of Information Science, Technology and Librarianship. Source: Redmond-Neal Hlava, 2005.
- 43. Knowledge Representation and Indexing: Background and Future 31 Figure 1.8. Excerpt from TESE: descriptor “university” with translations for ten languages. Source: http://eacea.ec.europa.eu/portal/page/portal/Eurydice/TESE Descriptor?descriptorId =2700. Figure 1.9. Term record for ‘Bilingualism’ in the UNESCO Thesaurus with translations for French (FR) and Spanish (SP). Source: http://www2.ulcc.ac.uk/unesco/ terms/list13.htm. The Eurydice Information Network on Education in Europe is currently develop- ing the Thesaurus for Education Systems in Europe18 (TESE) (partly based on ex- isting thesauri). It is intended to be published in 14 languages, but currently de- scriptors are translated to ten European languages (Figure 1.8). Another important example is the UNESCO Thesaurus19 (UNESCO, 1995), which covers the topics of education, science, culture, social and human sciences, information and com- munication, politics, law and economics. It includes French (FR) and Spanish (SP) equivalents of English preferred terms (Figure 1.9). 18 Thesaurus for Education Systems in Europe (TESE): http://eacea.ec.europa.eu/portal/page/portal/Eurydice/TESEHome. 19 UNESCO Thesaurus: http://www2.ulcc.ac.uk/unesco/.
- 44. 32 Knowledge Representation and Indexing: Background and Future 1.1.3 Knowledge Organization Systems in Practice As we have seen, classical knowledge organization systems are developed for the practical aim of indexing certain document collections for retrieval purposes. Oth- ers also focus on the standardization of domain representations, e.g. for organiza- tional or statistical purposes – and thus constitute a shared point of view on this domain. Let us have a look at some of the most important fields of application. Library Catalogs and Publication Databases Library catalogs are the traditional application area of classification schemes. We have already seen some universal classifications used in library indexing. Mitchell (2000) points out: “Today, the Dewey Decimal Classification is the world’s most widely used library classification scheme.” While physical libraries require a knowledge organization system which con- siders the actual placement of books in the library shelves, this is not needed for digital publication databases. Most professional information providers focus on domain specific databases. These, in return, require specialized, domain specific KOS. Figure 1.10. Highest hierarchical level of MeSH 2008 with subclasses for one of the top classes. Source: http://www.nlm.nih.gov/mesh/2010/mesh_browser/ MeSHtree.B.html.
- 45. Knowledge Representation and Indexing: Background and Future 33 The Medical Subjects Headings20 (MeSH) are a domain specific thesaurus used for document indexing (Gaus, 2005). MeSH is produced and maintained by the National Library of Medicine, Bethesda, Maryland, USA. It has reached a consid- erable degree of popularity due to it being used for the reference database MED- LINE21 . MeSH registers approximately 23.000 headings (which correspond to de- scriptors). The hierarchical structure starts with 16 top categories (Figure 1.10). The system is polyhierarchic, which means that every term may belong to more than one superordinate concept (Figure 1.11). Figure 1.11. Polyhierarchy in MeSH. Source: http://www.nlm.nih.gov/cgi/mesh/ 2008/MB_cgi. MeSH headings may further be combined with qualifiers. Qualifiers are used to render the headings more precisely and to capture certain document features. Ex- amples for MeSH qualifiers are ‘Abnormalities’ or ‘Administration Dosage’. Every qualifier can only be used in combination with a given division from the 20 Medical Subjects Headings (MeSH): http://www.nlm.nih.gov/mesh. 21 MEDLINE, accessible via PubMed: http://www.ncbi.nlm.nih.gov/sites/entrez.
- 46. 34 Knowledge Representation and Indexing: Background and Future whole thesaurus. This also imposes a certain form of implicit structure on concept interrelations. Figure 1.12. The highest hierarchical level of the ICD-10. Source: http://www.who.int/ classifications/apps/icd/icd10online/. Health Care Some classifications have been established in the sectors of medicine and public health service. They are used as standardized representations to enable compara- bility for health statistics (e.g. causes of death) or for the use in administrative fields (e.g. accounting in hospitals) (Stock Stock, 2008). The World Health Or- ganization oversees the following classifications for diseases, health and disabil- ity, and health interventions (Gaus, 2005): x International Classification of Diseases (ICD, currently the tenth version: ICD-1022 ): It classifies diseases and other health problems and is used in dif- ferent types of health records, e.g. death certificates. It starts with 22 classes on the highest level (called chapters, Figure 1.12) and currently contains around 64,000 classes. It provides unitary notations for diseases, like J32.2 for ‘chronic frontal sinusitis’ or S62.7 for ‘multiple fractures of fingers’, which are often enriched with synonyms and directions for appropriate usage. ICD-10 is available in six official languages of WHO (Arabic, Chinese, Eng- 22 International Classification of Diseases (ICD): http://www.who.int/classifications/icd/en/.
- 47. Knowledge Representation and Indexing: Background and Future 35 lish, French, Russian and Spanish) as well as in 36 other languages. Accord- ing to Gaus (2005) it is the most intensively used KOS in the world. x International Classification of Health Interventions23 (ICHI, formerly Interna- tional Classification of Procedures in Medicine, ICPM): This classification covers medical interventions such as surgeries, therapies and diagnostics. It provides a standardized tool for reporting and analyzing the distribution and evolution of health interventions for statistical purposes. x International Classification of Functioning, Disability and Health24 (ICF): The ICF classifies the domain of health as described from a body, individual and societal perspective. For this purpose it contains two lists: a list of body functions and structure, and a list of domains of activity and participation. Additionally, the ICF includes a list of environmental factors. x The Diagnosis Related Groups (DRG): The DRG is a classification system developed for classifying hospital patients. It was created in the early 1980s for statistical purposes and to compare hospital costs for different types of pa- tients. The DRG consists of approximately 500 classes (called groups). Based on the US version, translations have been created for use in other countries (e.g. the German G-DRG). Furthermore the DRG has provided the basis for more specific classification systems in health care politics. The DRG is updated yearly. In 2007 major changes where performed for the 25th version of DRG, including the re-sequencing of groups. As of October 1, 2008 the 26th version is available with only minor changes. x TNM Classification of Malignant Tumours25 (TNM) (Sobin Wittekind, 2002): The TNM is a faceted classification for classifying the state of cancer in a patient’s body. Its development began in the early 1940s and it has been constantly updated since, in accordance with progress in the field of medi- cine. Since 1993, the TNM Supplement is available; a set of rules that should promote the uniform use of the TNM and lists practical examples. The TNM aims at providing a “common language” for oncologists all over the world in comparing their clinical material and in assessing the result of a treatment. The International Union against Cancer (UICC) therefore works in close co- operation with national and international organizations in order to maintain and promote the TNM. The abbreviation TNM stands for the three different categories that are needed for cancer staging: T stands for ‘primary Tumor’ and is used to de- scribe the size of the tumor (Figure 1.13); N stands for ‘regional lymph Nodes’ and classifies those regional lymph nodes which are involved in the disease to describe the absence or presence of regional lymph metastasis; and 23 International Classification of Health Interventions (ICHI): http://www.who.int/classifications/ichi/en/. 24 International Classification of Functioning Disabilty and Health (ICF): http://www.who.int/classifications/icf/en/. 25 TNM Classification of Malignant Tumours: http://www.uicc.org/tnm.
- 48. 36 Knowledge Representation and Indexing: Background and Future M stands for ‘distant Metastasis’ and includes those regions where metastasis may appear to describe the absence or presence of distant metastasis. One distinguishes between clinical classifications (pre-treatment, cTNM) and pathological classifications (post-surgical, pTNM). Other facets of de- scribing tumors can be added, for example the ‘histopathological Grading’ (G), ‘Lymphatic Invasion’ (L), ‘Venous Invasion’ (V), and the ‘Certainty Factor’ (C-facor, C). The TNM includes mappings to ICD. T – Primary Tumour TX Primary tumour cannot be assessed T0 No evidence of primary tumor Tis Carcinoma in situ T1, T2, T3, T4 Increasing size and/or local extent of the primary tumor Figure 1.3. General definitions for classifying primary tumours with TNM. Source: Sobin Wittekind, 2002. Figure 1.14. AO Classification of Fractures: Classification of body regions. Source: www.aofoundation.org. x Müller AO Classification of Long Bone Fractures (AO Classification) (Mur- phy Leu, 2000): The AO Classification is developed by the AO Founda-
- 49. Knowledge Representation and Indexing: Background and Future 37 tion26 , an organization led by an international group of surgeons who special- ize in the treatment of trauma and disorders of the musculoskeletal system. They also provide additional classifications for cranio-maxillofacial fractures (CMF classification), scapula fractures, and spinal fractures. Figure 1.15. AO Classification: examples for fractures 21-B2.1 to 21-B2.3. Source: Fundación Maurice E. Müller, Fracture Atlas, from http://www.muller foundation.org/atlas/main/21/b2.htm. Figure 1.16. AO Classification Training Game. Source: http://www.muller foundation.org/ccs/tool/game.php. The AO longbone fracture classification system describes the localization and morphology of fractures. Each type of fracture is represented by a five digit notation. The first digit is a number from 1 to 9, representing the body region (Figure 1.14). Each region is again divided into sections, numbered 1 to 3 (sometimes 4), which make up the second digit. The last three digits classify the type of fracture and the severity (always with three subtypes). 26 AO Foundation: http://www.aofoundation.org.
- 50. 38 Knowledge Representation and Indexing: Background and Future Figure 1.15 shows examples of fractures classified 21-B2.1 to 21-B2.3. The Maurice E. Müller Foundation also provides a training game including about 1,000 cases to help young health professionals improve their fracture diag- nosing skills (Figure 1.16). Intellectual Property In the field of intellectual property, classifications are used for indexing technical (patents and utility models) and non-technical documents (trademarks and de- signs) (Stock Stock, 2006; Stock Stock, 2008). The following classifications for international usage in the intellectual property field are maintained by the World Intellectual Property Organization (WIPO). They are all intensively used on international levels and are supported by official agreements between international patent offices. x International Patent Classification27 (IPC) (IPC, 2005): The IPC is used for the thematic classification of patents and utility models (Figure 1.5, above). The current IPC (2009 edition) lists approximately 70,000 classes. The WIPO provides an English and a French version; translations into other lan- guages are available from certain national patent offices. Based on the Strasbourg Agreement from 1971, international patent of- fices consistently classify patent literature according to the International Pat- ent Classification. Commercial database producers for legal and technical in- formation (like Derwent28 and Questel29 ) have also adopted this standard. Figure 1.17 shows an excerpt from a US patent including the IPC number A63D 15/00 for ‘Billiards; Billiard tables; Pocket Billiards’. x Nice Classification30 (International Classification of Goods and Services for the Purposes of the Registration of Marks under the Nice Agreement) (Nice Classification, 2007): The Nice Classification classifies goods and services for the purpose of registering trademarks and service marks. A trademark may be registered for one or more Nice classes (and thus for one or more product sectors). Thus the same product names may exist in parallel if they have been registered for different Nice classes. On the top level, the Nice Classification comprises 34 classes of goods (e.g. class 15: Musical Instruments) and 11 classes of services (e.g. class 35: Advertising; Business Management; Business Administration; Office Func- tions). x Vienna Classification31 (International Classification of the Figurative Ele- ments of Marks under the Vienna Agreement) (Vienna Classification, 2008): Trademarks often consist of or contain figurative elements. To enable the 27 International Patent Classification (IPC): http://www.wipo.int/classifications/ipc/en/. 28 Derwent: http://www.derwent.co.uk. 29 Questel: http://www.questel.com. 30 Nice Classification: http://www.wipo.int/classifications/nice/en/index.html. 31 Vienna Classification: http://www.wipo.int/classifications/vienna/en/.
- 51. Knowledge Representation and Indexing: Background and Future 39 unified description of such figurative elements and thus the retrieval of these visual trademarks, the Vienna Classification is used. An exemplary class hi- erarchy is: Category 21 Games, toys, sporting articles, roundabouts 21.3 Sporting articles, roundabouts 21.3.1 Footballs and other balls, shuttlecocks x Locarno Classification32 (Locarno Classification, 2004): The Locarno Classi- fication provides classes of goods and products. Trademark offices use these classes for registering industrial designs (by naming the classes into which the goods incorporating the designs belong). The use of the Locarno Classifi- cation by national offices has the advantage of filing industrial designs with reference to a single classification system. Figure 1.17. Excerpt from a US Patent (Patent Nr. 6942576), including notations from the IPC (51) and USPC (52). For patent indexing, some individual other classifications are used in addition to the IPC. The European Patent Office uses the European Classification (ECLA), which provides a downward expansion of the IPC (Dickens, 1994). The United States Patent and Trademark Office (USPTO) also uses its own classification sys- tem; the US Patent Office Classification System (USPC). The USPTO has also es- 32 Locarno Classification: http://www.wipo.int/classifications/locarno/en/.
- 52. 40 Knowledge Representation and Indexing: Background and Future tablished mapping tables that point from a class in the USPC to a comparable class in the IPC. The exemplary patent in Figure 1.17 includes both an IPC class and the USPC class 473/2 (class 473: ‘Games using tangible projectile; 1. Billiards or pool; 2. Practice device or device to aid in aiming cue or cue ball during shots (e.g. to assist in bank shots, etc.)’). Economics In the field of economics, several classification systems exist parallely in order to classify industrial sectors or products (Krobath, 2004; Stock Stock, 2008). Some are provided by national or international institutions, others by commercial information hosts. For the classification of industrial sectors, the most important models are NACE33 (Nomenclature general des activités économiques dans les Communautés Européennes) and NAICS34 (North American Industry Classification System). NACE is maintained and used by the European Union and its member states; NAICS is used for official statistics in North America. Yet, the NAICS’ predeces- sor, the Standard Industrial Classification35 (SIC), is also still widely used, particu- larly by commercial information providers. SIC was officially replaced by NAICS in 1997, as it was based on the structure of the USA as an industrial society in the 1930s and had thus become too out of date to be restructured. In its 2007 version, NAICS consists of the following 20 top level classes: 11 Agriculture, Forestry, Fishing, and Hunting 21 Mining, Quarrying, and Oil and Gas Extraction 22 Utilities 23 Construction 31-33 Manufacturing 42 Wholesale Trade 44-45 Retail Trade 48-49 Transportation and Warehousing 51 Information 52 Finance and Insurance 53 Real Estate and Rental and Leasing 54 Professional, Scientific, and Technical Services 55 Management of Companies and Enterprises 56 Administrative and Support and Waste Management and Remediation Services 33 NACE : http://ec.europa.eu/eurostat/ramon/index.cfm?TargetUrl=DSP_PUB_WELC. 34 NAICS : http://www.census.gov/epcd/www/naics.html. 35 Standard Industrial Classification (SIC): http://www.census.gov/epcd/www/sic.html.
- 53. Knowledge Representation and Indexing: Background and Future 41 61 Educational Services 62 Health Care and Social Assistance 71 Arts, Entertainment, and Recreation 72 Accommodation and Food Services 81 Other Services (except Public Administration) 92 Public Administration Figure 1.18 shows an excerpt from the subclasses of the information sector. Figure 1.18. Excerpt from NAICS. Source: http://www.census.gov/naics/ 2007/NAICOD07.HTM. Industry classifications can be enriched with product classifications. Currently, additional representations of product groups are being developed for the NAICS in the North American Product Classification System36 (NAPCS). So far it is avail- able for products of service industries (sectors 51-81); an excerpt of product classes with references to the NAICS industries producing them (separated by countries) is depicted in Figure 1.19. NAPCS also contains cross-references to the 36 North American Product Classification System (NAPCS): http://www.census.gov/eos/www/napcs/napcs.htm.
- 54. 42 Knowledge Representation and Indexing: Background and Future United Nations Central Product Classification37 , the product classification system hosted by the United Nations Statistics Division. Figure 1.19. Excerpt from NAPCS. Source: http://www.census.gov/eos/www/napcs /napcstable.html. The information producer Dun Bradstreet (a provider of business records) has developed a product classification based on SIC: the DB-SIC. They have multi- plied the number of classes by more than 15 in order to enable detailed product representation. Together with the United Nations Development Programme (UNDP), Dun Bradstreet also developed the United Nations Standard Products and Services Codes38 (UNSPSC); a classification of goods and services that can provide a uni- fied set of descriptions for e-commerce. 37 United Nations Central Product Classification (CPC): http://unstats.un.org/unsd/cr/registry/regcst.asp?Cl=3. 38 United Nations Standard Products and Services Codes (UNSPSC): http://www.unspsc.org.
- 55. Random documents with unrelated content Scribd suggests to you:
- 56. circunda la arenosa llanura, agitando su cola en el vacío, y levantando el venenoso dardo de que tenía armada su extremidad, como la de un escorpión. Mi Guía me dijo: —Ahora conviene que dirijamos nuestros pasos hacia la perversa fiera que allí está tendida. Por lo cual descendimos por la derecha, y dimos diez pasos sobre la extremidad del margen, procurando evitar la arena abrasada y las llamas: cuando llegamos donde la fiera se encontraba, vi a corta distancia sobre la arena mucha gente sentada al borde del abismo. Allí me dijo mi Maestro: —A fin de que adquieras una completa experiencia de lo que es este recinto, anda y examina la condición de aquellas almas, pero que sea corta tu conferencia. Mientras vuelves, hablaré con esta fiera, para que nos preste sus fuertes espaldas. Continué, pues, andando solo hasta el extremo del séptimo círculo, donde gemían aquellos desgraciados. El dolor brotaba de sus ojos, mientras acá y allá se defendían con las manos, ya de las pavesas, ya de la candente arena, como los perros, en el estío, rechazan con las patas o con el hocico las pulgas, moscas o tábanos, que les molestan. Mirando atentamente el rostro de muchos de aquellos a quienes azota el doloroso fuego, no conocí a ninguno; pero observé que del cuello de cada cual pendía una bolsa de cierto color, marcada con un signo, en cuya contemplación parecían deleitarse sus miradas. Aproximándome más para examinar mejor, vi en una bolsa amarilla una figura azul, que tenía toda la apariencia de un león. Después, prosiguiendo el curso de mis observaciones, vi otra, roja como la sangre, que ostentaba una oca más blanca que la leche. Uno de ellos, en cuya bolsa blanca figuraba una puerca preñada, de color azul, me dijo: —¿Qué haces en esta fosa? Véte; y puesto que aún vives, sabe que mi vecino Vitaliano debe sentarse aquí a mi izquierda. Yo soy paduano, en medio de estos florentinos, que muchas veces me
- 57. atruenan los oídos gritando: Venga el caballero soberano, que llevará la bolsa con los tres picos. Después torció la boca, y sacó la lengua como el buey que se lame las narices. Y yo, temiendo que mi tardanza incomodase a aquél que me había encargado que estuviera allí poco tiempo, volví la espalda a tan miserables almas. Encontré a mi Guía, que había saltado ya sobre la grupa del feroz animal, y me dijo: —Ahora sé fuerte y atrevido. Por aquí no se baja sino por escaleras de esta clase: monta delante; quiero quedarme entre ti y la cola, a fin de que ésta no pueda hacerte daño alguno. Al oír estas palabras, me quedé como aquel que, presintiendo el frío de la cuartana, tiene ya las uñas pálidas, y tiembla con todo su cuerpo tan sólo al mirar la sombra; pero su sentido amenazador me produjo la vergüenza que da ánimo a un servidor delante de un buen amo. Me coloqué sobre las anchas espaldas de la fiera, y quise decir: Ten cuidado de sostenerme; pero, contra lo que esperaba, me faltó la voz; si bien él, que ya anteriormente me había socorrido en todos los peligros, apenas monté, me estrechó y me sostuvo entre sus brazos. Después dijo: —Gerión, ponte ya en marcha, trazando anchos círculos y descendiendo lentamente: piensa en la nueva carga que llevas. Aquel animal fué retrocediendo como la barca que se aleja de la orilla, y cuando sintió todos sus movimientos en libertad, revolvió la cola hacia donde antes tenía el pecho, y extendiéndola, la agitó como una anguila, atrayéndose el aire con las garras. No creo que Faetón tuviera tanto miedo, cuando abandonó las riendas, por lo cual se abrasó el cielo, como se puede ver todavía; ni el desgraciado Icaro, cuando, derritiéndose la cera, sintió que las alas se desprendían de su cintura, al mismo tiempo que su padre le gritaba: Mal camino llevas, como el que yo sentí, al verme en el aire por toda partes, y alejado de mi vista todo, excepto la fiera. Esta empezó a marchar, nadando lentamente, girando y descendiendo;
- 58. pero yo no podía apercibirme más que del viento que sentía en mi rostro y en la parte inferior de mi cuerpo. Empecé a oír hacia la derecha el horrible estrépito que producían las aguas en el abismo; por lo cual incliné la cabeza y dirigí mis miradas hacia abajo, causándome un gran miedo aquel precipicio; porque vi llamas y percibí lamentos, que me obligaron a encogerme tembloroso. Entonces observé, pues no lo había reparado antes, que descendíamos dando vueltas, como me lo hizo notar la proximidad de los grandes dolores, amontonados por doquier en torno nuestro. Como el halcón, que ha permanecido volando largo tiempo sin ver reclamo ni pájaro alguno, hace exclamar al halconero: ¡Eh! ¿Ya bajas?, y efectivamente desciende cansado de las alturas donde trazaba cien rápidos círculos, posándose lejos del que lo amaestró, desdeñoso e iracundo, así nos dejó Gerión en el fondo del abismo, al pie de una desmoronada roca; y libre de nuestras personas, se alejó como la saeta despedida por la cuerda.
- 59. CANTO DECIMOOCTAVO AY un lugar en el Infierno, llamado Malebolge,[18] construído todo de piedra y de color ferruginoso, como la cerca que lo rodea. En el centro mismo de aquella funesta planicie se abre un pozo bastante ancho y profundo, de cuya estructura me ocuparé en su lugar. El espacio que queda entre el pozo y el pie de la dura cerca es redondo, y está dividido en diez valles, o recintos cerrados, semejantes a los numerosos fosos que rodean a un castillo para defensa de las murallas; y así como estos fosos tienen puentes que van desde el umbral de la puerta a su otro extremo, del mismo modo aquí avanzaban desde la base de la montaña algunas rocas, que atravesando las márgenes y los fosos, llegaban hasta el pozo central, y allí se reunían quedando truncadas. Tal era el sitio donde nos encontramos cuando descendimos de la grupa de Gerión: el Poeta echó a andar hacia la izquierda, y yo seguí tras él. A mi derecha vi nuevas causas de conmiseración, nuevos tormentos y nuevos burladores, que llenaban la primera fosa. En el fondo
- 60. estaban desnudos los pecadores; los del centro acá venían de frente a nosotros; y los de esta parte afuera seguían nuestra misma dirección, pero con paso más veloz. Como en el año del Jubileo, a causa de la afluencia de gente que atraviesa el puente de San Angelo, los romanos han determinado que todos los que se dirijan al castillo y vayan hacia San Pedro pasen por un lado, y por el otro los que van hacia el monte, así vi, por uno y otro lado de la negra roca, cornudos demonios con grandes látigos, que azotaban cruelmente las espaldas de los condenados. ¡Oh! ¡Cómo les hacían mover las piernas al primer golpe! Ninguno aguardaba el segundo ni el tercero. Mientras yo andaba, mis ojos se encontraron con los de un pecador, y dije en seguida: No es la primera vez que veo a ése. Por lo que me detuve a observarlo mejor: mi dulce Guía se detuvo al mismo tiempo, y aun me permitió retroceder un tanto. El azotado creyó ocultarse bajando la cabeza; mas le sirvió de poco, pues le dije: —Tú, que fijas los ojos en el suelo, si no son falsas las facciones que llevas, eres Venedico Caccianimico. Pero ¿qué es lo que te ha traído a tan picantes salsas? A lo que me contestó: —Lo digo con repugnancia; pero cedo a tu claro lenguaje, que me hace recordar el mundo de otro tiempo. Yo fuí aquel que obligó a la bella Ghisola a satisfacer los deseos del Marqués, cuéntese como se quiera la tal historia. Y no soy el único boloñés que llora aquí; antes bien este sitio está tan lleno de ellos, que no hay en el día entre el Savena y el Reno tantas lenguas que digan sipa,[19] como en esta fosa; y si quieres una prueba de lo que te digo, recuerda nuestra codicia notoria. Mientras así hablaba, un demonio le pegó un latigazo, diciéndole: Anda, rufián; que aquí no hay mujeres que se vendan. Me reuní a mi Guía; y a los pocos pasos llegamos a un punto de donde salía una roca de la montaña. Subimos por ella ligeramente, y volviendo a la derecha sobre su áspero dorso, salimos de aquel
- 61. eterno recinto. Luego que llegamos al sitio en que aquel peñasco se ahueca por debajo a modo de puente, para dar paso a los condenados, mi Guía me dijo: —Detente, y haz que en ti se fijen las miradas de esos otros malnacidos, cuyos rostros no has visto aún, porque han caminado hasta ahora en nuestra misma dirección. Desde el vetusto puente contemplamos la larga fila que hacia nosotros venía por la otra parte, y que era igualmente castigada por el látigo. El buen Maestro me dijo, sin que yo le preguntara nada: —Mira esa gran sombra que se acerca, y que, a pesar de su dolor, no parece derramar ninguna lágrima. ¡Qué aspecto tan majestuoso conserva aún! Ese es Jasón, que con su valor y su destreza robó en Cólquide el vellocino de oro. Pasó por la isla de Lemnos, después que las audaces y crueles mujeres de aquella isla dieron muerte a todos los habitantes varones; y allí, con sus artificios y sus halagüeñas palabras, engañó a la joven Hisipila, que antes había engañado a todas sus compañeras, y la dejó encinta y abandonada; por tal culpa está condenado a tal martirio, que es también la venganza de Medea. Con él van todos los que han cometido igual clase de engaños: bástete, pues, saber esto de la primera fosa, y de los que en ella son atormentados. Nos encontrábamos ya en el punto donde el estrecho sendero se cruza con el segundo margen, que sirve de apoyo para otro arco. Allí vimos a los que se anidan en una nueva fosa, dando resoplidos con sus narices y golpeándose con sus propias manos. Las orillas estaban incrustadas de moho, producido por las emanaciones de abajo, que allí se condensan, ofendiendo a la vista y al olfato. La fosa es tan profunda, que no se puede ver el fondo, sino mirando desde la parte más alta del arco, que lo domina perpendicularmente. Allí nos pusimos, y desde aquel punto vimos en el foso unas gentes sumergidas en un estiércol, que parecía salir de las letrinas humanas; y mientras tenía la vista fija allí dentro, vi uno con la
- 62. cabeza tan sucia de excremento, que no podía saber si era clérigo o seglar. Aquella cabeza me dijo: —¿Por qué te muestras tan ávido de mirarme a mí, con preferencia a los otros que están tan sucios como yo? Le respondí: —Porque, si mal no recuerdo, te he visto otra vez con los cabellos enjutos, y tú eres Alejo Interminelli de Luca; por eso te miro más que a todos los otros. Entonces, él, golpeándose la calabaza, exclamó: —Aquí me han sumergido las lisonjas que no se cansó de prodigar mi lengua. Después de esto, mi Guía me dijo: —Procura adelantar un poco la cabeza, a fin de que tus miradas alcancen las facciones de aquella sucia esclava desmelenada, que se desgarra las carnes con sus uñas llenas de inmundicia, y que tan pronto se encoge como se estira. Esa es Thais, la prostituta, que cuando su amante le preguntó: ¿Tengo grandes méritos a tus ojos?, ella le contestó: Sí, maravillosos. Y con esto queden saciadas nuestras miradas.
- 63. CANTO DECIMONONO H Simón el mago! ¡Oh miserables sectarios suyos, almas rapaces, que prostituís a cambio de oro y plata las cosas de Dios, que deben ser las esposas de la virtud! Ahora resonará la trompa para vosotros, puesto que os encontráis en la tercera fosa. Estábamos ya junto a ésta, subidos en aquella parte del escollo que cae justamente sobre su centro. ¡Oh suma Sabiduría! ¡Cuán grande es el arte que demuestras en el cielo, en la tierra y en el mundo maldito, y con cuánta equidad se reparte tu virtud! Vi en los lados y en el fondo la piedra lívida llena de pozuelos, todos redondos y de igual tamaño, los cuales me parecieron ni más ni menos anchos que los que hay en mi hermoso San Juan para servir de pilas bautismales; uno de éstos rompí yo no ha muchos años, por salvar a un niño que dentro de él se ahogaba; y baste lo que digo, para desengañar a todos.[20] Fuera de la boca de cada uno de aquellos pozuelos salían los pies y las piernas de un pecador, hasta el muslo, quedando dentro el resto del cuerpo. Ambos pies estaban
- 64. encendidos, por cuya razón se agitaban tan fuertemente sus coyunturas, que hubieran roto sogas y cuerdas. Del mismo modo que la llama suele recorrer la superficie de los objetos untados de grasa, así el fuego flameaba desde el talón a la punta en los pies de los condenados. —¿Quién es aquél, Maestro, que furioso agita los pies más que sus otros compañeros—dije entonces—, y a quien corroe y deseca una llama mucho más roja? A lo cual me contestó: —Si quieres que te conduzca por aquella parte de la escarpa que está más cercana al fondo, él mismo te dirá quién es y cuáles son sus crímenes. Le respondí: —Me parece bien todo lo que a ti te agrada: tú eres el dueño y sabes que yo no me separo de tu voluntad, así como también conoces lo que me callo. Subimos entonces al cuarto margen; después volvimos y bajamos por la izquierda hacia la estrecha y perforada fosa, sin que el buen Maestro me hiciera separar de su lado, hasta haberme conducido junto al hoyo de aquel que daba tantas señales de dolor con los movimientos de sus piernas. —¡Oh! Quienquiera que seas, tú, que tienes enterrada la parte superior de tu cuerpo; alma triste, plantada como una estaca— empecé a decir—, habla, si puedes. Yo estaba como el fraile que confiesa al pérfido asesino, que, metido en la tierra, le llama para que cese su muerte. Y él gritó: —¿Estás ya aquí derecho, estás ya aquí derecho, Bonifacio?[21] Me ha engañado en algunos años lo que está escrito. ¿Tan pronto te has saciado de aquellos bienes, por los cuales no temiste apoderarte con
- 65. embustes de la hermosa Dama,[22] y gobernarla después indignamente? Quedéme, al oír esto, como aquellos que, casi avergonzados de no haber comprendido lo que se les ha dicho, no saben qué contestar. Entonces Virgilio dijo: —Respóndele pronto: yo no soy, yo no soy el que tú crees. Y yo contesté como se me ordenó. Por lo cual el espíritu retorció sus pies; y luego, suspirando y con llorosa voz, me dijo: —¿Pues qué es lo que me preguntas? Si te urge conocer quién soy, hasta el punto de haber descendido para ello por todos estos peñascos, sabrás que estuve investido del gran manto, y fuí verdadero hijo de la Osa, tan codicioso, que, por aumentar la riqueza de los oseznos, embolsé arriba todo el dinero que pude, y aquí mi alma. Bajo mi cabeza están sepultados los demás papas, que antes de mí cometieron simonía, y se hallan comprimidos a lo largo de este angosto agujero. Yo me hundiré también luego que venga aquel que creí fueses tú, cuando te dirigí mi súbita pregunta. Pero desde que mis pies se abrasan, y me encuentro colocado al revés, ha transcurrido más tiempo del que él permanecerá en este mismo sitio con los pies quemados; porque en pos de él vendrá de poniente un pastor sin ley, por causa más repugnante, y ése deberá cubrirnos a entrambos. Será un nuevo Jasón, parecido al de que se habla en el libro de los Macabeos; y así como el rey de éste fué débil para con él, así con el otro lo será el que rige la Francia. No sé si en tal momento fué demasiada audacia la mía; pues le respondí en estos términos: —¡Eh!, dime: ¿cuánto dinero exigió Nuestro Señor de San Pedro, antes de poner las llaves en su poder? En verdad que no le pidió más sino que le siguiera. Ni Pedro ni los otros pidieron a Matías oro ni plata cuando por suerte fué elegido en reemplazo del que perdió su alma traidora. Permanece, pues, ahí, porque has sido castigado justamente, y guarda bien la mal adquirida riqueza, que tan atrevido
- 66. te hizo contra Carlos. Y si no fuese porque aun me contiene el respeto a las llaves soberanas, que poseíste en tu alegre vida, emplearía palabras mucho más severas; porque vuestra avaricia contrista al mundo, pisoteando a los buenos, y ensalzando a los malos. Pastores, a vosotros se refería el Evangelista, cuando vió prostituída ante los reyes a la que se sienta sobre las aguas; a la que nació con siete cabezas, y obtuvo autoridad por sus diez cuernos, mientras la virtud agradó a su marido.[23] Os habéis construído dioses de oro y plata: ¿qué diferencia, pues, existe entre vosotros y los idólatras, sino la de que ellos adoran a uno y vosotros adoráis a ciento? ¡Ah, Constantino! ¡A cuántos males dió origen, no tu conversión al cristianismo, sino la donación que de ti recibió el primer papa que fué rico! Mientras yo le hablaba con esta claridad, él, ya fuese a impulsos de la ira, o porque le remordiese la conciencia, respingaba fuertemente con ambas piernas. Creo que complací a mi Guía; porque escuchó siempre con rostro satisfecho el sonido de mis palabras, expresadas con sinceridad. Entonces me cogió con los dos brazos, y teniéndome en alto bien afianzado sobre su pecho, volvió a subir por el camino por donde habíamos descendido, sin dejar de estrecharme contra sí, hasta llegar a la parte superior del puente que va de la cuarta a la quinta calzada. Allí, depositó suavemente su querido fardo sobre el áspero y pelado escollo, que hasta para las cabras sería un difícil sendero, desde donde descubrí una nueva fosa.
- 67. CANTO VIGESIMO IS versos deben relatar un nuevo suplicio, el cual servirá de asunto al vigésimo canto del primer cántico, que trata de los sumergidos en el Infierno. Me hallaba ya dispuesto a contemplar el descubierto fondo, que está bañado de lágrimas de angustia, cuando vi venir por la fosa circular gentes que, llorando en silencio, caminaban con aquel paso lento que llevan las letanías en el mundo. Cuando incliné más hacia ellos mi mirada, me pareció que cada uno de aquellos condenados estaba retorcido de un modo extraño desde la barba al principio del pecho; pues tenían el rostro vuelto hacia las espaldas, y les era preciso andar hacia atrás, porque habían perdido la facultad de ver por delante. Quizá, por la fuerza de la perlesía, se encuentre un hombre de tal manera contrahecho; pero yo no lo he visto ni creo que pueda suceder. Ahora bien, lector, ¡así Dios te permita sacar fruto de esta lectura! Considera por ti mismo si mis ojos podrían permanecer secos, cuando vi de cerca nuestra humana figura tan torcida, que las lágrimas le caían por la espina dorsal. Yo
- 68. lloraba en verdad, apoyado contra una de las rocas de la dura montaña, de suerte que mi Guía me dijo: —¿Tú también eres de los insensatos? Aquí vive la piedad cuando está bien muerta. ¿Quién es más criminal que el que se apasiona contemplando la justicia divina? Levanta la cabeza, levántala y mira a aquel por quién se abrió la tierra en presencia de los tebanos, que exclamaban: ¿Adónde caes, Anfiarao? ¿Por qué abandonas la guerra? Y no cesó de caer en el Infierno hasta llegar a Minos, que se apodera de cada culpable. Mira cómo ha convertido sus espaldas en pecho: por haber querido ver demasiado hacia adelante, ahora mira hacia atrás, y sigue un camino retrógrado. Mira a Tiresias, que mudó de aspecto cuando de varón se convirtió en hembra, cambiando también todos su miembros, y hubo de abatir con su vara las dos serpientes unidas, antes que recobrara su pelo viril. El que acerca sus espaldas al vientre de aquél es Aronte, que tuvo por morada una gruta de blancos mármoles en las montañas de Luni, cultivadas por el carrarés que habita en su falda, y desde allí no había nada que limitara su vista, cuando contemplaba el mar o las estrellas. Aquella que, con los destrenzados cabellos, cubre sus pechos, por lo cual se ocultan a tus miradas, y tiene en ese lado de su cuerpo todas las partes velludas, fué Manto, que recorrió muchas comarcas, hasta que se detuvo en el sitio donde yo nací; por lo cual deseo que me prestes un poco de atención. Luego que su padre salió de la vida, y fué esclavizada la ciudad de Baco,[24] Manto anduvo errante por el mundo durante mucho tiempo. Allá arriba, en la bella Italia, existe un lago al pie de los Alpes que ciñen la Alemania por la parte superior del Tirol, el cual se llama Benaco. Mil corrientes, y aun más, según creo, vienen a aumentar, entre Garda, Val-Camonica y el Apenino, el agua que se estanca en dicho lago. En medio de éste hay un sitio, donde el Pastor de Trento, y los de Verona y Brescia, podrían dar su bendición si siguiesen aquel camino. En el punto donde es más baja la orilla que le circunda, está situada Peschiera, bello y fuerte castillo, a propósito para hacer frente a los de Brescia y a los de Bérgamo. Allí afluye necesariamente toda el agua que no puede estar contenida en el
- 69. lago de Benaco, formando un río que corre entre verdes praderas. En cuanto aquella agua sigue un curso propio, ya no se llama Benaco, sino Mincio, hasta que llega a Governolo, donde desemboca en el Po. No corre mucho sin que encuentre una hondonada, en la cual se extiende y se estanca, y suele ser malsana en el estío. Pasando, pues, por allí la feroz doncella, vió en medio del pantano una tierra inculta y deshabitada. Se detuvo en ella con sus esclavas, para huír de todo consorcio humano, y para ejercer su arte mágica, y allí vivió y dejó sus restos mortales. Entonces los hombres, que estaban dispersos por los alrededores, se reunieron en aquel sitio, que era fuerte a causa del pantano que le circundaba: edificaron una ciudad sobre los huesos de la difunta, y del nombre de la primera que había elegido aquel sitio, la llamaron Mantua, sin consultar para ello al Destino. En otro tiempo fueron sus habitantes más numerosos, antes de que Casalodi se dejara engañar neciamente por Pinamonte. Te lo advierto a fin de que, si oyes atribuir otro origen a mi patria, ninguna mentira pueda obscurecer la verdad. Le respondí: —Maestro, tus razonamientos son para mí tan verídicos, y me obligan a prestarles tanta fe, que cualesquiera otros me parecerían carbones apagados. Pero dime si entre la gente que va pasando hay alguno digno de notarse, pues eso solo ocupa mi alma. Entonces me dijo: —Aquél, cuya barba se extiende desde el rostro a sus morenas espaldas, fué augur cuando la Grecia se quedó tan exhausta de varones, que apenas los había en las cunas, y junto con Calcas dió la señal en Aulide para cortar el primer cable. Se llamó Euripilo, y así lo nombra en algún punto mi alta tragedia. Aquel otro que ves tan demacrado fué Miguel Scott, que conoció perfectamente las imposturas del arte mágica. Mira a Guido Bonatti, y ve allí a Asdente, que ahora desearía no haber dejado su cuero y su bramante; pero se arrepiente demasiado tarde: contempla las tristes que abandonaron la aguja, la lanzadera y el huso para convertirse en
- 70. adivinas, y para hacer maleficios con hierbas y con figuras. Pero ven ahora, porque ya el astro en que se ve a Caín con las espinas ocupa el confín de los dos hemisferios, y toca el mar más abajo de Sevilla. La luna era ya redonda en la noche anterior; debes recordar bien que no te molestó a veces por la selva umbría. Así me hablaba y entre tanto íbamos caminando.
- 71. CANTO VIGESIMOPRIMERO SI, de un puente a otro, y hablando de cosas que mi comedia no se cuida de referir, fuimos avanzando y llegamos a lo alto del quinto, donde nos detuvimos para ver la otra hondonada de Malebolge y otras vanas lágrimas, y la vi maravillosamente obscura. Así como en el arsenal de los venecianos hierve en el invierno la pez tenaz, destinada a reparar los buques averiados que no pueden navegar, y al mismo tiempo que uno construye su embarcación, otro calafatea los costados de la que ha hecho ya muchos viajes; otro recorre la proa, otro la popa; quién hace remos; quién retuerce las cuerdas; quiénes, por fin, reparan el palo de mesana y el mayor; de igual suerte, y no por medio del fuego, sino por la voluntad divina, hervía allá abajo una resina espesa, que se pegaba a la orilla por todas partes. Yo la veía, pero sin percibir en ella más que las burbujas que producía el hervor, hinchándose toda y volviendo a caer desplomada. Mientras la contemplaba fijamente, mi Guía me atrajo hacia sí desde el sitio en que me encontraba, diciéndome:
- 72. Ten cuidado, ten cuidado. Entonces me volví como el hombre que ansía ver aquello de que le conviene huír, y a quien asalta un temor tan grande y repentino, que ni para mirar detiene su fuga; y vi detrás de nosotros un negro diablo, que venía corriendo por el puente. ¡Oh! ¡Cuán feroz era su aspecto, y qué amenazador me parecía con sus alas abiertas y sus ligeros pies! Sobre sus hombros, altos y angulosos, llevaba a cuestas un pecador, a quien tenía agarrado por ambos jarretes. Desde nuestro puente dijo: —¡Oh! Malebranche, ved aquí uno de los ancianos de Santa Zita: ponedle debajo; que yo me vuelvo otra vez a aquella tierra, que está tan bien provista de ellos. Allí todos son bribones, excepto Bonturo; y por dinero, de un no hacen un ita.[25] Le arrojó abajo, y se volvió por la dura roca tan de prisa, que jamás ha habido mastín suelto que haya perseguido a un ladrón con tanta ligereza. El pecador se hundió y volvió a subir hecho un arco; pero los demonios, que estaban resguardados por el puente, gritaban: —Aquí no está el Santo Rostro; aquí se nada de diferente modo que en el Serchio. Si no quieres probar nuestros garfios, no salgas de la pez. Después le pincharon con más de cien harpones, diciéndole: —Es forzoso que bailes aquí a cubierto, de modo que, si puedes, prevariques ocultamente. No de otra suerte hacen los cocineros que sus marmitones sumerjan en la caldera las viandas por medio de grandes tenedores, para que no sobrenaden. —A fin de que no adviertan que estás aquí—me dijo el buen Maestro —, ocúltate detrás de una roca, que te sirva de abrigo; y aunque se me haga alguna ofensa, no temas nada; pues ya conozco estas cosas por haber estado otra vez entre estas almas venales.
- 73. En seguida pasó al otro lado del puente, y cuando llegó a la sexta orilla, tuvo necesidad de mostrar su intrepidez. Con el furor y el ímpetu con que salen los perros tras el pobre que de pronto pide limosna donde se detiene, así salieron los demonios de debajo del puente, volviendo todos contra él sus harpones; pero les gritó: —Que ninguno de vosotros se atreva. Antes que me punce vuestra orquilla, adelántese uno que me oiga, y después medite si debe perdonarme. Todos gritaron: —Vé, Malacoda. Por lo cual uno de ellos se puso en marcha, mientras los otros permanecían quietos, y se adelantó diciendo: —¿Qué te podrá salvar de nuestras garras? —¿Crees tú, Malacoda, que a no ser por la voluntad divina y por tener el destino propicio—dijo mi Maestro—, me hubieras visto llegar aquí, sano y salvo, a pesar de todas vuestras armas? Déjame pasar, porque en el cielo quieren que enseñe a otro este camino salvaje. Entonces quedó tan abatido el orgullo del demonio, que dejó caer el harpón a sus plantas, y dijo a los otros: —Que no se le haga daño. Y mi guía a mí: —¡Oh tú, que estás agazapado tras de las rocas del puente! Ya puedes llegar a mí con toda seguridad. Entonces eché a andar, y me acerqué a él con prontitud; pero los diablos avanzaron, de modo que yo temí que no observaran lo pactado: así vi temblar en otro tiempo a los que por capitulación salían de Caprona, viéndose entre tantos enemigos. Me acerqué cuanto pude a mi Guía, y no separaba mis ojos del rostro de aquéllos, que no era nada bueno. Bajaban ellos sus garfios, y:
- 74. ¿Quiéres que le pinche en la rabadilla?, decía uno de ellos a los otros. Y respondían: Sí, sí clávale. Pero aquel demonio, que estaba conversando con mi Guía, se volvió de repente, y gritó: Quieto, quieto, Scarmiglione. Después nos dijo: —Por este escollo no podréis ir más lejos, pues el sexto arco yace destrozado en el fondo. Si os place ir más adelante, seguid esta costa escarpada: cerca veréis otro escollo por el que podréis pasar. Ayer, cinco horas más tarde que en este momento, se cumplieron mil doscientos sesenta y seis años desde que se rompió aquí el camino. [26] Voy a enviar hacia allá varios de los míos para que observen si algún condenado procura sacar la cabeza al aire: id con ellos, que no os harán daño. —Adelante, Alichino y Calcabrina—empezó a decir—; y tú también, Cagnazzo; Barbariccia guiará a los diez. Vengan además Libicocco, y Draghignazzo; Ciriatto, el de los grandes colmillos, y Graffiacane, y Farfarello, y el loco de Rubicantondad en torno de la pez hirviente: éstos deben llegar salvos hasta el otro escollo, que atraviesa enteramente sobre la fosa.[27] —¡Oh Maestro! ¿Qué es lo que veo?—dije—; si conoces el camino, vamos sin escolta; yo, por mí, no la solicito. Si eres tan prudente como de costumbre, ¿no ves que rechinan los dientes, y se hacen guiños que nos amenazan algún mal? —No quiero que te espantes—me contestó—; deja que rechinen los dientes a su gusto. Si lo hacen, es por los desgraciados que están hirviendo. Se pusieron en camino por la margen izquierda; pero cada uno de aquéllos de antemano se habían mordido la lengua en señal de inteligencia con su jefe, y éste se sirvió de su ano a guisa de trompeta.
- 76. CANTO VEGISIMOSEGUNDO E visto alguna vez a la caballería levantar el campo, empezar el combate, pasar revista, y a veces batirse en retirada; he visto ¡oh, aretinos! hacer excursiones por vuestra tierra y saquearla; he visto luchar en los torneos y correr en las justas, ya al sonido de las trompetas, ya al de las campanas, al ruido de los tambores, con las señales de los castillos, y con todo el aparato nacional y extranjero; pero lo que no he visto nunca es que tan extraño instrumento de viento haya indicado la marcha a jinetes ni peones; jamás, ni en la tierra, ni en los cielos, guió semejante faro a ningún buque. Marchábamos juntamente con los diez demonios (¡oh terrible compañía!); pero en la iglesia con los santos, y en la taberna con los borrachos. Sin embargo, mi atención estaba concentrada en la pez para distinguir todo lo que contenía la fosa y los que se abrasaban dentro de ella. Así como saltan los delfines fuera del agua, indicando a los marinos que precavan la nave de la tempestad, así también algunos condenados, para aliviar su tormento, sacaban la espalda y la volvían a esconder más rápidos
- 77. que el relámpago; y lo mismo que en un charco las ranas sacan la cabeza a flor de agua, aunque teniendo dentro de ella sus patas y el resto del cuerpo, así estaban por todas partes los pecadores; pero en cuanto Barbariccia se aproximaba, volvían a sumergirse en aquel hervidero. Yo vi, y aun se estremece por ello mi corazón, a uno de aquellos que había tardado más tiempo en hundirse, como sucede con las ranas, que una queda fuera del agua, mientras otra se zambulle; y Graffiacane, que estaba más cerca de él, le enganchó por los cabellos enviscados de pez, y lo sacó fuera como si fuese una nutria. Yo sabía el nombre de todos aquellos demonios, por haberme hecho cargo de ellos cuando los eligió Malacoda. Rubicante, plántale encima tu garfio y desuéllalo, gritaban a un tiempo todos aquellos malditos. Yo dije: —Maestro mío, si puedes, procura saber quién es ese desgraciado que ha caído en manos de sus adversarios. Mi Guía se le acercó, y le preguntó de dónde era, a lo que respondió: —Yo nací en el reino de Navarra. Mi madre me puso al servicio de un señor: ella me había engendrado de un pródigo, que se destruyó a sí mismo y disipó su fortuna. Después fuí favorito del buen rey Tebaldo, y me lancé a comerciar con sus favores; crimen de que doy cuenta en este horno. Y Ciriatto, a quien salía de cada lado de la boca un colmillo como el de un jabalí, le hizo sentir lo bien que uno de ellos hería. Entre malos gatos había caído aquel ratón; porque Barbariccia lo sujetó entre sus brazos, diciendo: Quedaos ahí mientras que yo le ensarto. Y volviendo el rostro hacia mi Maestro, añadió: Pregúntale aún si deseas saber más, antes que otros lo destrocen. Mi Guía preguntó: —Dime, pues, si entre los otros culpables que están sumergidos en esa pez, conoces algunos que sean latinos.
- 78. A lo que contestó: —Acabo de separarme de uno que fué de allí cerca. ¡Así estuviera, como él, bajo la pez; no temería ahora ni las garras ni los garfios! Y Libicocco: Ya hemos tenido demasiada paciencia, dijo; y le enganchó por el brazo con su harpón, arrancándole de un golpe todo el antebrazo. Draghignazzo quiso también cogerle por las piernas; pero su Decurión se volvió hacia todos ellos lanzando una mirada furiosa. Cuando se hubieron calmado un poco, mi Guía no tardó en preguntar a aquel que estaba contemplando su herida: —¿Quién es ése de quien dices que te has separado, por tu desgracia, para salir a flote? Y le respondió: —Es el hermano Gomita, aquel de Gallura, vaso de iniquidad, que tuvo en su poder a los enemigos de su señor, e hizo de modo que todos le alabasen. Aceptó su oro y los dejó libres, según él mismo dice; y con respecto a los empleos, no fué un pequeño, sino un soberano prevaricador. Con él conversa a menudo don Miguel Zanche de Logodoro, y sus lenguas no se cansan nunca de hablar de las cosas de Cerdeña. ¡Ay de mí! Ved a ese otro cómo aprieta los dientes. Aun hablaría más, pero temo que se prepare a rascarme la tiña. El gran jefe de los demonios se dirigió a Farfarelo, que movía sus ojos en todas direcciones buscando donde herir, y le dijo: Quítate de ahí, pájaro malvado. —Si queréis ver u oír a toscanos y lombardos—empezó a decir en seguida el desgraciado pecador—, haré que vengan. Pero que esas malditas garras se mantengan un poco apartadas, a fin de que ellos no teman sus venganzas: yo, sentándome en este mismo sitio, por uno que soy haré venir siete, silbando como acostumbramos cuando uno de nosotros saca la cabeza fuera de la pez.
- 79. Al oír estas palabras, Gagnazzo levantó el hocico meneando la cabeza, y dijo: ¡Oigan el medio malicioso de que se ha valido para volver a sumergirse! A lo cual contestó aquél, que tenía abundancia de estratagemas: ¡En verdad que soy muy malicioso, cuando expongo a los míos a mayores tormentos! No pudo contenerse Alichino, y en contra de lo dicho por los otros, respondió: Si te arrojas en la pez, no correré al galope detrás de ti, sino que emplearé mis alas para ello. Te damos de ventaja la escarpa, y el ribazo por defensa, y veamos si tú solo vales más que todos nosotros. ¡Oh tú, que lees esto, ahora verás un nuevo juego! Todos los demonios se volvieron hacia la pendiente opuesta, y el primero de ellos, el que se había mostrado más renitente. El navarro aprovechó bien el tiempo; fijó sus pies en el suelo, y precipitándose de un solo salto, se puso al abrigo de los malos propósitos de aquéllos. Contristados se quedaron los demonios ante esta treta, pero mucho más el que tuvo la culpa de ella; por lo cual se lanzó tras de él gritando: Ya te tengo. Pero de poco le valió, porque sus alas no pudieron igualar en velocidad al espanto de Ciampolo: éste se lanzó en la pez, y aquél cambió la dirección de su vuelo, llevando el pecho hacia arriba. No de otro modo se sumerge instantáneamente el pato cuando el halcón se aproxima, y éste se remonta furioso y fatigado. Calcabrina, irritado contra Lichino por aquel engaño, echó a volar tras él, deseoso de que el pecador se escapara para tener un motivo de querella. Y cuando hubo desaparecido el prevaricador, volvió sus garras contra su compañero, y se aferró con él sobre el mismo estanque. Pero éste, gavilán adiestrado, hizo uso también de las suyas, y los dos cayeron en medio de la pez hirviente. El calor los separó bien pronto; pero todo su esfuerzo para remontarse era en vano, porque sus alas estaban enviscadas. Barbariccia, descontento como los demás, hizo volar a cuatro desde la otra parte con todos sus harpones, y bajando rápidamente hacia el sitio designado, tendieron sus garfios a los dos demonios, que estaban medio
- 80. cocidos en la superficie de aquella fosa. Nosotros los dejamos allí enredados de aquella manera.
- 81. Welcome to our website – the perfect destination for book lovers and knowledge seekers. We believe that every book holds a new world, offering opportunities for learning, discovery, and personal growth. That’s why we are dedicated to bringing you a diverse collection of books, ranging from classic literature and specialized publications to self-development guides and children's books. More than just a book-buying platform, we strive to be a bridge connecting you with timeless cultural and intellectual values. With an elegant, user-friendly interface and a smart search system, you can quickly find the books that best suit your interests. Additionally, our special promotions and home delivery services help you save time and fully enjoy the joy of reading. Join us on a journey of knowledge exploration, passion nurturing, and personal growth every day! ebookbell.com