Make our Scientific Datasets Accessible and Interoperable on the Web

1
Make our
Scientific Datasets
Accessible and
Interoperable
on the Web
Franck MICHEL
I3S - UMR 7271, CNRS - Univ. Nice Sophia
RBDD2015, CNRS Database Network
Oct. 21st 2015

2Franck Michel, RBDD 2015
Make our datasets
accessible and
interoperable on the Web…
 Not only because of H2020 requirement
 Linking data increases its value
• Mash up with related data
• Produce new knowledge
• Opportunity for new (unexpected) usage
 Citizenship demand for access to public
data (scientific, government…)
 …

 Publication/interlinking of open datasets
• Publish heterogeneous data in a common format
• Using common vocabularies
 Driven by major initiatives, e.g.:
• Linking Open Data
• W3C Data Activity
• Open Data hosting services... OpenAIRE, Zenodo...
 As well as other domain-specific projects
• Bio2RDF, BioPortal
Towards a Web of Data
From a Web of Documents
...to a Web of Data

4
May 2007
April 2008 Sept. 2008
March 2009
Sept. 2010
Linked Open Data Cloud
Sept. 2011
Aug. 2014
Linked Datasets as of Aug. 30th 2014. (c) R. Cyganiak & and A. Jentzsch

5
May 2007
April 2008 Sept. 2008
March 2009
Sept. 2010
Linked Open Data Cloud
Sept. 2011
Aug. 2014
Linked Datasets as of Aug. 30th 2014. (c) R. Cyganiak & and A. Jentzsch

Visible Web
Deep Web

Heterogeneous data models
Tabular
ID NAME
Directory
o=fr
cn=Franck Michel
ou=cnrs
objectclass=user
userid=65286
Graph
Object-Oriented
Documents
XML Native DBs,
Document stores…

 RDBs, NewSQL => SQL
 XML native DBs => XPath/XQuery
 Graph - Neo4J: Cypher, Allegrograph: SPARQL
 Document - MongoDB, CouchDB:
similar JS/JSON-based QL
 Column – Cassandra (CQL), HIVE (HQL)
 Key-value – Riak, DynamoDB: low level API
 …
Heterogeneous query capabilities

Make our datasets
accessible and

© NeuroLOG project
To you, your data may mean this…

To others,
your data may mean that…

The key is Metadata
Finding, understanding, and reusing scientific datasets
requires consistent, high-quality metadata
 Context: identification, authors, dates, license, version,
reference articles
 Access: format, structure, location (dwld), query method
 Meaning:
• What does it represent? What concepts, entities, semantics?
 Interpretation: units (cm or inches, left/right)…
 Provenance:
• Acquired with what equipment? Parameters, protocols?
• Derived from what dataset? With what processing?
• Dataset-level or entity-level provenance
 Statistics
 Etc.

CSV on the Web*
 Help access and understand CSV tabular data available
on the web
• Recommendations for Metadata vocabulary for CSV data
• Access methods for CSV Metadata
• Mapping mechanism to transforming CSV into various
Formats (e.g., RDF, JSON, or XML)
 Annotations on a table or group of tables, columns…
*https://www.w3.org/standards/techs/csv#w3c_all

CSV on the Web*
*https://www.w3.org/standards/techs/csv#w3c_all
GET tree-ops.csv
Content-Type: text/csv
Link: <http://example.org/tree-ops.json>; rel="…"
GID, Street, Species,Trim Cycle, Inventory Date
1, Addison Av, Celtis australis, 2010/10/18
2, Emerson St, Liquidambar styraciflua, 2010/06/02
GID, Street, Species,Trim Cycle, Inventory Date
1, Addison Av, Celtis australis, 2010/10/18
2, Emerson St, Liquidambar styraciflua, 2010/06/02
{ "@context":["http://www.w3.org/ns/csvw",{"@language":"en"}],
"url": "tree-ops.csv",
"dc:title": "Tree Operations",
"dc:license": { "@id":
"http://opendefinition.org/licenses/cc-by/"},
"dc:modified": {"@value":"2010-12-31","@type":"xsd:date"},
"tableSchema": {
"columns": [{
"name":"GID","titles":["GID", "Generic Identifier"],
"dc:description":"...",
"datatype": "string","required":true },
{
"name":« Street","titles":"On Street",
"dc:description":"The street that the tree is on.",
"datatype":"string" }, ...
],
"primaryKey": "GID","aboutUrl": "#gid-{GID}" }}

HCLS Profile*
 Health Care and the Life Sciences
 Consensus among participating stakeholders on the
description of datasets using RDF
 Data description, versioning, provenance, discovery,
exchange, query, and retrieval
*http://www.w3.org/TR/hcls-dataset/
RDF, RDFS, XSD
Citation Typing Ontology
Data Catalog
Dublin Core Metadata Types, Dublin Core Metadata Terms
Friend-of-a-Friend
Collection Description Frequency Vocabulary
Identifiers.org vocabulary
Lexvo.org - Lexical Vocabulary
Provenance Authoring and Versioning ontology (PAV)
PROV Ontology
Semanticscience Integrated Ontology (SIO)
Vocabulary of Interlinked Datasets (VoID)
Used
vocabularies

Challenges of publishing Metadata and/or Data?
Metadata
Data
Web
Syntax?
Shared meaning?
Link to others?
describe
Raw data?
Convert?

 Have a common representation format
 Structural heterogeneity
 Have common ways to describe the data
• Vocabularies, ontologies, thesaurus…
 Semantic heterogeneity
 Have common ways to query the data
Make our datasets
accessible and

 The Web of Data and the Semantic Web
 Create, reuse and link vocabularies
 Populate the Web of Data
 Publish Linked Open Data on the Web
Agenda

The Web of Data
And
the Semantic Web
Source: C. Faron Zucker[1], O. Corby[1]. Introduction au web de données et au web sémantique. Séminaire INRA Open Data Dec. 2014.
[1] INRIA Sophia Antipolis, CNRS, UNS.

Standards of the Semantic Web

Web of Data

RDFis a model based on triples, i.e. any fact
consists of 3 components:
( subject, predicate, object )
The Resource Description Framework

websem.html is a texte
websem.html has as author Fabien
websem.html has as author Olivier
websem.html has as author Catherine
websem.html has as subject Semantic Web
websem.html was written in 2011

websem.html
SemanticWeb
Texte
Catherine
Olivier
Fabien
type
date
author
subject
author
author
2011

http://ns.inria.fr/
ex/websem.html
http://en.wikipedia.org/
wiki/Semantic_Web
dt:Text
catherine.faron
olivier.corby
fabien.gandon
rdf:type
dc:date
dc:author
dc:subject
dc:author
dc:author
2011

N-Triples syntax
<http://inria.fr/ex/websem.html>
<http://purl.org/dc/elements/1.1/author>
<http://ns.inria.fr/catherine.faron> .
<http://purl.org/dc/elements/1.1/theme> “Semantic Web" .
@prefix dc: <http://purl.org/dc/elements/1.1/> .
dc:author <http://ns.inria.fr/catherine.faron> ;
dc:theme "Semantic Web" .
Turtle syntax

XML Syntax
<rdf:RDF
xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:dc="http://purl.org/dc/elements/1.1/">
<rdf:Description rdf:about="http://inria.fr/ex/websem.html">
<dc:author
rdf:resource="http://ns.inria.fr/catherine.faron"/>
<dc:theme>Semantic Web</dc:theme>
</rdf:Description>
</rdf:RDF>

Linked Open Data

chemas define
classes of resources,
their properties,
and organize their hierarchies
RDFS
RDF Schema

igeo:TerritoireAdministratif
igeo:Commune
rdfs:subClassOf rdfs:Class
rdf:type
rdf:type
RDF Schema
http://id.insee.fr/geo/
commune/34301
rdf:type
@prefix igeo: <http://rdf.insee.fr/def/geo#> .

igeo:codeINSEE
igeo:codeCommune
rdfs:subPropertyOf rdf:Property
rdf:type
rdf:type
RDF Schema

igeo:Commune
rdfs:range
igeo:chefLieu
igeo:PaysOuTerritoire
rdfs:domain
RDF Schema
departement/34
igeo:chefLieu
rdf:typerdf:type
commune/34172
Montpellier

SPARQL
Query RDF with SPARQL
SPARQL Protocol and RDF
Query Language

SPARQL 1.1 Rec. 21 Mar. 2013
 Query Language (using the Turtle syntax)
• SPARQL 1.1 Query Language
• SPARQL 1.1 Update
 Representation of query results
• SPARQL Query Results Format XML, CSV/TCV, JSON
 Protocols
• SPARQL 1.1 Protocol
• SPARQL 1.1 Graph Store HTTP Protocol
 Entailments
• SPARQL 1.1 Entailment Regimes

SPARQL: triple patterns
Turtle syntax with « ? » to mark variables:
?x rdf:type ex:Person
Describe patterns of triples that we look for:
SELECT ?subject ?type
WHERE { ?subject rdf:type ?type }
Default pattern: conjunction of triple patterns:
SELECT ?x WHERE
{ ?x rdf:type ex:Person .
?x ex:name ?name . }
?x
rdf:type
ex:Person
?name
ex:name

SPARQL: namespace prefixes
Declare prefixes of used vocabularies:
PREFIX mit: <http://www.mit.edu#>
PREFIX foaf: <http://xmlns.com/foaf/0.1/>
SELECT ?student
WHERE {
?student mit:registeredAt ?x .
?x foaf:homepage <http://www.mit.edu> .
}
Declare a base namespace for relative URIs:
BASE <http://www.example.org/people#>
SELECT ?student
WHERE { ?student foaf:knows <Ted> . }
?student
mit:registeredAt
?x
http://www.mit.edu
foaf:homepage

SPARQL: language and typed literals
SELECT ?x ?f WHERE {
?x foaf:name "Fabien"@fr ; foaf:knows ?f .
}
SELECT ?x WHERE {
?x foaf:name "Fabien"@fr ;
foaf:age "21"^^xsd:integer .
}

SPARQL: optional pattern
SELECT ?person ?name
WHERE {
?person foaf:homepage <http://fabien.info> .
OPTIONAL { ?person foaf:name ?name . }
}
Variable ?name is potentially unbound.

SPARQL alternative pattern
Merge the results of two graph patterns:
WHERE {
?person foaf:name ?name .
{ ?person foaf:homepage <http://fabien.info> . }
UNION
{ ?person foaf:homepage <http://bafien.org> . }
}

SPARQL filters
PREFIX ex: <http://inria.fr/schema#>
WHERE {
?person rdf:type ex:Person; ex:name ?name; ex:age ?age .
FILTER (xsd:integer(?age) >= 18)
}
Other examples:
FILTER(?name IN ("fabien", "olivier", "catherine"))
FILTER(if (langMatches(lang(?name),"FR"), ?age>=21)
FILTER NOT EXISTS {?x foaf:age -1}

SPARQL additional features
 Substract results
WHERE { ?x a ex:Person MINUS { ?x a ex:John } }
 Bind values
?person foaf:name ?name .
VALUES ?name { "Peter" "Pietro" "Pedro" "Pierre" }
 Property paths
?x foaf:knows+ ?friend .
 From
FROM <http//www.mit.edu/data.rdf>
SELECT ?student
WHERE { ?student mit:registeredAt ?x . }

SPARQL XML results
<?xml version="1.0"?>
<sparql xmlns="http://www.w3.org/2005/sparql-results#">
<head><variable name="student"/></head>
<results>
<result>
<binding name="student">
<uri>http//www.mit.edu/data.rdf#ndieng</uri>
</binding>
</result>
<result>
<binding name="student">
<uri>http//www.mit.edu/data.rdf#jdoe</uri>
</binding>
</result>
</results>
</sparql>

RDF/SPARQL @ French national inst. of statistics

RDF/SPARQL @ French National Library (BnF)

RDFS Entailment: infer new knowledge
PREFIX igeo:<http://rdf.insee.fr/def/geo#>
SELECT ?x
WHERE { ?x rdf:type igeo:TerritoireAdministratif }
igeo:TerritoireAdministratif
igeo:Commune
rdfs:subClassOf
ex:Sète
rdf:type

PREFIX igeo:<http://rdf.insee.fr/def/geo#>
SELECT ?x ?code
WHERE { ?x igeo:codeINSEE ?code}
igeo:codeINSEE
igeo:codeCommune
rdfs:subPropertyOf

igeo:Commune
rdfs:range
igeo:chefLieu
igeo:PaysOuTerritoire
rdfs:domain
SELECT ?x WHERE { ?x rdf:type igeo:Commune }
SELECT ?x WHERE { ?x rdf:type igeo:PaysOuTerritoire }
departement/34
igeo:chefLieu
commune/34172

So far, so good…

def. by enumeration
def. by intersection
def. by union
def. by complement
 class disjunction
def. by restriction!
def. by cardinality
1..1
symetric prop.
def. by equivalence
[>=18] def. by value restrict.
prop. disjunction
cardinality
1..1
!
indiv. prop. negation
chained prop.


…
Need for more? OWL in one slide…

Web of Data vs. Semantic Web
Web of Data: first step
in the deployment of
the Semantic Web

Make our datasets
accessible and

Definitions
 Taxonomy:
• Practice and science of classification
• Hierarchical categorization of controlled classes/terms
• Nested classes under broader categories
 Thesaurus
• Networked collection of controlled vocabulary terms, grouped according to
various types of relationship, e.g. similarity of meaning (synonyms,
antonyms),
 Ontology
• Formal semantic description for the taxonomy terms, properties and
interrelationships between categories in a domain of discourse, to facilitate
conceptual search and natural language queries.
 Folksonomy
• Collaborative/social tagging, social classification…
• Tag category schemes
• No (not necessarily) hierarchical categorization

Create, reuse and link vocabularies
 May seem easier: “I do whatever I want”
 Can be derived from an existing schema, e.g.:
• RDB: table -> class, column -> property,
primary key -> resource URI
• Thesaurus -> list of classes or SKOS Concepts
 But modeling implies choosing a point of view…
• E.g. biologist vs. geneticist, surgeon vs. anatomist, history…
• Domain experts must be involved
 Risk: create an island of data
 How to link my vocabulary/dataset
with other related ones?
Create my own vocabulary

Create, reuse and link vocabularies
 Where to look for: vocab./ontologies catalogs(see later)
 Difficulty to find the appropriate description
• Partial coverage of the domain I’m dealing with
• E.g. geographical area
• Granularity: level of details
• Too many (cumbersome), not enough (useless)
• Different points of view
Frequently, a mixed approach is used
• Reuse + create
• Need for interlinking => alignment
Reuse existing vocabularies

My vocabular Third-party vocabular
Link ontologies (very basic)
websem.html
SemanticWeb
ex:Book dc:Text
rdf:type
ex:topic
owl:equivalentProperty
ex:topic dc:subject
WSbook.html
Web
Sémantique
dc:subject
owl:equivalentClass
rdf:type
owl:sameAs

Link ontologies (basics)
 Classes
• owl:equivalentClass, owl:disjointWith, rdfs:subClassOf,
 Properties
• owl:equivalentProperty, owl:inverseOf, rdfs:subPropertyOf
 Individuals
• owl:sameAs, owl:differentFrom, owl:allDifferent
 rdfs:seeAlso
• Indicate a resource that might provide additional information about the
subject resource
 SKOS concepts
• skos:exactMatch: transitive
• skos:closeMatch, skos:relatedMatch
• skos:narrowMatch, skos:broaderMatch

Link ontologies … a complex topic
 Discovery of matching between classes,
properties
 Discovery of matching between
individuals
 Named Entity Recognition,
Entity matching, text-mining…
Ontology matching: “representing declaratively relations
between heterogeneous models”

SKOS: Simple Knowledge Organization System
RDF-based standard to represent controlled vocabularies:
glossary, dictionary, taxonomy, thesaurus…
Bridge the gap between existing KOSs and the SW and LD
Definition and documentation of classification systems
 SKOS concepts
• skos:Concept
 Labels and classification codes
• skos:prefLabel, skos:altLabel, skos:notation…
 Documentation
• skos:definition, skos:changeNote, skos:editorialNote, skos:example, etc.
 SKOS schemas
• skos:ConceptScheme, skos:hasTopConcept, skos:isTopConceptOf

SKOS: Simple Knowledge Organization System
 Hierarchy of collections of concepts
• skos:Collection, skos:OrderedCollection, skos:member…
 Semantic network and Hierarchies of concepts
• skos:related
• skos:broader, skos:narrower
 Alignment of schemas
• skos:closeMatch, skos:exactMatch
• skos:relatedMatch, skos:broadMatch, skos:narrowMatch
Semantic relations between concepts

Linked Open Vocabularies
 522 curated vocabularies
 Quality requirements
• URI stability and availability,
• Quality metadata and
documentation,
• Identifiable and trustable
publication body,
• Proper versioning policy,
• …
“Vocabularies provide the semantic glue
enabling Data to become meaningful Data.”
http://lov.okfn.org/dataset/lov/

BBC Wildlife Ontology
UniProt: protein sequence and
functional information.

Other catalogs of vocabularies
 Schemapedia (?) http://schemapedia.org
 schema.org
“Create, maintain, and promote schemas for structured data on
the Internet, on web pages, in email messages, and beyond”.
Controlled set of curated vocabularies: cars, TV series, arts,
administrations, diseases…
 DERI Vocabularies http://vocab.deri.ie/
URI space for RDFS vocabularies and OWL ontologies maintained at DERI.
No search interface.
 NCBO BioPortal http://bioportal.bioontology.org/ontologies/
 TDWG - Biodiversity Information Standards
http://www.tdwg.org/standards/
TDWG - Taxonomic Databases Working Group
And your favorite web search engine…
GeneralDomainspecific

Practical use case: TAXREF
CD_NOM : Unique identifier of the scientific name
CD_SUP : Identifier of the upper taxon in the classification
CD_REF : Identifier of the reference taxon
RANG : taxonomical rank

Practical use case: Taxref
CD_NOM : Unique identifier of the scientific name
CD_SUP : Identifier of the upper taxon in the classification
CD_REF : Identifier of the reference taxon
RANG : taxonomical rank
How to translate this table into a thesaurus
exploitable as a semantic reference using
semantic web technologies?

TAXREF SKOS Modelling
Taxon
skos:Concept
Reference name
skosxl:Label
Synonym
skosxl:Label
skosxl:prefLabel
Habitat
skos:Concept
taxref:habitat
Biogeo. status
skos:Concept
Taxonomical Rank
skos:Concept
"Linnaeus, 1758”
nt:has_rank
“Delphinus delphis”
skos:broader
<http://inpn.mnhn.fr/taxref/v8/taxon/60878>
txn:authority
“Van Bree, 1971”
“Delphinus tropicalis”
<http://inpn.mnhn.fr/espece/cd_nom/60881>
txn:authority
skosxl:literalForm
"Short-beaked
common dolphin"@en
taxref:
vernacularName
skosxl:literalForm
<http://inpn.mnhn.fr/espece/cd_nom/60878>
taxref:bioGeoStatusIn
skosxl:altLabel

TAXREF: Alignments with domain ontologies
 TaxonConcept Ontology: properties habitat and authority, taxonomic ranks
 NCBI Organismal Classification: property has_rank, taxonomic ranks
 GeoSpecies Knowledge Base: taxonomic ranks
 ENVO Environment Ontology: habitats
 Geonames: mainland France and overseas territories
 Darwin Core Terms: properties occurenceStatus, locationID
 TDWG Occurence Status Terms: biogeographical statuses
 World Geographical Scheme for Recording Plant Distributions
“Static” (hand-made) alignments:
predicates, reference values (habitats, taxonomical ranks)

TAXREF: Alignments with domain ontologies
Alignment of taxa and names done in a second step
 Automate search of matches within other taxonomical references
• DBpedia
• NCBI Organismal Classification
• Agrovoc
• BnF
• Encyclopedia of Life
• Vertebrate Taxonomy Ontology
 Difficulties
• Spelling differences
• Disagreements: reference vs. synonym, taxonomical rank
• owl:sameAs (individuals), owl:equivalentClass (classes), rdfs:seeAlso,
skos:exactMatch, skos:closeMatch, skos:relatedMatch (concepts)...

Populate
the
Web of Data

 HTML: RDFa, Microformats
 XML, XHTML
• XPath: RML
• XQuery: XSPARQL, SPARQL2XQuery
• XSLT: Gleaning Resource Descriptions from Dialects of Languages
(GRDDL), Scissor-Lift
 CSV/TSV/Spreadsheets: CSV on the web (W3C WG)
 RDBs (next slides)
 NoSQL stores (MongoDB…): xR2RML (next slides)
 Integration frameworks
• DataLift, Asio Tool Suite, Talend (with Semantic Web plugin?)
Many methods for many data sources

<body vocab="http://schema.org/">
<div resource="/jrbdd2015" typeof="Event">
<h2 property="title">RBDD 2015</h2>
<p>Date: <span property="startDate">2015-10-20</span></p>
...
<p>Conduire et construire un plan de gestion des données.
<a property="url"
href="http://rbdd.cnrs.fr/spip.php?article179">More…</a>
</p>
</div>
</body>
prefix sch: <http://schema.org/>
<http://rbdd.cnrs.fr/jrbdd2015>
rdf:type sch:Event ;
sch:title "RBDD 2015";
sch:startDate "2015-10-20" ;
sch:url <http://rbdd.cnrs.fr/spip.php?article179> .
RDFa: RDF in HTML attributes
: http://rbdd.cnrs.fr

 Various initial motivations
• Web of Data, Linked Data
• OBDA
• Ontology learning
• Schema mapping…
 Historical products: D2RQ, Virtuoso…
 R2RML
• 2012 W3C recommendation, mapping language,
several implementations
 Several methods: direct mapping vs. domain-specific
Translation of RDBs to RDF (RDB2RDF)

Direct Mapping of RDB to RDF
<PEOPLE/ID=7> rdf:type <PEOPLE> .
<PEOPLE/ID=7> <PEOPLE#FNAME> "Catherine" .
<PEOPLE/ID=7> <PEOPLE#ADDR> <ADDRESS/ID=18> .
<PEOPLE/ID=8> rdf:type <People> .
<PEOPLE/ID=8> <PEOPLE#FNAME> "Olivier" .
<PEOPLE/ID=8> <PEOPLE#ADDR> <ADDRESS/ID=22> .
Table: PEOPLE
ID FNAME ADDR (FK ADDRESS/ID)
7 Catherine 18
8 Olivier 22
… … …

Customized Mapping of RDB to RDF
<#TriplesMap1>
rr:logicalTable [ rr:tableName "PEOPLE" ];
rr:subjectMap [
rr:template "http://i3s.wimmics.org/staff/{ID}";
rr:class ex:Teacher;
];
rr:predicateObjectMap [
rr:predicate foaf:name;
rr:objectMap [ rr:column "FNAME" ];
].
<http://i3s.wimmics.org/staff/7> rdf:type ex:Teacher.
<http://i3s.wimmics.org/staff/7> foaf:name "Catherine".
<http://i3s.wimmics.org/staff/8> rdf:type ex:Teacher.
<http://i3s.wimmics.org/staff/8> foaf:name "Olivier".

xR2RML: Mapping of heterogeneous DBs to RDF
 Uniform language to describe mappings from most
common types of DB to RDF
 Extends R2RML and RML
 Features:
• Allow any declarative query language
• Allow any syntax to reference data elements from query results
(column name, attribute name, JSONPath, XPath...)
• Generate RDF lists and containers (bag, sequence, alternate)
• Support mixed content, e.g. XML value in relational column
 Implementation for MongoDB
• Data Materialization
• Query rewriting

xR2RML mapping example
<#TriplesMap>
xrr:logicalSource [
xrr:query "db.studies.find({ studyid:{ $exists:true } )";
];
rr:subjectMap [
rr:template "http://example.org/study#{$.studyid}";
rr:class ex:Study
];
rr:predicateObjectMap [
rr:predicate ex:involves;
rr:objectMap [
xrr:reference "$.centres.*.name" ];
];

xR2RML in practice: data materialization
Morph-xR2RML
xR2RML mapping
description
MongoQL
JSON
documents
Domain
ontologies
refers to
uses

xR2RML in practice: query rewriting
Morph
xR2RML
xR2RML mapping
description
MongoQL
SPARQL
SPARQL to
Mongo query
rewriting
JSON
documents

Publish
Linked Open Data
on the Web

Linked Data rules
1. Use URIs as names for things
2. Use HTTP URIs so that people
can look up those names
3. When someone looks up a URI, provide useful
information, using the standards (RDF*, SPARQL)
4. Include links to other URIs, so that they can discover
more things

Using URIs to look up information resources
 Dereferencing the URI returns a representation of the
document
• Either a direct link to
the representation:
the URI is a URL
• Or content negotiation
links to an appropriate
representation
Source: http://www.w3.org/TR/cooluris/

GET /people/cv_alice HTTP/1.1
Host: www.example.com
Accept: text/html, application/xhtml+xml
Accept-Language: en, de
HTTP/1.1 200 OK
Content-Type: text/html
Content-Language: en
Content-Location: http://www.example.com/cv_alice.en.html
<html ...>
...
HTTP content negotiation

GET /people/cv_alice HTTP/1.1
Accept: text/html, application/xhtml+xml
Accept-Language: en, de
HTTP/1.1 302 Found
Location:
http://www.example.com/cv_alice.en.html
GET /cv_alice.en.html HTTP/1.1
...
HTTP/1.1 200 OK
Content-Type: text/html
Content-Language: en
<html ...>
...
HTTP content negotiation

Using URIs to look up real-world objects
 One option: use Hash URIs for non-document resources

Using URIs to look up real-world objects
 Use HTTP 303 forwarding to an information resource
• 303 See Other: the requested resource is not a regular Web doc.
There is no suitable representation for the resource, but we can
provide information about the resource

Reference documentation
 Cool URIs for the Semantic Web.
W3C Interest Group Note 03 December 2008
http://www.w3.org/TR/cooluris/
 Dereferencing HTTP URIs
Draft Tag Finding 04 October 2007
http://www.w3.org/2001/tag/doc/httpRange-14/HttpRange-14.html

Data curation
 Main idea: better publish less
data but publish useful data
• Choose an appropriate modeling
• Choose appropriate vocabularies
• Include high quality metadata,
provenance information
• Deal with privacy issues
• Interlink
 Time consuming activity => significant cost
 Need skilled scientists who know the data, the software…
• Under-valued, no reward: data must become citable like
any scientific publication => Data Paper

Thank
you!

Make our Scientific Datasets Accessible and Interoperable on the Web

More Related Content

What's hot (19)

Viewers also liked (20)

Similar to Make our Scientific Datasets Accessible and Interoperable on the Web (20)

More from Franck Michel (16)

Recently uploaded (20)

Make our Scientific Datasets Accessible and Interoperable on the Web