nodalities_issue7.pdf

Nodalities
The mAgAzine of The SemAnTic Web
www.talis.com/nodalities

INSIDE August / September 2009

5 Trends and Barriers
What kind of data is
6 The Greatest Challenge
Facing IT
on the Guardian’s
Datastore?
9
Wages
£30bn

Building a Civic Simon Rogers, General and acute

Semantic Web Editor of Guardian’s data Blog
£27.5bn

HM Revenue Mental illness

and Customs £6.6bn

11
£30.9bn
NHS
Child Trust
£90.7bn
Waiving Rights over
(9) Fund
£0.24bn
Community

Linked Data
health
Department of Health services
£5.6bn

Child bene t
£10.6bn £105.7bn (6)
£2bn Learning
di culties
Tax credits
£19.5bn (9) Other
£2bn UK trade &
investment
£1.7bn
Accident & £0.088bn (15)
£1.6bn

13
NHS pensions emergency
£14bn Maternity

Might Semantic Personal £2.1bn

Technologies Permit
social
services
National School
Communities and of Government
General schools' spending

Meaningful Brand Local Government £0.003bn
£31.7bn

£34.3bn
Relationships? Department for Children, O ce for National
Spending by local and
regional government
Schools and Families Statistics (ONS)
Schools

£60.9bn
£0.159bn
£23.7bn
Other (3) £41.2bn £4.0bn Investment in
spending school buildings
Improving
supply & quality £3.7bn
of housing
£6.9bn Sixth form funding
£0.3bn £2.0bn (through learning
and skills council)
Other

15
£1.8bn Other schools

RDFa and Linked Data
Children & spending
families Teachers' pension £0.6bn £1.1bn
£2.9bn scheme
£10.7bn ICT Academies &

in UK Government
Other spending Young people specialist
on services for £1.2bn
children & families £5.8bn schools
£1.8bn

Websites Sure Start £0.6bn

Education
£0.8bn
£4.5bn
Learning & Skills
Council (excluding
sixth form funding)

maintenance
allowance Other spending
on services for
young people

We are drowning in information. The web

17 Search Engines, User What kind of data is on the Datastore? has given us access to data we would
• MPs’ expenses—full details never have found before, from specialist
Interfaces for Data,
• The world in carbon emissions datasets to macroeconomic minutiae. But,
Wolfram Alpha • The US in plastic surgery look for the simplest fact or statistic and
and More... • Deaths in iraq and Afghanistan Google will present a million contradictory
• Breakdown of UK population by race and area ones. Where’s the best place to start?
• Public debt since the 1970s That’s how our new datajournalism
• The world’s biggest banks ranked by assets operation came about. Everyday we work

22
• The world in refugees with datasets from around the world. We
Garlik Announce
• Interest rates since 1694 have had to check this data and make
Open-Source of 4Store • Every Guardian/ICM poll ever sure it’s the best we can get, from the
continued on page 3

nodAliTieS mAgAzine ediToRiAl

SUBSCRIBE NOW
Nodalities Magazine is made available, free of charge, in print Editor’s Notes
and online. If you wish to subscribe, please visit www.talis.com/ Welcome to the seventh edition of Nodalities Magazine.
nodalities or email nodalities-magazine@talis.com. It’s been about a year since I took the mantle of editor,
and I’ve certainly noticed a growth in Semantic Web
awareness, building and technology adoption. Perhaps
more importantly, however, is a rising awareness of the
importance of data to the workings of our societies. From
MEET US the implementation of data.gov in the US to the promise of
the UK Prime Minister to open up British public data, it’s becoming clear
For further information visit: www.talis.com/platform/events that data is growing in coverage and importance.
Following this observation, I’m very happy to include an article which
is less Semantic Web, and more data-centric. In fact, it’s our cover story
this edition, and I think it highlights the importance of publishing data and
encouraging its reuse. Simon Rogers, whose background is in putting
together info-graphics for the Guardian, writes about his newspaper’s Data
project which is making information of public interest and record available
FOLLOW US in data form for people to mashup and reuse.
In order for data to be truly public, however, it is important for an
Follow Nodalities Magazine on Twitter: @nodalities organisation publishing its information to consider the future reuse of
its data. Talis’ CTO Ian Davis discusses data licensing and waivers—
explaining the difference and making it easy to embed machine-readable
waivers into published datasets.
Elsewhere, Mark Birbeck explores the potential of RDFa to enable
innovation in public-sector websites and government datasets. He gives
examples of the UK’s Civil Service website, which uses RDFa to embed
metadata for job vacancies on the site itself. Alongside this, Joshua
Tauberer discusses his ideas for civil/social hacking and the importance
the Semantic Web can play in dealing with public-sector information.
Lee Feigenbaum, of Cambridge Semantics introduces Nodalities
readers to Anzo, a metadata tool for Microsoft’s Excel application. Anzo
exists to help with data collaboration in the enterprise, and Lee tells the
story behind its development. Paul Miller, from his cloud-computing
perspective talks about the Semantic Web and it’s relationship to brands
in the future. Paul challenges the concept of brand ownership in an era of
social media and widely-available data.
Finally, we also have a transcript of a recent podcast in which Paul
Miller talks with Garlik’s Tom Ilube and Steve Harris about security on the
Semantic Web and Garlik’s decision to open-source their triple-store (4Store).
As always, I welcome feedback and story submissions. You can follow
Nodalities on twitter @nodalities, or me on @zbeauvais. Also feel free to
email ideas and feedback to zach.beauvais@talis com, which also works
Nodalities Magazine is a Talis Publication with MSN Messenger and GTalk.

ISSN 1757-2592 Zach Beauvais

Talis Group Limited
Knights Court, Solihull Parkway
Birmingham Business Park B37 7YB
United Kingdom
Telephone: +44 (0)870 400 5000
www.talis.com/platform
nodalities-magazine@talis.com

The views expressed in this magazine are those of the contributors for
which Talis Group Limited accepts no responsibility. Readers should take
appropriate professional advice before acting on any issue raised.
©2009 Talis Group Limited. Some rights reserved. Talis, the Talis logo,
Nodalities and the Nodalities logo where shown are either registered
trademarks or trademarks of Talis Group Limited or its licensors in the United
Kingdom and/or other countries. Other companies and products mentioned
may be the trademarks of their respective owners.

2 Nodalities Magazine | August / September 2009

nodAliTieS mAgAzine gUARdiAn dATASToRe

Continued from front page. accurately. Who knows how many scandals collaborative utopian fantasy.”
have been obscured by purposely confusing So, when the commons authorities decided
most credible sources. But numbers? to publish the 500,000 pages of receipts, with
then it lives for the moment of But now, publishing data has got much very little time for us to process the information,
the paper’s publication and easier. The web has given us easy access to we decided it was time to let some Kool-Aid
afterward disappears into a hard billions of statistics on every matter. And with it slurpers loose on it. Software architect Simon
drive, rarely to emerge again are tools to visualise that information, mashing Willison came up with an interface that allowed
before updating a year later. it up with different datasets to tell stories that users to rate each of the receipts and add how
Comment, as Guardian could never have been told before. much they were for. Even better, it was fun to
founding editor CP Scott said, is free. But the It brings with it confusion and inaccessibility. do. The results: over 150,000 pages reviewed
second part of his maxim holds equally true for How do you know where to look, what is in two days, several stories revealed and—
the Guardian today: facts are sacred. credible or up to date? Official documents are crucially—great metadata on how MPs had filed
That is where the Data Store and the
Datablog come in. Part of the Open Platform—
the Guardian’s new API—it is about freedom of
information. But now, publishing data has got much easier. The web has given us
We like to think it’s part of our tradition. easy access to billions of statistics on every matter.
Take issue number one of the Manchester
Guardian, Saturday 5 May, 1821. News was on
the back page, like all papers of the day. And,
amid the stories and poetry excerpts, a third of often published as uneditable pdf files—useless their expenses, claims by party and averages
that back page is taken up with, well, facts. A for analysis except in ways already done by the across all categories.
comprehensive table of the costs of schools in organisation itself. It’s not just public data that we’re scooping
the area never before “laid before the public”, Alternatively sometimes an avalanche of up here. The Guardian produces tonnes of
writes “NH”. facts is unleashed in order to bury the truth. the stuff itself. For instance we did a series of
NH wanted his data published because Journalists have to walk this tightrope everyday, 1000 songs to hear before you die. Based on
otherwise the facts would be left to untrained ensuring that the numbers we publish are right. a spreadsheet and with Spotify links, this is
clergymen to report. His motivation is clear: If you’re reading this in the US, say, or suddenly great data. We also have university
“Such information as it contains is valuable; Canada or Italy or France then Britain’s MP ranking tables, raw information about our tax
because, without knowing the extent to which expenses scandal may mean very little to series and so on. In the past, this stuff would
education, and particularly the education of the you. In the UK, however, the revelations for have just stayed on the reporter’s hard disk.
labouring classes, prevails, the best opinions how British Members of Parliament have Now, if it’s data, it will be on the site.
which can be formed of the condition and been fiddling—it really is the best word—the Increasingly reporters around the world are
future progress of society must be necessarily expenses system have caused nothing less making it their mission to make data truly free;
incorrect.” In other words, if the people don’t than a shockwave. They have plunged the to publish everything. We have done it with the
know what’s going on, how can society get any prime minister’s party into shock and resulted information behind our series on corporate tax
better? in the not-so-swift exit of the Commons speaker, avoidance. Our Free Our Data campaign has
something that has only ever happened once even prompted the government to review its
before. information access policies.
The story was revealed by The Telegraph, So, together with its companion site,
Alternatively sometimes which—having purchased the leaked the Data Store—a directory of all the stats
an avalanche of facts is documents early—had 25 journalists working on we post—we have opened up that data
unleashed in order to bury the project for weeks before any of the stories for everyone. Whenever we come across
saw the light of day. It was, said Telegraph something interesting or relevant or useful, we
the truth. Journalists have to commentator Milo Yiannopoulos, a riposte to post it up on the site.
walk this tightrope everyday, those who believe anyone can be a journalist: And, rather than uploading spreadsheets
ensuring that the numbers we “Collaborative investigative journalism… onto our server, for now we are using Google
publish are right. feels good because it’s messy,” said [4ip’s Tom] Docs. This means we can update the facts
Loosemore, “and could work better than the old easily and quickly, which makes sure you get
models.” Oh, yeah? I’d like to see a “messy” the latest stats, as we get them. We’ve chosen
collective of Kool-Aid slurping Wikipedians Google Spreadsheets to host these datasets as
It’s a fine tradition, but one neglected by conduct the sort of rigorous analysis necessary the service offers some nice features for people
newspapers until very recently. I hated maths at for the Telegraph’s recent MPs’ expenses who want to take the data and use it elsewhere.
school. If someone had told me I would spend investigation. Can you imagine social media It’s just a start. We’re looking with interest at
the most productive part of my professional life achieving anything like it? Of course you Google Fusion tables and developing our own
working with spreadsheets, I would have been can’t: great journalism takes discipline and visualization apps.
either alarmed or derisory. While journalists training—neither of which exists in Loosemore’s
accept our mission to inform, we often proudly
boasted of our lack of mathematical prowess as
if it was a bold character trait. As if the choice at
school was either learning maths or English.
And the availability of data reflected that:
without computers and accessible numbers,
official data was impossible to challenge

August / September 2009 | Nodalities Magazine 3

nodAliTieS mAgAzine gUARdiAn dATASToRe

It is not comprehensive—there will always We’ll be looking at other methods for making
be things we’ve missed—but we want our users data we publish useful both for people and for KEY LINKS:
to help us with that, by posting requests to see machines, but we’d love to get some insights
if anyone out there knows how to find it. Above from you, as well. Tell us how we can make data guardian.co.uk/datablog
all, it is selective. It is information that we find more useful. A place to get the latest datasets and talk
interesting and curious. This is not the end of the road. We need about those we’ve put up
to find ways to make all our data much more
useable. Countries are named differently in guardian.co.uk/data-store
different datasets, for example. A truly useful The full directory of all our publicly-
A key reason for choosing data source will have to get round that— available datasets
Google Spreadsheets to probably by using ISO country codes, for
example. We’re always trying to get round guardian.co.uk/obamas-america
publish our data is not just
compromises between making our data easy US datasets
the user-friendly sharing and pleasant to look at and making it useful for
functionality but also the developers to take and produce beautiful things mps-expenses.guardian.co.uk
programmatic access it offers with. Our unique crowdsourcing experiment
directly into the data. But what we have done is to say to the with the MPs’ expenses receipts
world: we are not some exclusive club that has
access to this special stuff that we and only we
are going to see.
A key reason for choosing Google It is not a one-way process—we want you
Spreadsheets to publish our data is not just the to tell us what you have done with the data and
user-friendly sharing functionality but also the what we should do with it. As CP Scott said, the
programmatic access it offers directly into the facts are sacred: And they belong to all of us.
data. There is an API that will enable developers
to build applications using the data too. Simon Rogers is editor of the Guardian’s datablog


nodAliTieS mAgAzine TRendS And bARRieRS

Trends and Barriers Nodalities Magazine
By Zach Beauvais
SUBSCRIBE NOW
For anyone following the chunks of public data? What about when data
Nodalities blog, you may have comes from universities, institutions, scientific
read some of my recent posts foundations and NGO’s? What about charities
discussing the trends boiling monitoring crime, CO2 emissions and family
up around Web 3.0 (other histories? Wouldn’t these make a useful piece
buzzwords are available). The in the web of social data? What resources have
Mobile Web and upgraded the governments themselves got, if they want
connectivity in general; the rise of ubiquitous to make their public-owned data available in a
computing from chips in every product useful format?
imaginable; Linked Data and the “Semantic These questions form a major part of the
Web” as an organising platform for this rising thinking behind Talis’ Connected Commons
tide of data—these are three very broad trends initiative (talis.com/cc). Basically, Talis has
seeing a lot of media attention presently. From made its Semantic Web platform (including
where I’m standing, I tend to see the next great data hosting and access tools) available free of
turning point of the Web as a convergence of charge for any datasets made available to the
some of these trends, and see it as a rise in the public. In doing so, we’re hoping to remove the
importance of and reliance upon data itself and barrier of cost entirely to publishing interesting
data tools generally. data in a Linked Data way. One major reason
The mobile web is bringing new sorts of for this is to promote reuse and mashups of
information to people, and they can make use this interesting data, and for people to be
of this info wherever they happen to be because able to “follow their noses” to the data that
of advances in devices ad connectivity. As completes their projects. But, from a publishers’
phones and web-enabled devices get better, so perspective, this is important, because it’s
to do the chips we seem to have embedded all removing a major reason not to bother with
over the place, and we can now begin to have making data useful, if not only public. So, with
a more clear picture of what we do through the this, data can be made public and useable and
information we gather from our heaters, cars, the developers and users get the benefit of
and pedometers. Also, as more objects become public SPARQL endpoints and API access to
connected, the grunt-work of number-crunching interesting data.
and storage is becoming commoditized into big, To keep the data open and public, datasets
efficient, utility-like cloud services, which host need to make use of either the Public Domain
and work with our collected information much Dedication and License (PDDL) or Creative
more effectively than the gadget in your hand Commons’ CC0 license. Ian Davis, in his article
could ever hope to do. Others, like us here at in this magazine, explains more about waivers
Talis, talk about the Semantic Web, which allows and the Connected Commons, and there is a lot
for an evolution from a bunch of connected more about this particular initiative over on the
documents to the explicit connections between Talis site (talis.com/platform/cc/faqs/).
bits of information. In a recent interview with the BBC, Sir Tim
Also fermenting in this mix is a strengthening said: “This is our data. This is our taxpayers’
trend of political transparency and a public, money which has created this data, so I would Nodalities Magazine is made
shared ownership of social data. Barack like to be able to see it, please.” I wonder available, free of charge, in
Obama’s new administration has clearly made if initiatives such as Connected Commons print and online. If you wish to
this a priority with the launch and work around will begin to remove excuses, hindrances, subscribe, please visit
data.gov; and in the UK, Sir Tim Berners-Lee and obstacles? As public awareness of the www.talis.com/nodalities
himself has been appointed to an Parliamentary importance of access gets hotter, this might or email
advisory role. There is growing pressure to be become a political issue, as well as a pragmatic nodalities-magazine@talis.com.
able to have access to public data, and to see one. I hope that in the rush to publish data,
it as belonging to the nation’s people rather and in the ensuing discussion and debate that
than allowed to be legitimately filed away in the follows, that the users, hackers and developers
great, locked bureau of the capitols. don’t get sidelined. I think the world is ready for
So, picking up two fairly obvious trends its data back.
Nodalities Blog
here: Social, Public Data and Linked Data;
it would seem to follow that people would Zach Beauvais is the editor of Nodalities and From Semantic Web to
begin to have access to previously unavailable Platform Evangelist at Talis Web of Data
information in usable, linked forms. And it’s
certainly beginning, as articles elsewhere in this
blogs.talis.com/nodalities/
magazine have illustrated. But, what about other


nodAliTieS mAgAzine gReATeST chAllenge

The Greatest Challenge Facing IT
As the old adage goes: time is money.
By Lee Feigenbaum and Mike Cataldo, of Cambridge Semantics

Ultimately,
information
systems are
about saving
time. One
could argue
that technology
enables analysis that facilitates competitive
differentiation or improved product quality, but
the fact of the matter is that these things and
others could all be done without computers;
they would just take much, much longer.
A lot has been said and written about
information overload. Ultimately, though, the
issue with ever-expanding data is that the data
we need becomes hidden in mountains of other
data. Typically, these mountains take the form
of relational databases where the data is neatly
stored in rows and columns, and we find the
data in one of two ways. Either we directly look
up data by its “address” within the database, or
else we use a simple text search. But if we don’t
know what table or column the data resides in,
we can’t look it up. And as the quantity of data
grows, text searching the mountain of data itself
yields a mountain of results. Combing through
these results then compromises the real benefit Pursuing any of these typical solutions With data collaboration, the data is much
of information technology: time savings. means spending 6-18 months at a time solving more granular, more accessible, and more
This leads to the greatest challenge facing IT a single problem. And even worse, all of these consumable. In contrast, data warehouse, BI,
organizations across industries: how to provide approaches are doomed to obsolescence from and portal solutions, in addition to contact
users the data they need when they need it, the start. As requirements change, the fixed tracking (CRM), supply-chain management
visualized in a way that is understandable schemas and the complex ETL processes (SCM), employee management (HR), and
and useful. Or put more simply: get the right inherent to data warehouses must be recreated all-in-one enterprise bundles (ERP), all fall into
data, for the right people, at the right time. from scratch. The canned queries and views the category of data containment. While these
Traditionally, this is much easier said than done, that define BI- and portal-based approaches applications (commonly known as data silos)
as the data lives in multiple databases, exists in must be constantly re-evaluated. And the excel in capturing extremely structured data,
various formats, and no user interface exists to limited search and query capabilities of a they make it almost impossible to get the data
present the information in a way that is helpful document management system mean that new out to be re-used by other users and in other
to the user. requirements demand a new installation. applications.
Typically, the approach to solving In short, traditional approaches all suffer
these problems involves some sort of data from the dreaded Shampoo Syndrome: the only
warehouse. Atop the warehouse, we’d probably workable long-term solution is to constantly With data collaboration, the
deploy a business intelligence (BI) solution to lather, rinse, and repeat. And when we do, we
data is much more granular,
surface the answers to common queries to the just create another mountain of data, another
people who need them. place where what we really need can hide.
more accessible, and more
Another tactic might be to install a document consumable.
management system that stores documents The solution is to find data by its meaning
in a central repository, where employees can rather than its location
use search and basic metadata to better locate The key to eliminating many of the inefficiencies Document management systems, on the
individual pieces of information. of today’s information technology solutions is to other hand, attempt to make information more
Or we might build a portal to allow people to access data by its meaning—what it is—rather shareable, but essentially end up creating many
view the right data from multiple silos in a timely than its location—where it is. With meaning, mini-silos in the form of Word documents, PDFs,
fashion. By defining a collection of portlets as we can quickly find what we need simply by Excel spreadsheets, or Web pages. This is
views into specific sources of data, we can describing what it is. This enables information the world of document collaboration, in which
provide a one-stop location for people to view to be shared and consumed at the data level, a information is readily shared, but the data we
information from business-critical data sources. paradigm known as data collaboration. need is locked within the min-silo.



Data collaboration is the best of both worlds. standards. In short, the Anzo products allow right data, the Anzo Data Collaboration Server
By combining the ease of access to information businesses to layer a semantic fabric over can connect to data sources including LDAP
that is the hallmark of document collaboration existing data that: directories, HTTP-accessible Linked Data, and
with the highly structured natured of data from 1. Virtualizes the data so that it is accessible by standard relational databases.
data containment solutions, we can begin to its description regardless of location. But perhaps one of the most useful
answer the IT challenge. The key to success 2. Lets users create their own views of data. connectors is Cambridge Semantics’ Anzo
is to ensure that the meaning of every data 3. Fills in the views by traversing the fabric and for Excel. With Anzo for Excel, data inside
element is surfaced so that it can be easily picking out the relevant information. spreadsheets with arbitrary layouts can be
accessed by any person or application that 4. Keeps everything in synch by allowing linked into the Anzo Data Collaboration Server.
needs it. updates that occur anywhere to update By breaking down the walls of spreadsheet
information everywhere. mini-silos, Anzo for Excel weaves information
Data Collaboration and the Semantic Web from thousands (or more) spreadsheets
It’s no coincidence that the technology scattered across a business, dramatically
standards developed over the past ten years increasing the availability of the right data.
in support of Tim Berners-Lee’s vision of a Context. It’s not enough simply
Semantic Web are the key elements for building to have the right data. People …For The Right People
data collaboration solutions. For as with data Getting the data in front of the right people relies
collaboration, the Semantic Web relies on
must have access to views of on three things: context, security, and “reach”.
explicitly capturing the meaning of data. As the data that depict exactly Context. It’s not enough simply to have the
such, the core Semantic Web standards pave what they need to see, whether right data. People must have access to views
the way for: it be an executive dashboard, of the data that depict exactly what they need
• Flexible, define-as-it-arrives, data structures to see, whether it be an executive dashboard,
• Explicit relationships that travel with the data
a regional summary map, a regional summary map, or a customer-
• Data that is accessed by its definition rather or a customer-by-customer by-customer detailed report. Cambridge
than its address detailed report. Semantics’ visualization product, Anzo on
• Distributed query the Web, allows the same information to be
rendered in many different ways via semantic
As with all standards, Semantic Web lenses. Lenses provide context-appropriate user
technologies lay the groundwork that makes The Right Data… interfaces to render a particular type of data,
improvement possible. It is up to application At the heart of the Anzo suite of products is the meaning that the right people see the right data
developers to build solutions that make the Anzo Data Collaboration Server. This acts as in the right way.
standards practical. a central gateway that provides a consistent Security. In many ways, security is the
interface for applications to read, write, and converse of context. While context ensures
Practical Data Collaboration to Solve query RDF data, regardless of the actual source that the right data surfaces properly to the right
IT’s Challenge of the data. While RDF provides the flexibility people, robust security makes sure data does
Cambridge Semantics is one of the first to incorporate new data as it is virtualized, it’s not surface to the wrong people. The Anzo
companies to develop practical business- all for naught without the proper adaptors for Data Collaboration Server provides security by
solution enablers based on Semantic Web existing data sources. To facilitate access to the layering a role-based access control model atop
the semantic fabric. All data access is gated
through this security model, which defers to the
permissions schemes of legacy data sources
where appropriate. The result is that only the
right people can ever see (or change) the right
data.
Reach. The right data needs to be able to
be brought to the right person, whether that
person is a technical staff member, a line-of-
business manager, a “power user,” or a senior
executive. As such, the software must be within
reach of all users, without the need to call on
IT. Research analysts must be able to collect
and share spreadsheet data themselves. Anzo
for Excel reaches these users by allowing
spreadsheets to be visually linked with just a
few clicks. Supply-chain managers must be able
to drill through data on warehouses, suppliers,
and distributors on their own terms. Anzo on
the Web reaches these users via a simple and
customizable faceted browsing paradigm,
whereby anyone can add their own filters, add
their own lenses, query their data however they
like, and save the results to re-run later or share
with colleagues.



…At The Right Time Data Collaboration in the Days to Come Lee Feigenbaum is VP of Technology and
Finally, it’s not enough to just bring the right Imagine a world in which this challenge has Standards and Cambridge Semantics and co-
data to the right people. It also needs to be been solved. End users—whether knowledge chairs the W3C SPARQL Working Group.
done in a timely fashion. workers, line of business managers, or
First, data access against existing data executives—can simply draw a picture of what Mike Cataldo is currently CEO of Cambridge
sources is accomplished via federated they want to see and then choose the data that Semantics and a veteran of multiple technology
(distributed) query. SPARQL is explicitly should fill in the picture. Within minutes rather start-up companies.
designed to enable queries that access multiple than months the right data shows up on the
data sources at once, and the Anzo Data right people’s screens. Now imagine that the
Collaboration Server includes a SPARQL engine data is live as well: you make a correction to the
that does exactly that. By querying the source data and your changes are reflected in real-time
data directly, Anzo eliminates the cycle time in whatever legacy database or application
typically associated with a data warehouse’s the data comes from. You’ve managed to
ETL processes. maintain a single source of truth for your key
Second, data updates performed via the information assets, while still preserving existing
Anzo Server are broadcast out in real-time to investments in legacy systems and applications.
anywhere the data resides. This means that What sounds miraculous is possible today,
if a value is changed in a spreadsheet cell, in software such as Cambridge Semantics’
the value instantly updates anywhere else Anzo. By combining the revolutionary enabling
it appears, including Web pages or within a capabilities of Semantic Web standards with
relational database. This is essential as many solid, practical engineering, we open the door
spreadsheets, Web pages, and databases will on a completely new paradigm for enterprise
share the same piece of data with confidence software: data collaboration.
as semantic tools are made available to users
across the business enterprise.


nodAliTieS mAgAzine civic SemAnTic Web

Building a Civic Semantic Web
By Joshua Tauberer of GovTracks.us

Technology is a new key player population demographics, etc. We establish sparql?query=DESCRIBE+%3Chttp://www.
in government accountability relations like sponsorship, represents, voted, rdfabout.com/rdf/usgov/congress/111/bills/
and transparency. It’s our own and population across entities of many types. A h1%3E) using URL rewriting in Apache (for a
defense against the threat web lets us ask new questions, and from there robust solution, see my explanation at the end
of government information transforming their answers into visualizations. of http://rdfabout.com/demo/census/). For more
overload. Take the U.S. And because the Semantic Web is a generic about GovTrack’s RDF data, see http://www.
Congress: More than 10,000 platform for all data, I actually think it has govtrack.us/developers/rdf.xpd.
bills are on the table for discussion at any the potential to radically and fundamentally When data gets big, it’s hard to remember
given time, and Members of Congress are transform the way we learn, share information, the exact relations between the entities
taking campaign contributions from thousands and live—but that’s still a bit far off. represented in the data set, so I start to think of
of sources. How can a representative be So for the purposes of my tinkering with the my area of the Semantic Web as several clouds.
accountable if his legislative actions are Semantic Web, GovTrack creates an RDF dump One cloud is the data I generate from GovTrack.
too numerous to track? How can financial of its database (13 million triples) covering Another cloud is data I separately generate
disclosure root out conflicts of interest if
the interesting ones are buried deep within
piles and piles of records? The thread to
transparency isn’t shear volume, however.
It’s the complex network of relationships that
makes up the U.S. Congress, and that makes it
an interesting case for applying Semantic Web
technology.
What the Semantic Web addresses is
data isolation, and this is a problem for
understanding Congress. For instance,
the website MAPLight.org, which looks for
correlations between campaign contributions
to Members of Congress and how they voted Figure 1
on legislation, is essentially something that is
too expensive to do for its own sake. Campaign bills, politicians, votes and more using a mix about campaign contributions from data
data from the Federal Election Commission of existing schemas and some new ones that I files from the government’s Federal Election
isn’t tied to roll call vote data from the House created. I chose URIs for entities in the Linked Commission (FEC): 10 million triples. This cloud
and Senate. It’s only because separate projects Open Data tradition, HTTP-dereferencable relates politicians to election campaigns and
have, for independent reasons, massaged the URIs that resolve to self-describing RDF/ elections, campaign donors with zipcodes, and
existing data and made it more easily meshable XML about the entity. Two good examples are contribution amounts. A third data set is based
that MAPLight is possible. The Semantic Web <http://www.rdfabout.com/rdf/usgov/congress/ on the 2000 U.S. Census, 1 billion triples. The
makes this process cheaper by addressing people/M000303> for Senator John McCain census data has population demographics
meshability at the core. The more government and <http://www.rdfabout.com/rdf/usgov/ for many geographic levels, including states,
data that is mashable, the easier it is to congress/111/bills/h1> for H.R. 1, the economic congressional districts, and postal zipcodes
investigate connections across independent recovery bill passed earlier this year. The HTML (actually “ZCTA”s but we can put that aside).
data sets, research the dynamics of the system, pages on GovTrack itself tie in to the RDF world (For more, see http://rdfabout.com. Through the
or teach others how Congress works. through <link> tags: bill pages include the URI Census cloud the data is linked to Geonames
Innovating the public’s engagement with I coined for the bill, for instance. and the rest of the the Linked Open Data
Congress by applying technology has been the I also have a sometimes-working- community.)
motivation behind my site www.GovTrack.us, a sometimes-not SPARQL endpoint set up, I’ve related the clouds together so we
free congress-tracking tool that I built and have SPARQL being the de facto query language can take interesting slices through them. The
been running since 2004. GovTrack amasses for RDF. SPARQL lets us ask questions of GovTrack data connects to the FEC data
a large XML database of congressional the data, such as how did politicians vote on through politicians. The Census data connects
information, including the status of legislation, bills (see example 1). The SPARQL endpoint to the GovTrack data through states and
voting records, and other bits, by screen runs off of a “triple store”, the equivalent of congressional districts (the regions represented
scraping official government websites that have a relational database for the semantic web, by senators and representatives) and to the
the data online already but in a less useful form. which is underlyingly a MySQL database with FEC data through zipcodes. That means we
If “metadata” is tabular, isolated, and about a table whose columns are “subject, predicate, ask questions that go beyond one data set
web resources, the Semantic Web goes far object”, i.e. a table of triples. (It uses my own such as: what are the census statistics of the
beyond that. It helps us encode non-tabular, C#/.NET RDF library: http://razor.occams. districts represented by congressmen, are
non-hierarchical data. It lets us make a web of info/code/semweb.) The RDF/XML returned votes correlated with campaign contributions
knowledge about the real world, connecting by dereferencing the URIs is actually auto- aggregated by zipcode, are campaign
entities like bills in Congress with Members of generated by redirecting the user to a SPARQL contributions by zipcode correlated with
Congress, what districts they represent, their DESCRIBE query (i.e. http://www.rdfabout.com/ census statistics for the zipcode, etc.? Once


nodAliTieS mAgAzine civic SemAnTic Web

the Semantic Web framework is in place, the Example 1 Example 2
marginal cost of asking a new question is much Get a table of how senators voted on all of the Get total campaign contributions to Rep. Steve
lower. We don’t need to go through heavy work Senate bills in 2009-2010: Israel by zipcode.
of meshing two data sets for each new question
once the data is already in RDF with connected
URIs.
My dream is to be able to plug in SPARQL PREFIX rdf: <http://www. PREFIX fec: <http://www.rdfabout.com/
queries into visualization websites like Many w3.org/1999/02/22-rdf-syntax-ns#> rdf/schema/usfec/>
Eyes, Swivel, and mapping tools and instantly PREFIX rdfs: <http://www.
get an answer to my question in a compelling SELECT ?zipcode ?value WHERE {
w3.org/2000/01/rdf-schema#>
form. For now, some copy-paste is necessary. ?campaign fec:candidate <http://www.
PREFIX bill: <http://www.rdfabout.com/
Let’s take an example. Did a state’s median rdf...ongress/people/I000057> .
income predict the votes of senators on H.R.
rdf/schema/usbill/>
1, the economic recovery bill? Perhaps the PREFIX vote: <http://www.rdfabout. ?campaign fec:cycle 2008 .
senators from the poorest states, likely the com/rdf/schema/vote/> ?zipcode
most affected by the economic trouble, were SELECT ?bill ?voter ?option WHERE { fec:zipAggregatedContribution [
more likely to want economic stimulus. This
query takes a path through two of my clouds, ?bill a bill:SenateBill . fec:toCampaign ?campaign;
depicted in Figure 1. The SPARQL query ?bill bill:congress “111” ; fec:amount ?value
mimics the picture: each edge corresponds
bill:hadAction [ ].
to a statement in the query. Except the real
query is more complicated (it’s given at http:// a bill:VoteAction ; ?zipcode fec:zcta ?uri .
www.govtrack.us/developers/rdf.xpd). It is bill:vote [ }
complicated not because RDF or SPARQL are
inherently complicated, but because the data vote:hasOption [
model that I chose to represent the information vote:votedBy ?voter ;
is complicated. That is, I made my data set
very detailed and precise, and it takes a
rdfs:label ?option ;
precise query to access it properly. If you run ]
it on the SPARQL form on that page, get the
];
results in CSV format, copy them into Excel,
and run a correlation test, you’d indeed find a ].
moderate correlation between median income }
and vote, but in the direction opposite to what
we expected. (I know why, but I’ll let you think
about it.)
Another interesting case is whether
campaign contributions to congressmen
mostly come from their district, or if they get
contributions from sources far away. The
SPARQL query listed in example 2 extracts the
relevant numbers for Rep. Steve Israel from
New York: for each zipcode, the total amount
of campaign contributions he received from
individuals with addresses in that zipcode in the
last election. Figure 2 puts these values on a
map, with congressional districts overlayed as
well. A form where you can submit a SPARQL
query like these examples and see the results
instantly on a map would be incredible for data
investigation.
So what is government transparency,
practically speaking? It’s more than just
information disclosure. Transparency means
the public can get answers to their burning
questions. The more questions they can
answer from a dataset, the more transparency
it provides. We can have more transparency
without necessarily more disclosure but instead Figure 2
with the ability to apply better tools. Meshing
and querying government datasets with RDF
and SPARQL could be a new way to reach Joshua Tauberer is a software developer and entrepreneur and runs www.GovTrack.us, which tracks
new heights of civic engagement and public what is happening in the U.S. Congress.
oversight.


nodAliTieS mAgAzine WAiving RighTS

Waiving Rights over Linked Data
By Ian Davis, Talis CTO

We love data at Talis and we granting specific people a limited right to make allow people to use data you have collated but
want as much of it to be freely copies without having to ask you first. Licensing your company goes bankrupt and the rights to
reusable as possible. In fact, of one right does not affect your possession the data collection are sold by the liquidators.
because we wanted to see of the others. For example you could grant If you hadn’t licensed your rights explicitly then
even more reusable data we the right to copy your work but retain the right every one of your users could be liable to be
recently launched the Talis to perform it. Creative Commons licenses are sued by the new rights holder!
Connected Commons offering mostly concerned with copyright, but they do This is where waivers of rights can help. By
completely free hosting of public domain data. not usually deal with the other rights such as explictly waiving your rights over your data then
We believe that dedicating data to the public database rights or trade secrets. you are giving your users the best guarantee of
domain is the best way to ensure that data is Waivers, on the other hand, are a voluntary safety that you can. Even if you lost control of
universally reusable and remixable. When data relinquishment of a right. If you waive your the data collection subsequent owners could
is public domain it means that it can be reused exclusive copyright over a work then you are not persue your users because the rights you
automatically without needing to check terms explictly allowing other people to copy it and held have already been waived.
and conditions or track the source of every you will have no claim over their use of it in that There are two waivers of rights that can be
statement to provide attribution. These kinds of way. It gives users of your work huge freedom applied to datasets:
things act as friction to reuse, wasting energy and confidence that they will not be persued for
that could be better spent creating inspiring license fees in the future. • PDDL from Open Data Commons
things. • CC0 from Creative Commons
We also firmly believe that, in the future, The Licensing Problem
there will a significant role for other forms of In general factual data does not convey any Both of these waivers can be used for data
data licensing, including commercial access. copyrights, but it may be subject to other rights intended for submission to the Talis Connected
We will support those efforts too when the time such as trade mark or, in many jurisdictions, Commons.
comes but today the Linked Data web needs database right. Because factual data is not
more and better data that is freely accessible. usually subject to copyright, the standard Community Norms
Creative Commons licenses are not applicable: When you apply a waiver like CC0 you are
Licensing vs Waivers you can’t grant the exclusive right to copy relinquishing all your rights over the work to
If you are not interested in the background to the facts if that right isn’t yours to give. It also the fullest extent possible under the law. That
data licensing then you can jump straight to the means you cannot add conditions such as means that you cannot force people to attribute
section How to Declare Your Waiver. share-alike. you or stop them from making commercial use
You are probably familiar with the process There isn’t a Creative Commons license for of your work.
of licensing a creative work, most likely through every possible right and there probably can’t be The preferred approach is to attach a set
the great job that Creative Commons have been because of the huge variation in rights granted of community norms to the work. These are
doing in recent years. However, the concept of in different jurisdictions around the world. Also, like a code of conduct for use of the work and
waivers is less well known but highly relevant to when we start to look at licensing compilations are usually self-policing. They are not legally
reuse of linked data. of data we find that the situation becomes enforceable but form part of the ethical or
complex because you have to consider both professional requirements for participating
the database and its contents seperately. in a community. The best known example of
For example a document of articles would community norms are the citation standards
Whenever you create be subject to database right over the whole used in the academic commnity. Citing pre-
something you have collection and individual copyrights for each existing work is not legally enforceable but
automatic rights over it article, quite possible to many different owners. those who abuse the norms can find themselves
The Open Data Commons has addressed this excluded from the academic community.
granted to you.
particular example with its Open Database The Open Data Commons has published a
License and Database Contents License set of attribution and share-alike norms which
(based on work originally donated by Talis). If asks that users of the data:
Whenever you create something you have a standard license doesn’t exist then you need • Share work derived from the data.
automatic rights over it granted to you. The to hire lawyers and write one for yourself—a • Give credit to the original data publisher.
best known of these rights is copyright, which potentially huge cost. • Point others at the source of the data.
gives you the exclusive right to make copies of Our collective goal for a successful Linked • Publish in open formats.
your creative work. There are many other rights Data web has to be to protect consumers of • Avoid using digital rights management.
which can be held over intellectual property the data: the people who are remixing many
such as design rights, trade marks, registered different sources of data. Our intentions may How to Declare Your Waiver
designs, performers rights, trade secrets, be very honourable, but people need certainty To delare your waiver in a machine readable
database rights, publication rights and many if they are to build enduring value on data. way, you should first create a voID description of
more. Creative Commons licenses are irrevocable so your dataset. VoID, or Vocabulary of Interlinked
Licensing is the process of granting others even if you lose control over your work through Datasets, is a vocabulary designed to describe
limited use of rights you possess. For example, some misfortune, the people reusing it will be key attributes of your dataset. We created a
when you license your copyright you are protected forever. Imagine this scenario: you waiver RDF vocabulary that can be used with


nodAliTieS mAgAzine WAiving RighTS

voID to declare any waiver of rights and the community norms around a dataset.
In this example we describe a dataset using the void:Dataset class and provide it with a dc:title
as a minimal human readable description. You should add other descriptive properties as necessary
(some suggestions can be found in the voID guide).
We then use the wv:waiver property (defined in the waiver RDF vocabulary) to link the dataset to
the Open Data Commons PDDL waiver. We use the wv:declaration property to include a human-
readable declaration of the waiver. This is purely informational, but can be immediately be used by
a person examining the voID description. Finally we use the wv:norms property to link the dataset to
the community norms we suggest for it, in this case the ODC Attribution and Share-alike norms.

<rdf:RDF xmlns:rdf=”http://www.w3.org/1999/02/22-rdf-syntax-ns#”
xmlns:dc=”http://purl.org/dc/terms/”
xmlns:wv=”http://vocab.org/waiver/terms/”
xmlns:void=”http://rdfs.org/ns/void#”>
<void:Dataset rdf:about=”{{uri of your dataset}}”>
<dc:title>{{name of dataset}}</dc:title>
<wv:waiver rdf:resource=”http://www.opendatacommons.org/odc-public-
domain-dedication-and-licence/”/>
<wv:norms rdf:resource=”http://www.opendatacommons.org/norms/odc-
by-sa/” />
<wv:declaration>
To the extent possible under law, {{your name or organisation}} has
waived all
copyright and related or neighboring rights to {{name of dataset}}
</wv:declaration>
</void:Dataset>
</rdf:RDF>

Alternatively if you were to choose the CC0 waiver without any particular norms then you should
use the following RDF:

<rdf:RDF xmlns:rdf=”http://www.w3.org/1999/02/22-rdf-syntax-ns#”
xmlns:dc=”http://purl.org/dc/terms/”
xmlns:wv=”http://vocab.org/waiver/terms/”
xmlns:void=”http://rdfs.org/ns/void#”>
<void:Dataset rdf:about=”{{uri of your dataset}}”>
<dc:title>{{name of dataset}}</dc:title>
<wv:waiver rdf:resource=”http://creativecommons.org/publicdomain/
zero/1.0/”/>
<wv:declaration>
To the extent possible under law, {{your name or organisation}} has
waived all
copyright and related or neighboring rights to {{name of dataset}}
</wv:declaration>
</void:Dataset>
</rdf:RDF>

These examples show that it is very simple to declare your waiver. However, before you do so be
sure to read carefully what rights you are irrevocably giving up. For example you would most likely
be waiving your publicity and privacy rights, so if your image is included in the dataset you could not
later complain that someone is using it in a way you do not approve of. If you are worried about how
your work will be used, if you want to legally require attribution, or if you don’t want people to make
money off of your work, then you should not use a waiver and instead seek legal advice on the
creation of a data license specific to your needs.



Might Semantic Technologies Permit
Meaningful Brand relationships?
By Paul Miller, Founder of cloudofdata.com

Much has been written about about an otherwise exemplary service, the Brands need to engage in this conversation,
growing Enterprise use of human touches that made us smile, the odd as we are beginning to see them do, but they
social media (usually Twitter, inconsistencies in a polished persona; none also need to discover the means to cost-
these days) to successfully are enough to make us pick up the phone, but effectively monitor and engage with a potential
track and mitigate customer we comment upon them endlessly in Twitter, flood of third party reaction whilst using the
complaint. Many have Facebook, FriendFeed and elsewhere, and by Business Intelligence tools available to them
been quick to spot that the tapping into this fundamentally honest stream in nimbly shaping public opinion to their
disproportionately high cost of satisfying (or, of consciousness there is much for those about advantage wherever possible.
more cynically, silencing) these early adopters whom we comment to learn. Good companies I spoke with Scott Brinker last year (http://bit.
is unlikely to scale effectively as an increasingly probably already know about fundamental ly/wlkwp), to explore his—then nascent—views
large cohort of customers move onto these failings in a product long before their customer on Semantic Marketing, and look forward to
services, and it must remain an open question support operation melts down under the weight hearing his latest thoughts at the Semantic
as to whether ComcastCares and its peers of complaints or their quarterly sales targets Technology Conference in San Jose in June.
can survive any move to the mainstream in are seriously under-achieved. Do they have as More recently, Eric Hillerbrand (http://bit.
recognisable form. good a handle on the things we love? Do they ly/XyAA9) talked about some of his ideas with
have a clue about the minor gripes of customers respect to ‘Social Commerce,’ and the ways
outside their pre-launch polling groups? Do in which commercial organisations might seek
they know about the gut reaction to a colour, to strengthen and exploit relationships with
a touch, a smell, or a careless word that their customers, aided by a range of semantic
Collectively, we’ve moved from persuaded a likely prospect to buy a technically technologies.
simply complaining about the or aesthetically inferior product from the
worst failures of companies, competition instead? All this and more is there
their products and their for the taking in the stream of online chatter
employees, toward emitting freely directed their way. We’re just beginning to
an impressive stream of FYIs.
Semantic Technologies aren’t often directly grasp the realities of a world
associated with the worlds of marketing
and commerce, yet individuals such as Eric
in which tightly controlled
Hillerbrand and Scott Brinker are hard at work and fiercely guarded brand
to show just what might be possible when the attributes become increasingly
It appears, though, that Enterprise experiences of the Semantic Web are applied to permeable.
engagement in the social sphere changes this space. Brands are no longer owned by the
the game far more significantly than merely companies in whose name they were created.
enabling a select few twitterati to jump the Increasingly, ownership of various forms is
Customer Support queue, and that this change being asserted by the multitude of stakeholders We’re just beginning to grasp the realities of
is worth effort and investment in order to ensure with effort and attention invested in the brand. a world in which tightly controlled and fiercely
that it does scale. What’s actually happening They care about it, they care about what it says guarded brand attributes become increasingly
is that a relationship is being enabled between about them, and they play a clear role in the permeable. For those companies with the
a brand and those that Seth Godin might brand’s evolution whether its managers want confidence and foresight to loosen their grip,
recognise as its tribe; a relationship in which them to or not. whilst simultaneously exploiting the wealth of
interactions are no longer driven predominantly data and new opportunities to engage, there is
by the desire to seek redress. Rather than only much to be gained. For the dinosaurs that hang
raising those issues serious enough for us Brands are no longer owned on to ‘their’ brand in spite of the world around
to have written letters or endured telephone them, there is everything to lose.
muzak in the past, we now comment on issues
by the companies in whose
at the periphery of a brand. Collectively, we’ve name they were created. Paul Miller is the founder of The Cloud of
moved from simply complaining about the Increasingly, ownership Data blog.
worst failures of companies, their products and of various forms is being
their employees, toward emitting an impressive
stream of FYIs. Individually insignificant, and
asserted by the multitude of
possibly unimportant, together these light stakeholders with effort and
touches on and around a brand build into an attention invested
ever-changing and valuable commentary that in the brand.
brands and the corporations they front would
do well to take notice of. The minor niggles


Liberate your data and let
a thousand ideas bloom.

Talis provides a Software as a Service (SaaS) platform for creating applications that use Semantic
Web technologies. By reducing the complexity of storing, indexing, searching and augmenting
linked data, our platform lets your developers spend less time on infrastructure and more time
building extraordinary applications. Find out how at talis.com/platform.

shared innovation
TM

nodAliTieS mAgAzine RdfA And linked dATA

RDFa and Linked Data in UK Government
Websites
By Mark Birbeck, of Backplane Ltd

The UK government’s Central these consultations are on departmental sites, development work, retraining of users, porting
Office of Information had a rather than being available on a central site; data from older systems, and so on.
straightforward problem to from the Department of Energy and Climate The second ‘traditional’ solution at least has
solve: how could they create Change (DECC) seeking feedback on clean the merit of keeping existing systems intact, but
a centralised website of coal, to the Ministry of Justice (MOJ) providing would have required additional interfaces to be
information that the public an opportunity for people to comment on created to move the data from the departmental
could search and access, prisoners’ voting rights, each department servers to the centre; each department would
when the source of that information could be manages its own publication of consultations. have had to create an interface between their
any government department database or any own system and the central one.
public sector website? Traditional solutions Just getting one department into a situation
For example, different organisations, such as Traditional answers to these problems would where they could centralise their information
Her Majesty’s Revenue and Customs (HMRC) or have been to either (a) impose on each of the would have been a major undertaking—not
the National Health Service (NHS) would each departments that they should key their data only were there lots of departments to consider,
post job vacancies to their own websites, but directly into a new central database (which but each department was using a different
technology to publish their vacancies or
consultations to the web. For example, some
departments with only a small number of job
vacancies would likely use static HTML pages.
Other departments, perhaps with larger IT
departments, might use ASP .NET or a Java-
based system.

Enter RDFa
The RDFa answer to this set of problems is
simple—both conceptually, and to implement.
RDFa allows HTML publishers to embed
RDF into their pages, so using the HTTP and
HTML infrastructure to publish their information.
This simple method of publishing data in turn
means that any system can import this data, just
by obtaining (or creating) an RDFa parser.
In short, each department can keep their
own data management system, and simply
add code to their existing web-page publishing
step to augment the HTML with the data as
RDFa. The central system in turn only needs
one import mechanism—something that
understands RDFa.
Adding this facility to an individual
departments publishing system proved to be
very quick and straightforward. But it’s not just
there was no central site that the public could would in turn drive the central website), or (b) UK government departments that are finding it
go to, to find all public sector vacancies. This create complex communication pipelines that straightforward to add RDFa to their pages. It
would be a problem at any time, but in the midst would allow the decentralised databases to was interesting to hear at SemTech in June that
of attempts by the government to help people communicate with the central system. Google’s rich snippet launch partners (such
through the recession, it’s crucial to ensure that And either of these solutions would almost as Yelp), were able to add RDFa support in
the public knows what vacancies are available. certainly have turned out to have been a non- “roughly a day”.
It might not occur to someone looking for a job starter.
as a plumber or an electrician they they should The first solution was unlikely to ever get RDF publishing techniques
visit the NHS or Army websites, so a centralised off the ground, because it would have required Adding data to web-pages might seem quite an
site could make a big difference. each department to replace their existing obvious technique, but there are two important
Similarly, as in most modern democracies, technology with something new. Even if there things to note here.
government departments are constantly seeking was agreement on what that technology should First, the COI has to be commended for
feedback from the public and interested parties, be—and that in itself could take an age to having the vision to publish RDF at all. Of
about specific issues. But as with job vacancies, resolve—there would have been a need for new course, now that Gordon Brown has asked


nodAliTieS mAgAzine RdfA And linked dATA

RDFa turns this on its head, and says that As more servers both consume RDFa from
the HTML page is the data. One and the same one set of servers, and publish RDFa again to
page can be read as an HTML page, or as a variety of other servers, we enter the exciting
an RDF page, which in turn means that the world of Linked Data, and it’s ‘turtles all the way
changes required to the existing publication down’.
system are minimal. The COI once again
showed its far-sightedness by adopting this Conclusion
technique. By using RDFa to address the challenge of
making distributed data available in one place,
Turtles all the way down the COI avoided having to make changes to
But the benefits of RDFa don’t just stop there. each department’s systems. But once each
for Sir Tim Berners-Lee’s help in making Firstly, because the data is being published department is publishing RDFa, it becomes
government data publicly available, it seems via HTTP and HTML, it’s possible for anyone possible for third parties to consume that
pretty obvious—indeed it may even become to read the same data, not just the centralised information however they see fit. Such a flexible
fashionable! But the COI were planning this website that was being planned. This means architecture is crucial in the age of open
project at least a year ago, and at that time RDF that third party job vacancy sites, for example, government, and is a cornerstone of linked
was by no means a done deal (and you could could import vacancies from relevant open data.
say it’s still not). departments, to add to their databases. In fact,
But the second important thing is that one of the main drivers for the consultations Mark is managing director of Backplane Ltd.
even after deciding to publish RDF, it’s still not project was to try to help improve the accuracy (http://webBackplane.com/), a London-based
immediately obvious that the solution should of an already existing website (set up by a company involved in a number of RDFa/linked
involve RDFa, especially not a year ago. member of the public) that used ‘screen- data projects for UK government departments. He
The usual means of publishing RDF is to scraping’ to try to keep up with the available is the original proposer of RDFa.
provide a distinct source of data in the form consultations—RDFa provides much more
of RDF/XML (and perhaps other formats, too, accurate information.
such as N3). If there is an HTML version it In addition, the centralised website will not
usually exists for the purpose of describing the only import RDFa but publish it too. This means
data itself. In other words, the RDF/XML format that third-party servers are also able to import
is primary, which means that anyone who is some or all of the centralised data, into their
publishing HTML pages but wants to publish own sites.
RDF as well, will need to add an extra piece of And thirdly, by using RDFa the sites could
infrastructure that exists alongside their web- provide information to search applications such
pages. as SearchMonkey.


nodAliTieS mAgAzine beRneRS-lee And RWW

Search Engines, User Interfaces for Data,
Wolfram Alpha, and More...
Interview of Sir Tim Berners-Lee by Richard MacManus of ReadWriteWeb. Reprinted with kind permission.

During my recent trip to a direct inspiration for me when I launched How Linked Data Relates to The Semantic
Boston, I had the opportunity ReadWriteWeb back in 2003.1 Web
to visit MIT. At the end of a long After shaking hands, I told Tim Berners-Lee RWW: Earlier this year you gave an inspiring
day of meetings with various that this blog’s name was in part inspired by talk at TED about Linked Data. You described
MIT tech masterminds, I made the first browser, which he developed, called Linked Data as a sea change akin to the
my way to the funny shaped “WorldWideWeb”. That was a read/write invention of the WWW itself - i.e. we’ve gone
building where the World Wide browser; meaning you could not only browse from a Web of documents to a Web of data. Can
Web Consortium (W3C) and its director Tim and read content, but create and edit content you please explain though how Linked Data
Berners-Lee work. Berners-Lee is of course too. It was a shame then when Mosaic, a read- relates to the Semantic Web, is it a subset of it?
the man who invented the World Wide Web 20 only browser, became the first mainstream web TBL: They fit in completely, in that the linked
years ago. browser in the mid-90s. It wasn’t until the rise of data actually uses a small slice of all the various
This was my first meeting with the Web’s Web 2.0 that the read/write philosophy gained technologies that people have put together and
creator, whose work and philosophy was widespread acceptance.2 On that note, we standardized for the Semantic Web.
launched into the interview...

1. The very first sentence written on this blog, on 20 April, 2003, was: “The World Wide Web in 2003 is beginning to fulfill the hopes that Tim Berners-Lee had
for it over 10 years ago when he created it.”
2. For more on read/write browsers, you can read another early RWW post entitled ‘What became of the Browser/Editor’.



We started off with the Semantic Web Linked Data and Governments [their data] into Linked Data. One of the nice
roadmap, which had lots of languages that we RWW: In a recent Design Issues note, you things about Linked Data, when they have a pile
wanted to create. [However] the community urge governments to put their data online as of it, is that they could run a SPARQL server on
as a whole got a bit distracted from the idea Linked Data (although you’d also be happy for it. SPARQL servers are a commodity product,
that actually the most important piece is the governments to just make available the raw data a solution for all of the people who say ‘but
interoperability of the data. The fact that things - presumably so that others can then structure actually I wanted to have XML.’ A SPARQL
are identified with URIs is the key thing. it). What do you realistically expect, for example, server will generate an XML file [and] allow
The Semantic Web and Linked Data connect the U.S. or U.K. governments to do over the somebody to write out, effectively, a URL for the
because when we’ve got this web of linked next year? And in the near future, do you XML file.
data, there are already lots of technologies foresee different governments interconnecting In fact, I don’t see why SPARQL servers
which exist to do fancy things with it. But it’s their Linked Data sets? shouldn’t provide CSV files, something which
time now to concentrate on getting the web of TBL: One can’t generalize, governments as far as I know isn’t in the standards. But I’d
linked data out there.

How Linked Data Has Evolved via
Grassroots
RWW: Linked Data has had a lot of grassroots
support, which you mentioned in your TED
speech. This is something Semantic Web
technologies, such as RDF, have struggled
to get over the years. Has the W3C been
pushing the more bottom-up Linked Data world,
because of the frustration over lack of take-up of
top-down Semantic Web?
TBL: A lot of the initial RDF and OWL
projects came out of the academic world; and
some of them were projects to show what you
could do in a closed world. And the files were
zipped up and left on a disc. While they were
interesting projects, and while the systems were
useful systems, the semantic web community
maybe missed the point of the ‘web’ bit and
focused too much on the ‘semantic’. However
the work that’s been done in the Semantic Web,
the standards, was really valuable. It’s relatively are (like most big organizations) fascinatingly recommend it, certainly in government context,
recently for example that SPARQL [an RDF diverse inside them. So you’ll find that there because CSV files are what people have and
query language] has been developed. are places inside governments where you get a what people want.
Somebody drew an analogy the other day: champion who gets linked data and who’s just So the message [for government] is to use
can you imagine trying to promote a world written a script and produced some linked data. RDF. Linked Data is the backplane, it’s the thing
of databases without SQL? Even though it’s So in the UK government for example, you’ll find that you connect to in both directions. As a
not an interoperable protocol, it’s just a query there’s RDFa [in the code of its website] for civil [web] producer your job is to make sure that
language. So similarly, all that’s been put into service jobs. So if somebody wants to make a you produce Linked Data one way or another.
RDF, rdfs and OWL is very valuable to the linked database of all the jobs, they can do that very And as a consumer, there are lots of ways to
data community. easily. consume that data once it’s out there as Linked
The Linked Data community tend to use a There are other cases where the easiest Data.
subset of that [Semantic Web technologies], thing for somebody to do is to just put data
of OWL for example. But they certainly use up in whatever form it’s available. Comma Semantic Web and Search Engines Like
SPARQL. So you could argue that really it separated values (CSV) files are remarkably Google, Yahoo
wasn’t ready to be deployed widely. popular. They’re exported sometimes from RWW: You’ve been talking about the Semantic
Linked Data started as a very informal spreadsheets. It’s remarkable how much Web for many years now. Generally the view is
Design Issues note that I put in; it was a information is in spreadsheets. Or sometimes that Semantic Web is great in theory, but we’re
grassroots movement from very early on. So pulled out of a database and then put up on still not seeing a large number of commercial
yes W3C has been emphasizing the importance the web. It’s not as good, not as useful to the web apps that use RDF (we’ve seen a number
of Linked Data. It’s been the Semantic Web community, as if Linked Data had been put up of scientific or academic ones). However we
Interest Group of course, and various [other there and linked. But the first step of actually have begun to see some traction with RDFa
Semantic Web] activities, which has been putting the data out there is the one that nobody (embedding RDF metadata into XHTML Web
pushing it. But also Linked Data has been else can do. content), for example Google’s Rich Snippets
seized on - a group of people for example put Data.gov, a catalog of public data, was and Yahoo’s SearchMonkey. Has the takeup of
together DBpedia.3 That wasn’t commissioned, launched in May by the U.S. government RDFa taken you by surprise?
that was that they just thought it would be a The way to go is for government TBL: Not really, but the takeup by the search
really cool idea. departments to go the extra step and convert engines is interesting. In a way I was happy to

3. DBpedia is a community project to extract structured information from Wikipedia; see ReadWriteWeb’s profile of this and similar resources.



see that, it was a milestone for those things to different thing to be able to do. It really doesn’t interfaces or designs for semantic applications
come out of the search engines. The search compare to a search engine. in recent months - if so which ones and why did
engines had typically not been keen on the You’ve got search for text phrases on one they impress you?
Semantic Web - maybe you could argue that side (which is a useful tool) and querying of the TBL: I think that whole area is very exciting
their business is making order out of chaos, data on the other. I think that those things will at the moment. The only piece of hacking I’ve
and they’re actually happy with the chaos. And connect together a lot. done over the past few years has been on a
if you provide them with the order, they don’t So I think people will search using a search thing called the Tabulator [a data browser and
immediately see the use of it. text engine, and find a webpage. On the front editor], which is addressing exactly that. Partly
because I wanted to be able to look at this
data. And now there are lots of different ways
that people need to be able to look at data. You
need to be able to browse through it piece by
piece, exploring the world of data. You need to
be able to look for patterns of particular things
that have happened. Because this is data, we
need to be able to use all of the power that
traditionally we’ve used for data. When I’ve
pulled in my chosen data set, using a query,
I want to be able to do [things like] maps,
graphs, analysis, and statistical stuff.
So when you talk about user interfaces for
this, it’s really very very broad. Yes I think it’s
Also I think there was misunderstanding in of the webpage they’ll find a link to some data, important. There’s also the distinction we can
the search engine industry that the Semantic then they’ll browse with a data browser, then make between the generic interfaces and the
Web meant metadata, and metadata meant they’ll find a pattern which is really interesting, specific interfaces.
keywords, and keywords don’t work because then they’ll make their data system go and find There will always be specific interfaces;
people lie. Because traditionally in information all the things which are like that pattern (which for example if you’re looking at calendar
retrieval systems, keywords haven’t proven is actually doing a query, but they’ll not realize data, there’s nothing else like a calendar that
up to the task of finding stuff on the Web. One it), then they’ll be in data mode with tables and understands weeks, months and years. If
of the reasons is that people lie, the other is doing statistical analysis, and in that statistical you’re looking at a genome, it’s good to have a
that they can’t be bothered to enter keywords. analysis they’ll find an interesting object which genetics-specific user interface.
So keywords have gotten a bad reputation, has a home page, and they’ll click on that, and However you also need to be able to
then metadata in general was tarred with this go to a homepage and be back on the Web connect that data, through generic interfaces.
‘keywords don’t work’ brush. Because a lot of again. So if my genome data was taken during an
Semantic Web data included metadata, then
people thought that with Semantic Web data
-- again, that people will lie and won’t have the
time to produce it.
Now I think there’s a realization that when
you’re putting data online, that people are
motivated NOT to lie. For example when
your band is going to produce its next
album, or when your band is going to play
next downtown, you’re motivated to put that
information up there on the Semantic Web.
There’s an awful lot of cases when actually data
is really important to people; and it’s on the web
anyway. So I think it’s great that some of the
search engine companies are starting to read
RDFa.
Does this mean that they [search engines]
will start to absorb the whole RDF data model?
If they do, then they will be able to start pulling
all of the linked data cloud in.
Will they know what to do with it? Because So the web of linked data and the web of experiment which happened over a particular
when it’s data in a very organized form, I think documents actually connect in both directions, period, I need to be able to look at that in the
some people have been misunderstanding the with links. calendar - so I can connect the genetics to the
Semantic Web as being something that tries calendar.
to make a better search engine - i.e. when you User Interfaces for Semantic Content So one of the things I hope to see is domain-
type something into a little box. But of course RWW: At the recent SemTech conference, specific things for various different domains,
the great thing about the Semantic Web is that Tom Tague of Thomson Reuters’ Calais project and the generic user interfaces. And hopefully
you can query it, you can ask a complicated suggested that user interfaces for semantic the generic interfaces will be able to tie together
query of the Semantic Web, like a SQL query content are key in getting more take-up. With all of the domains.
(we call it a SPARQL query), and that’s such a that in mind, I wonder if you’ve seen some great



addition. I imagine that quite a lot of it may have
come from the web of Linked Data.

e-Commerce and Linked Data
RWW: There have been reports recently that
both Google and Yahoo will be supporting the
Good Relations ontology and linked data for
e-commerce. Companies such as Best Buy are
already putting out product information in RDFa.
What would be your advice to e-commerce
vendors right now, to help them transition to this
world of structured data on the Web. The same
question could be asked across many verticals,
but e-commerce seems like one area which has
some momentum right now. Would you advise
them just to put out their data as Linked Data?
TBL: Yup! Certainly this year is the year
to do it. I’ve been advising governments to
do it and when you look at an enterprise, you
find that a lot of the issues are the same. But
when you put your data from government
or enterprise out there, make sure you don’t
disturb existing ecosystems. Don’t threaten
those systems, because you’ve spent years
building them up.
Maybe there’s an analogy with when the
Web first started and the first bookshops went
online. They were more or less a flyer, saying
‘hey we have a great bookshop at 23 Main
St, come on down!’. Let’s say that a person
named Joe owned one of these early online
bookshops. If somebody had suggested to
Joe that he should put his catalog online, Joe
would’ve felt that that was very proprietary data.
And he’d be worried that other bookshops
would see where he was weak, so they’d be
Wolfram Alpha and Natural Language yes, those interfaces will become important. able to advertise themselves as filling that niche
Interfaces Conversational interfaces have always he’s weak in.
RWW: An interesting new product was launched been a really interesting avenue. We’ve had But when his competitors Fred and Albert
this year called Wolfram|Alpha, described as a voice browser work in W3C, that has been an put their catalogs online, then Joe can check
‘computational knowledge engine.’ It’s kind of interesting alternative avenue. It’s possible which books people are browsing at Fred and
a mix between Google (search) and Wikipedia that as compute power goes up, we’ll see a Albert’s websites. So Joe would [finally] be
(knowledge), and its key attribute is that enables prolifieration of machines capable of doing pursuaded to put his book catalog up online.
you to compute something. The founders think voice. It’ll move from the mainframe to being But he doesn’t put up the prices... until Albert
that ‘computing’ things on the fly is something able to run on a laptop or your phone. As that and/or Fred does. And even if catalog and
we’re going to see a lot of in future. What’s your happens, we’ll get actual voice recognition pricing is up there, nobody puts their stock
take on Wolfram|Alpha? and pattern natural language at the front end. levels online. And there was a period of time
TBL: There are two parts to that sort of That will perhaps be an important part of the when nobody [i.e. online booksellers] had their
technology. One of them is a sort of stilted Semantic Web. stock levels up. But people got fed up with
natural language interface. We’ve seen those We talked before about what a great ordering stuff that wasn’t in stock. So the first
sort of natural language queries for years. challenge the Semantic Web is going to be from book shop to actually tell you about stock levels
Boris Katz [from W3C] created a system a user interface point of view. Conversational suddenly was then unbelievably attractive to its
called START [a software system designed to interfaces are going to be part of [solving] that. customers.
answer questions that are posed to it in natural Of course it’s also going to be really valuable to So there’s this syndrome of progressive
language]. I think with the Semantic Web out have compositional interfaces - for the visually competitive disclosure. This happens when
there, those sorts of interfaces are going to impaired and so on. people realize that if you’re going to do
become important, very valuable, because Wolfram|Alpha is also a large curated business with somebody, if you’re going to have
people will be able to ask more complicated database of data sets. Obviously I’m interested your partners up and down the supply chain,
things. The search engine has traditionally been in the big data set which is out there, which is really it’s useful to check the data web - and life
limited to just a phrase, but some of the search Linked Data. This everybody can connect to. goes much more quickly and open.
engines are now starting to realize that if they I don’t really know a lot about the internals of Best Buy may be what starts the ball rolling
put data behind them and have computation Wolfram|Alpha’s data set. I don’t know whether [among e-commerce vendors]. Now if I want
engines, then you can ask things like ‘what’s they’re likely to put any of it out on the web to look out for what [products are] available,
this many pounds in dollars?’ and so on. So as Linked Data - that might be an interesting



Nodalities Magazine
SUBSCRIBE NOW

I can write a program to see what there is. If identifier of the thing into data about the thing, is
somebody wants to compete with Best Buy, to yet to be seen. But it’s a very exciting idea.
my program they’ll be invisible unless they can Similarly, I’d like to be able to scan a
get their data up in RDF. Doesn’t matter whether barcode and get back nutritional information
they use RDFa or RDF XML, as long as it maps about what’s in - for example - a can of food.
in a standard fashion to the RDF model, then But we don’t have that yet. To get that sort of
they will be visible. thing, which is very powerful, we need to build
look-up systems, which allow you to translate an
The Internet of Things RFID code or a barcode into an HTTP address.
RWW: I’m fascinated by how the Internet is The Semantic Web is a web of things,
becoming more and more integrated into conceptually. Tying an actual thing down to a
the real world. For example the Internet of part of the web is the last link - the last mile.
Things, where everyday objects become Give the thing a notion of its own identity in the
Internet connected via sensors. Have you been web.
following this trend closely too, and if so what
impact do you think this will have on the Web in Conclusion
say 5 years time? RWW: The over-riding message in both Part
TBL: It connects very much with Semantic 1 and 2 of our interview with Tim Berners-Lee,
Web [and] with linked data. With Linked Data is for companies and organizations to make
you’ve got the ability to give a thing a URI. So their data available online. Preferably as Linked
I can give a URI to my phone, and I can say Data, which uses a subset of Semantic Web
that’s my phone in Linked Data. And also the technologies. But Berners-Lee noted, in Part 1 Nodalities Magazine is made
company that made it can give a URI to the of our interview, that he’d even be happy with available, free of charge, in
model of the phone. They can also put online all the data in CSV (comma separated values) print and online. If you wish to
the specs of the phone, and then I can make a format. subscribe, please visit
link to say that my phone is an example of that It’s clear that we’ve seen a lot of progress www.talis.com/nodalities
product. So now any system which is dealing in linked data already in 2009. In upcoming or email
with me and has access to that data will be able posts on ReadWriteWeb, we’ll continue to track nodalities-magazine@talis.com.
to figure out the sorts of things I can do with this trend and explain how organizations can
my phone, which actually is really valuable. contribute their data.
Especially if the phone breaks.
The Semantic Web has already given URIs to Orginal article is available at tinyurl.com/tblrww
things, and to types of things. When the things
themselves have an RFID chip in them, then
Nodalities Blog
I think it’s a very exciting world. One can take From Semantic Web to
that RFID chip, go to the Internet and find out Web of Data
the data about the thing. Whether we’ll be able
to do that, whether the manufacturers will be
blogs.talis.com/nodalities/
open enough to allow me to turn data about the


nodAliTieS mAgAzine gARlik

Garlik Announce Open-Source of 4Store
Podcast with Paul Miller

Tom: Garlik is an identity protection and fraud Paul: OK, good. Despite the fact that this data
prevention company aimed at consumers. We comes from a lot of different databases, a lot
help consumers to look after their personal of different websites, and comes out in a lot of
information in the online world. And we set the different forms, you’re actually storing it all in
company up about three years ago, working RDF, aren’t you?
with and for consumers as we saw huge
Paul Miller: Hello and welcome to this podcast, amounts of personal information appearing in Tom: Part of the reason we went down the path
with your host Paul Miller. Today I talk with the online world, and consumers being taken of semantic technologies was because we knew
Tom Ilube and Steve Harris from Garlik. We advantage of, particularly by identity theft and that in order to deliver the type of service we
discussed their decision to open source their so forth. wanted to, we were going to be wrestling with
RDF triple store, 4store, and make it freely So Garlik helps consumers look after their information from a huge range of sources and
available for download. personal information. coming at us in different formats. Some of these
For more information on this show, please sources of data are quite transient as well. So
see the notes on the website. Thank you. Paul: And how does it do that? they come up very quickly. We have to analyse
them, understand how to interact with them,
Paul: Tom, Steve, thank you very much for and then six months later they’ve disappeared
joining me for this podcast today. Before we talk again.
about Garlik and open source triple stores and
We created a service called And one of the challenges, in fact probably
all the other things we’re going to get to, I think Data Patrol which essentially the biggest challenge that I’ve found as chief
it would be worthwhile to have you both briefly patrols or harvests and looks information officer of Egg for a number of
introduce yourselves. So first off, Tom. Who are for information on your years—and I note that other CIOs find—is being
you? And what do you do? able to change your databases, being able to
behalf across millions of web
wrestle with heterogeneous data sources that
Tom Ilube: Sure. I’m Tom Ilube. I’m chief pages of dark markets where change constantly it seems. It’s a huge problem
executive of Garlik and founder of the company. information is traded. in large organisations.
I’ve been in technology for 25 years and was What I’ve found is the types of technologies
previously chief information officer of one the we’re working with now gives us a real
world’s largest Internet banks. Tom: We created a service called Data Patrol advantage in that area. It gives us a degree of
which essentially patrols or harvests and looks flexibility and adaptability that I just didn’t have
Paul: OK, good, and that’s Egg for those who for information on your behalf across millions of working with the previous set of technologies.
are listening along and anyone who wants to web pages of dark markets where information
get their Egg card out and say, “I’ve got one of is traded, on structured databases that have
them.” [laughs] been made publicly available, and analyses
OK, good. Thank you. And Steve, Steve that information to see whether there’s anything
Harris. Hello. Who are you, and what do you that puts you particularly at risk of identity theft
What I’ve found is the types
do? or exposing your privacy in a way you didn’t of technologies we’re working
intend. with now gives us a real
Steve Harris: Hi. I’m Garlik’s head of And then it alerts you and advises you on advantage in that area. It
architecture, and I’ve designed RDF systems what you should do about it. It’s a continuous
gives us a degree of flexibility
and some of our systems as well. And I’m proud scanning, monitoring process. We make it
about—I was a researcher at the University of available to people both direct and also through and adaptability that I just
Southampton working in the IAM Group. business partners such as banks, credit card didn’t have working with the
companies, Internet service providers, and so previous set of technologies.
Paul: OK, good. Thank you, Steve. And we’ll forth.
dig into some of the detail of that RDF stuff in a
moment. First through, you both work at Garlik. Paul: OK, and when we last I spoke, I think
Tom, what does Garlik do? because of the nature of the data, this service Paul: OK, and this underlying data storage is
was just available in the UK. Is that still the entirely homegrown, or have you brought things
case? in from outside?
Garlik is an identity protection
and fraud prevention Tom: It’s now available in the UK and in the Tom: The RDF storage is entirely homegrown.
US. We launched it in the US towards the Do you want to comment on that, Steve?
company aimed at consumers.
back end of last year—set up an office in New
We help consumers to York. We’ve both launched it there, and we’re Steve: Yes. We use Dave Beckett, who works
look after their personal talking to a number of partners. And we hope for Yahoo!, we use his SPARQL and RDF parser,
information in the to be launching in a couple of other European but the actual storage and the execution engine
countries—certainly one, possibly two—later on are all completely homegrown.
online world.
this year.



Paul: OK. I guess that is sort of the meat of developed at Egg. Maybe some of those could want to do this. Here are the reasons why. Is
the story today, is that you’re taking the step of have been released as well. this something you’d be willing to support us in
taking this system that you’ve developed over a doing?”
number of years, presumably at a fair degree of Paul: One of the things in the press release, you And it was really only about a week or so
expense, and you’re making it freely available mentioned that Garlik may be willing to provide ago that we had the final board meeting to
for everyone to have. Are you mad? sort of support to help other companies get discuss it and decide that we were happy to do
going with this. Is that a new business direction it, and we had their blessing, and let’s get on
Tom: [laughs] I was almost going to say for Garlik? Are you going to do that for free as with it. Actually, they’re quite excited about it.
absolutely, then I heard the question at the end. well?
“Are you mad?” Paul: OK, good. So the desire to do this would
Actually, we’ve been thinking about this for Tom: We won’t be able to do that for free, have been very clearly on the table in your last
quite a while and talking to our investors and so unfortunately, but what we felt is it’s useful to funding round, for example?
forth. And we decided take our RDF store, which some organisations to know there is a company
we call 4store, and make it freely available to that’s willing to stand behind the software Tom: Absolutely. We’ve talked to our VCs over
anyone anywhere—businesses, governments, and provide some support and guidance as the last year, and we did our last funding round
about six months ago—maybe less than that
now. So this has been in the conversation all
The reason why we wanted to do this was because as a company we along.

placed quite a big bet. We are betting that this set of technology— Paul: OK, good. Now. Steve, one of the things
RDF, the semantic web, the web of linked data—are going to that always comes up as soon as we start
become hugely important. talking about triple stores is how big will it go.
So for 4store, how big will it go?

universities, academics, developers—as open required. Therefore we decided that we would Steve: That question is always a bit hard to
source software. And we’re making it available do that. answer because it’s always more a case of, how
today for people to download and really do We’re not trying to build up a whole fast do you want it to be than hard limits. But the
whatever they want with it free of charge. major business line in that area. We’re not production system that backs Data Patrol has
The reason why we wanted to do this was transforming into a Red Hat or one of the currently about 15 billion triples, long-five. And
because as a company we placed quite a big other organisations that supports open source that runs fast enough to supply that service to
bet. We are betting that this set of technology— software. But we have developed the software. not many users.
RDF, the semantic web, the web of linked data— We understand it inside out.
are going to become hugely important. And I’m sure there’ll be a lot of people quite
really, as we look forward, are going to define happy downloading it, installing it, and getting
what the web looks like in many years to come. on with it without any support. And there’ll be
It’s a little difficult to predict,
And therefore as a company it’s actually some organisations that’ll want to be able to because it depends on the
in our interest to help make that happen. So if turn to us and ask for support. And we’ll be shape of the data and how
we have technology that can really help other there to provide that. complex your data is and the
people to store and publish data in RDF form,
then that’s what we want to be able to do. And Paul: OK. And you’re venture capital backed—
kind of queries you
so once there was some computations to be Garlik. want to run.
had along the way to get everyone aligned, the
idea of doing, from our company’s point of view Tom: That’s right.
makes a lot of sense. So it really depends how much hardware
Paul: How did you get his past VCs? you throw at it, and how fast you want it to be.
Paul: But you’re not a software company. It’s a little difficult to predict, because it depends
You’re a company solving identity management Tom: [laughs] The one thing I would say is if on the shape of the data and how complex your
problems that happens to need to use software you’re going to do something like this, you don’t data is and the kind of queries you want to run.
to get there. Is this not a dilution of your core spring it on your VCs. It’s sort of a long process. Paul: OK, but in the press release there’s the
mission? So we probably started talking about potentially figure of 60. 60 billion is mentioned. Is that a
doing something like this maybe a year or so projection, presumably, of how far you think it
Tom: No, it isn’t really. One of the advantages, in ago with our VCs. Amongst ourselves, we’ve could go?
a way, is we’ve built this software to be used for probably been talking about it earlier than that,
our own purposes and it’s been in production but with our VCs perhaps a year ago I would of Steve: That’s based on partly projection, and
for the last three years. So it’s a good, hardened started to raise it at board meetings. partly based on non-production test unit we’ve
bit of software and I think people will see that To say, “Look I think we might want to do done.
when they try and download it, install it, and something like this.” But I wouldn’t push it
make use of it. too far straightaway. What I was essentially Paul: OK. And that makes it pretty big actually.
It’s not a distraction for us at all to release doing was seeing whether my VCs were aware I mean, if you look there’s a page on the World
it like this. In fact, I think more companies of these sorts of approaches, what their gut Wide Web Consortium site that I’ll link to in the
ought to be looking at their software portfolio reaction was, and so forth. show notes that talks about large triple stores.
and thinking about what they can release. And Essentially, we’ve taken them on a journey. And most of them are sort of 10, 12, 15 billion.
perhaps with hindsight, I could have looked Particularly over the last six months it’s sort of You’re bigger than that already, and project you
at some of the great bits of software that we accelerated, where we’ve said, “Look, we do could go three or four times that size.



Steve: Yes. Scalability was one of the major Steve: Internally, we’ve only ever run it on Steve: Yes, that’s how we use it in our... So the
design features. We knew we were going to Red Hat derived Linux machines and on Mac QDOS QDOS.com services and also Data Patrol
have a lot of data to support and service. So desktop machines for our developers, just to use the RESTful API on top of their SPARQL
from the outset we accepted some of those use single machines. It’s a fairly portable piece endpoints to add and remove data.
scale problems. 4store not only works at of software. The clustering support depends
that size, it’s also still quite fast. If you have a on a multicast DNS library which is not all that Paul: OK. And how fast will that work?
sufficiently powerful machine, you could import portable, but other than that it should run on
data at 120 or 150 thousand triples a second all pretty much any Unix system. Steve: So for bulk imports, it’s slower to put
the way up to tens of billions size. The only thing we’ve really got any individual HTTP RDF files over HTTP than it is
experience running it on is Red Hat derived to use the command line tools. Again, it’s really
systems. We tend to use Centos. difficult to tell. It depends on the exact structure
Scalability was one of the of the file. But for sort of normal sized RDF files,
Paul: OK. And presumably people can it’s not really a problem—40 to 50 thousand
major design features. download it for Ubuntu or whatever it may be triples a second should be easily attainable.
and test it. And if it doesn’t work, find out why it
doesn’t work and contribute those fixes back to
Paul: So to achieve these scales, have you the code line.
done something fundamentally different to the
Once it’s up and running, it’s
others? Or have you simply optimized existing Steve: Yeah, I think because it runs on OS10 very simple to use.
practices better than they have? and on Red Hat, it wouldn’t have any problems
on something like Ubuntu. I think if you wanted
Steve: Yes. It’s built from the ground up to be to try and run it on, say, Open Solaris you might Paul: OK. Good. Now one of the domains
a clustered system. Although it does run on a find a couple of challenges, but I think any Linux obviously where’s there’s quite a lot of interest
single machine, a cluster of one. When you take system would be very straightforward. in semantic web stuff at the moment is the
that approach, it leads you down a different Paul: OK, good. So you’ve intrigued me. whole area of government. We were talking
design route. You’ve got me interested in this massive triple before the recorder went on about some of the
Another thing we’ve done is, rather than store. I’ve got a big data problem in my office. work that Tim Berners-Lee and others are doing
building it out of a modern more trendy What do I have to do to get going once I’ve in that space at the moment. It seems to be
language, we built it out of C just to get that downloaded it? I’m probably use to Oracle or attracting quite a lot of attention from outside
raw performance, especially in the IL area. We MySQL or something. Presumably it’s quite a the semantic web community.
found it very hard to get the same amount of shift to move from thinking about a database to Is something like 4store—are we going to
performance out of higher level languages. thinking about an RDF triple store. see it running in the home office any time soon?

Paul: OK, and you talked about designing Tom: I think it will be up to the various
it for clustering. Is 4store designed for high government departments. If they find it
performance machines in a cluster, or is it Rather than building it out interesting or not, it’s available to anyone who
designed to run on commodity hardware? of a modern more trendy wants to use it.
I think there are a few different stores and
Steve: We run it on commodity one looped
language, we built it out open source RDF stores out there. Yeah,
Linux service. It does like a lot of RAM, because of C just to get that raw some of them are sort of more feature-rich and
it tries to put as much of the index in RAM performance, especially perhaps slower, and ours is fast and large and
as possible. But really it’s not all the fussy. in the IL area. basic. And I think different organisations will
Other people that are running it are running decide what they’re requirements are, what
it on smaller clusters of much more powerful they’re trying to do, and decide whether ours
machines. It self-adapts to whatever hardware lends itself to what they want to do.
environment it’s installed on. Steve: Well, not really. We designed it so once But I think overall we would be happy. The
it’s installed it’s actually very easy to use. It more organisations that are making raw data
Paul: OK. And we’ve been talking quite a lot doesn’t have exactly the same model as an available in RDF form, the better. So if our store
about downloading it, and clearly it’s available Oracle or a MySQL. If you’ve got a cluster set contributes to some organisations doing that,
on the website now to download. Do you have up or a single machine, you start an individual then that’s a jolly good thing.
any plans to offer a hosted arrangement for backend for every database that you want to
people who don’t feel capable or who don’t query. Paul: Absolutely. “Raw data now,” as Tim
want to download and run software of their And then you just run an HTTP server, which Berners-Lee has been saying recently.
own? attaches to the database itself to provide the So the store itself will allow an organisation,
Tom: We’re not really planning to go in that SPARQL endpoint. And once you’ve done that, whatever they may be, to clearly use that
direction. I think that there are a number of you can use RESTful HTTP operations to add as the backend to some application of their
companies that do that, and I suspect they and remove data. Or there’s command line tools own. Is it also capable of contributing the
might be interested in taking our software and if you prefer that kind of thing. data that’s in the store back out to this wider
providing that sort of hosted arrangement. But Once it’s up and running, it’s very simple to linked data cloud? Or does it simply sit behind
it’s not sort of a service line that Garlik itself will use. applications?
get into.
Paul: OK, and you said RESTful APIs for Steve: So the QDOS.com service is also a
Paul: OK. And what sort of machines will it run pushing the data in and out? So that’s how I linked data application. And the linked data
on? would attach it to my existing applications, is it? interfaced to that is provided by 4store. When



a linked data press comes in, it’s just rewritten Paul: OK. Now you’re making this available But I think people will find they can go a long
into a SPARQL construct statement. And that’s under an open source license. So presumably way in the direction they want to with 4store and
very fast. It’s quite highly optimised. people are going to take it and they’re going other stores that are out there in similar sort of
The SPARQL interface was designed so that to start tinkering with it, and they’re going to scale.
it could be exposed to the outside world. It’s start adding things in that they feel they need
got a lot of features in there to—some of the and want. Are you going to contribute those Paul: Yeah, good. So are you going to be
SPARQL features disabled by default to make it additions back into the code line? actively and aggressively pushing 4store, or is
safe to deploy inside enterprise environments. it simply going to be put out there? Built it and
A typical SPARQL system is not safe to be Tom: Yes, absolutely. Yes. So our needs inside they will come?
deployed inside a secure enterprise system Garlik in some ways are quite better. Although
because of some of the features of SPARQL. So we have a lot of data, we don’t do things like Tom: We’ll certainly want people to know about
that stuff’s disabled by default. natural language search and so on, although it it, so I think we would really like it if people were
And also it has some features in there that wouldn’t be too hard to add that kind of thing to thinking about using an RDF store and having
enable you to limit the amount of effort you’re the engine. a look at what was out there and what was out
willing to put into arbitrating any particular So if people have those needs and they’re there in the open source world, then we’d hope
query, dependent on where it’s coming from. willing to make those patches available, we’ll that 4store would be one of the things that they
But we’ve never done that on Angus so we’re definitely include them in the main line. look at.
not prepared to say it’s safe to expose the Paul: OK. And do you think continued But I don’t think we’re going to be sort of
SPARQL endpoint and let random people on the development will make it get bigger as well, or trying to compete with others, in a sense. I
street come up and try a SPARQL query at the would you almost have to start again to build think that the more options there are for new
gates. For that, for certain, you want to design something that was capable of scaling further? organisations and new people looking at the
criteria. whole linked data world, the better.
Steve: Well there are two things on that really. I know when I was a CIO, if I was trying to
Paul: OK, so the intention there would be It could probably improve and get bigger in make a technology decision, I always liked it
that if I or someone else either deliberately or certain respects, as it is. But what we’ve done to be three or four or five options with different
accidentally sent a very expensive query to your internally—we’ve been working on building characteristics so that I could decide what was
right for my organisation. So we’re keen that
as people are looking at making this sort of
decision of moving in this direction, this is one
of the options that they look at.
The store itself will allow an organisation, whatever they may be, to And we would really like to see the more
clearly use that as the backend to some application of their own. organisations that are moving in this direction
that are using tools like ours and others to
publish their data, the better. Because as a
company what we’re excited about is what
system, it should be able to say, after a certain another generation of RDF store. The scale is the world looks like in three or four years time
period of time, that it’s going to stop and fill the probably on order of magnitude bigger than when thousands and thousands of mainstream
query. 4store does. organisations are publishing their data as RDF,
And as their systems and our internal part of the Linked data cloud.
Steve: Yeah, it does IL accounting, so it’ll needs grow, then we’ll start to use that store to That will be a fascinating world to play in.
allocate out a certain amount of effort. And underpin them. And that one will scale much, And we’re hoping that what we have done is to
you can say how much effort you’re willing much, much larger than 4store will. help to contribute towards that.
to expend on a query. And then you get only
a partial result and a warning saying that it’s Paul: And is this called 5store?
partial, and that it was limited by the complexity.
[laughter]
We would really like it if people
Paul: Right. And presumably then if you’re a
valid query, you can contact the system provider Steve: Internally we call it Keep, but we haven’t
were thinking about using an
and ask them to give you more time to run your got an external name for it yet. RDF store and having a look
query. at what was out there and
Steve: Yeah, you have to provide a separate Paul: OK. And presumably there’s also this sort what was out there in the open
endpoint that has slightly more resources of related school of thought that suggests we
allocated to it. don’t necessarily need to lump everything into
source world.
one store. So 60 billion triples or there about
Paul: Right, OK. I guess one of the more real- may actually be big enough for any one store,
world applications of that would be perhaps that and we need to be getting better at creating the
internal systems get more time than external connections between stores working together. Paul: What would that world look like? If all
systems do. these claims certainly for government data
Steve: Absolutely. Yeah, yeah, and we’re become real and if organisations of various
Steve: Yes. That’s what we’ve done. We have thinking about... other kinds follow along and start exposing
provided SPARQL endpoints to a few people It think it’ll vary between—there’ll be some data. Wearing your identity hat for a second,
that want to run applications on top of QDOS, huge organisations that decide that everything I guess, as a corollary perhaps to the brave
but our internal ones go through a different they want to do internally and actually it turns vision, what would that world look like?
endpoint with a much higher limit on it. about that they do need big centralised stores.



Tom: Gosh, I think the organisation would be a Paul: OK. And presumably if people keep using the key thing. That is the main benefit we see
fascinating world. I try not to predict too much them then there is still a need for them. is the flexibility and the ease of use of relational
what it would look like and just try and help databases.
to make it happen and then get stuck into it. Steve: Yeah, exactly. Yeah.
But I think the issues of identity and privacy Paul: And you actually have real world
and control and power over your personal examples to point to as well, because you have
information shoot up the agenda hugely in that been using it in Angus yourselves for a number
world. So it’s going to be an of years?
In a world of masses interlinked data interesting time. And as you
wherein who you are and who said what and Steve: Absolutely, it makes a real difference
say, 4Store is there on the
what personal data there it is very visible, then being able to say to organisations—I’ve been
it becomes a real challenge, but also a real table for people who want a CIO myself. You can be quite nervous. You
opportunity for individuals to understand what to start looking at how they are thinking about introducing a new set
data is out there about them and what they can can make use of all this data. of technology and it looks like there’s only
and can’t do with it. downsides. [laughs]
Now, within the world of triple
I think my view is that the world of personal It looks like if you introduce this new
information is changing in a really sort of
stores, there are an awful technology your chief executive or your
fundamental way. And we’re going to have to lot of them. marketing director is going to come and slap
stop talking about wouldn’t it be great if things you around the head and say, what on earth
were how they used to be? And how do we hold did you do that for? And the old stuff was OK,
back the tide? And how do we get our privacy wasn’t it?
back in the box, as it were? Paul: OK. Good. So just before we finish off
And recognise that the world is changing then, do either of you have any closing thoughts
quite fundamentally and the challenge is how on where we go from there?
do you as an individual survive and thrive and There are hundreds of
take advantage of opportunities in that world. Tom: I think that what we need to do, those of thousands of organisations
I think it is sort of two sides of the coin, both us that are close to this area, is to try and bring out there that will need to
protecting yourself and taking advantage of the wider technology world into it and to make
the information that is out there about you and them aware of the advantages of adopting these
be looking at these sorts of
others. technology sets and at least exploring them and technology sets, from us,
Sort of two sides of the same coin. We’re trying to get their minds around them. from other people, thinking
going to have some really interesting times over There are hundreds of thousands of about what it means to their
the next few years, thinking about how you live organisations out there that will need to be
in that world. looking at these sorts of technology sets,
businesses.
from us, from other people, thinking about
Paul: Indeed we are. And it’s going to be an what it means to their businesses. And I think
interesting journey and interesting finding out one of our challenges, if you like, is both to So before you introduce something new you
how we cope and how we make the most of it make our toolsets available for those sorts of need to be pretty convinced that it is going to
because, as you say, it is a big opportunity for organisations to engage with without making give you some real benefits. And if you can talk
all of us and for organisations building systems too much fuss, but to give them a way in which to someone who has been using it to support
and technologies to take advantage of it. they can start to explore without making a tens of thousands of users in a production
So it’s going to be an interesting time. And massive commitment. system day to day, even if you don’t use that
as you say, 4Store is there on the table for And also figure out how to speak to the same technology set that they are using, the
people who want to start looking at how they wider world in a language that they understand. fact that there are people out there like us and
can make use of all this data. Now, within the I think that sometimes you can get so wrapped others using this technology in production
world of triple stores, there are an awful lot of up in the world that you live in that you can day in and day out, to support large scale and
them. forget the 90 odd percent of the rest of the world growing businesses. It means a lot to the main
Most tend to be an awful lot smaller. Do you has no idea what you are talking about. stream technology guys.
think we are going to see some harmonisation I think sometimes we run into that problem
in that market? I’m not suggesting that it will be in semantic technology world as well. But I think Paul: Yeah, definitely. Good, well thank you
just one. But do you think we are going to see over the next sort of couple of years, getting both very much for that. and I look forward to
some of the smaller, older ones falling by the more and more organisations, more and more seeing what people make of 4store. They can
wayside? smart technology people who just haven’t go to your website and download it and try it for
happened to encounter this world particularly to themselves.
Steve: I guess there will be some that fall out come and play and get involved and experiment
of use, but if you look at the relational database and try things. It’s what we need to do. So thank you and good luck.
market, which is the nearest equivalent. There’s
really an awful lot of software in that domain, Paul: Yeah. Steve? Steve: Thank you very much.
just spanning different needs and different
emphasis and catering different languages and Steve: No, what Tom said, really. I think getting Tom: Thanks.
so on. the message out there that this real world
So I think the number of RDF stores we’ve practical problems that can be solved quickly
got at the minute certainly isn’t unsustainable. and efficiently and sustainably. Technology is


nodAliTieS mAgAzine

Nodalities Magazine
SUBSCRIBE NOW
Nodalities Magazine is made available,
free of charge, in print and online.

If you wish to subscribe,
please visit www.talis.com/nodalities
or email nodalities-magazine@talis.com.

Read Previous Issues of Nodalities Magazine Online
Want to contribute? Please visit www.talis.com/nodalities or email nodalities-magazine@talis.com.


nodalities_issue7.pdf

More Related Content

What's hot (14)

Viewers also liked (9)

Similar to nodalities_issue7.pdf (20)

More from Hiroshi Ono (20)

Recently uploaded (20)

nodalities_issue7.pdf