Second day

In fact, there is much less to say… In the morning I was on two workshops; I was at the Uncertainty Reasoning on the SW one for a while, but then I was asked to participate at a panel at the Semantics for the Rest of Us one, so I had to switch. This was a bit unfortunate, because I could not really ‘dive in’ to any of the two. And my afternoon was taken up by ‘networking’, catching up with some people on many many issues that are not worth blogging (yet?).

I listened to Kathryn Laskey’s presentation on how to combine probability theory in the mathematical sense (the good old Kolmogorov axiomatic theory on probability that I learned at university in a distant past…) with first order logic. I cannot claim to have really understood all the details but it made me curious enough to put reading her paper on my to do list…

As for the panel “Little vs Large Semantics: What’s next for the Semantic Web languages?”, with Leigh, Kendall and Ora on the panel besides me… it was not that exciting, I must admit. Maybe the main message I take away from it was the passionate request of Chris Welty to re-open RDF (see also Pat’s keynote below!).

Third day (well, first real conference day)

Preamble: I would have wanted to add links to papers. And I couldn’t: I have not found the papers on the Web. Neither on Springer’s site nor elsewhere. I may have missed a reference somewhere, if somebody knows then tell me. But if the papers are not available, I think it is a shame…

The conference began with a keynote of Pat Hayes. Entertaining and also thought provoking; Pat is a great speaker. What really interested me is his talk on ‘RDF Redux’; I was actually anxious to listen to that one at SemTech last June but he had to call this off back then. So he repeated it here. This is typically the kind of talk that needs more thinking afterwards to understand it (and Pat has promised to write it down!), but he essentially proposed to re-think and re-do some of the fundamentals of RDF semantics. Instead of set-based model theory which we have today, and which makes the treatment of b-nodes, shall we say, a bit complicated (some would use harsher words:-) we should consider RDF graphs as ‘things’ on a ’surface’ (think of it as a real surface on a sheet of paper) and b-nodes are just ’scratches on that  surface’. (A bit like ‘context’ of a graph?) Because these surfaces are different from one graph to the other, when a merge occurs then in fact a new surface is created where the unified graph is put, and the issue of b-nodes becomes natural (instead of the ‘renaming’ procedure that the current semantics document describes). Pat claims that the whole semantics could be re-written that way and none of the current RDF implementations would change. But one can go one step further: there may be different kinds of surfaces (eg, negations) and surfaces can have a name (a bit like named graphs) and all can be put together to provide a powerful semantics for these entities. His further claim was that such an extended semantics of RDF could be powerful enough to describe, conceptually, RDFS or even OWL, ie, the semantics should not be layered any more.

No way I would accept all this argumentation on face value:-), so I have to think about this and, mainly, read whatever Pat may want to write down to understand it. In the meantime, I may have to look into the concepts of conceptual graphs, and the Peircian notation of logic that Pat referred to as inspiration…

A more general take away (see also Chris’ remark above): maybe it is time to look into RDF again? A scary thought. Touching to something that is fundamental on the SW has to be done with extreme care… We will see.

There were two papers in the same session that were very close in subject and topic: one of Jesse Weaver and Jim Hendler on the parallel materialization of RDFS graphs and the one of Jacopo Urbani et al on using MapReduce for RDFS reasoning. (Sigh…, this is where I would like to put a reference!) Both aimed at similar challenges, namely the materialization of RDFS inference results of a graph using parallel computing methods. And there was one more similarity: both had some sort of a classification of the rules in the rule set described in the RDF Semantics document to help improving the processing. (Eg, to analyze which rules should be duplicated among processing nodes and which one can be handled without, or which one need a special treatment for a map-reduce pair). It seems that it would be worthwhile to see if some of these classifications (‘ontology rules’ and the like) could be extended to OWL 2 RL (Jesse Weaver told me afterwards that they want to look into this).  But, to put things into perspective: we are the points when billions of triples can be expanded with relative ease. Who would have thought a few years ago? There was also a remark on one of Jesse’s slide (I do not remember the exact wording) which said that RDFS is insanely parallelizable:-)  It was a really interesting session.

The SW in use session included  a paper from Landong Zuo et al “Supporting multi-view network analysis to understand company value chains”.  Integrating a bunch of data in the UK on companies, integrating them in an RDF store, and let users get information on the ‘value chains’, ie, how companies relate to one another as producers/consumers. Technically, the interesting point was the fact that users had the possibility to interactively add new relationships, new classifications to the system, essentially new rules that could be evaluated. The whole system seemed to be a really cool, a well engineered and well functioning machinery. As the speaker put it, although all conclusions drawn from the system could be found by the users by analyzing databases, but it would take weeks to do what this system can give them in a few minutes. This is exactly the kind of message we need for the outside  world about the usefulness of Semantic Web technologies.

On another session Martin Szomszor presented an experiment they conducted at the ESWC conference, combining RFID-based personal badges with an underlying SW system. The resulting system could be used to show personal contacts among delegates, could help people find others with similar interest, could retrace later whom one met at what point (“I remember talking to that chap, but I do not remember his name!”), etc. Lots of privacy issues, for example, but I would have liked to see that in practice, that is for sure!

Stéphane Corlosquet’s presentation on SW and Drupal was really exciting. I already knew about the plans of Drupal 7 to incorporate RDF management from the start, that all Drupal 7 pages will be annotated via RDFa. The RDFa community has been  fairly excited about that for a while now. But the work done by Stéphane and others provide some additional modules that makes it easy to add a SPARQL endpoint to a Drupal based site easily, to import other RDF content, or to manage the vocabularies used on the pages and the like. They already have such a system running with the current Drupal, but these modules will become part of the standard Drupal 7 module set that one can download from the drupal site. And that is cool.  It significantly lowers the barrier to build Web sites that are prepared to be part of the Linked Data cloud, even if the system administrators are not SW experts. I expect this to open up quite a lot of possibilities…

Off to the next day! More paper and the presentation of the Semantic Web challenge finalists…

Posted in Semantic Web, Work Related Tagged: drupal, parallel computing, probabilistic reasoning, RDF, RDFa, RDFS, SPARQL

This morning, I gave the keynote address (my slides) at RuleML 2009. I assumed the audience would be fairly familiar with rule systems and rule technologies, but not necessarily with RIF, the Semantic Web in general, or my sense of the future of the Semantic Web (for which RIF is important).

Those slides may be boring for some of you. The interesting new bits:

1. I made a new diagram for the RIF dialects. During the talk, I presented it as a successive reveal: the chaos of rule system features, the BLD grouping, the PRD grouping, and the Core intersection. Here’s the final slide:

   

2. I wanted to convey that the Semantic Web means real change, and I wanted emotional impact, so I took a shotgun approach and enumerated a list of likely data sources that was big enough to have some surprises for most folks; then I moved into a long list of things you could do with that data. Some things on the list were boring (having impact only by showing how long the list is), but some got the desired wide eyes and shocked sounds as people realized this just might happen. I think it went over well.

   My list of types of data we’ll be seeing:

   • From producers: product information
   • From sellers: product and service offerings
   • Customer support (instructions, upgrades, …)
   • Social network (who you trust, who interests you)
   • Personal information shared with friends
   • Public records (financial, legal, political, …)
   • Science (medical, environmental, economic, …)
   • News, Blogs, Public photos, videos
   • Event listings (performances, meetings)
   • Review, opinions, product experiences, preferences
   • Personal location, location history
   • Financial transactions

   Which led into my general scenario: you’re in a store about to buy something, but first you scan it with your phone and look up a little more information about it. What might you look up?

   • Its price at other stores, nearby
   • Its price for delivery, and how how long you’d have to wait
   • Maybe: where it was made, and under what conditions
   • Is it’s producer a good corporate citizen?
   • Does its producer agree with your political views (uh oh)
   • How many houses does its CEO own?
   • Did your spouse/housemate just buy some? Or something like it?
   • How your friends feel about this product
   • How Consumer Reports (or some such service) reviewed it
   • Product liability suits
   • Endorsements
   • Maybe: Payment for Endorsements
   • For electronics: compatibility information
   • For mechanical items: How repairable is it? MTTF, MTTR
   • For food: nutritional information, health benefits and risks
   • Demographics of this brand, these products

That’s all for now. I saw the Bellagio Fountain from my 29th floor
hotel room late last night, but I’d like to see it up close.

I was at the SemTech conference in June when Evan Sandhaus from the New Your Times gave a keynote and when he announced that the NYT would gradually publish many of their data as Linked Data using Semantic Web technologies. Unfortunately, I had to leave on the last day of the ISWC2009 last week when they announced to keep their promise, and release the first 5,000 subject headings tags to the LOD. Which is really great news.

I remember Evan saying in Santa Clara (maybe privately, I do not remember that detail) that they are newcomers in this area, and it will be difficult to get it right (and, well, there are bugs, as, for example, Eric Hellman or Richard Cyganiak pointed out in their respective blogs). But I think we should really applaud when such a promise is held…

Posted in Semantic Web, Work Related Tagged: Linked Data, new york times

Here’s what I think should be standardized at some point, soon, in the Semantic Web infrastructure. These items are at various levels of maturity; some are probably ready for a W3C Working Group right now, while others are in need of research. They are mostly orthogonal and most can be handled in independent efforts. (I would lean against forming a single RDF Working Group to handle all of this; that would be slower, I think.)

To be clear, when I say “RDF 2″ I mean it like OWL 2: an important step forward, but still compatible with version 1. I’m not interested in breaking any existing RDF systems, or even in causing their users significant annoyance. In some traditions, where the major version number is only incremented for incompatible changes, this would be called a 1.1 release. In contrast, at W3C we normally signal a major, incompatible change by changing the name, not the version number. (And we rarely do that: the closest I can think of is CSS->XSL, PICS->POWDER, and HTML->XHTML). The nice thing about using a different name is it makes clear that users each decide whether to switch, and the older design might live on and even win in the end. So if you want to make deep, incompatible changes to RDF, please pick a new name for what you’re proposing, and don’t assume everyone will switch.

This is partially a trip report for ISWC, because the presentations and especially the hallway and lounge conversations helped me think about all this.

Note that although I work for W3C, this is certainly not a statement of what W3C will do next. It’s not my decision, and even if it were, there would be a lot of community discussion first. This is just my own opinion, subject to change after a little more sleep. Formally the decisions about how to allocate W3C resources among the different possible standards efforts are made by W3C management guided by the the folks who provide those resources, via their representatives on the Advisory Committee (AC). If the direction of the W3C is important to you or your business, it may be worthwhile to join and participate in that process.

1. RDF and XML interoperation

There’s a pretty big divide between RDF and XML in the real world. It’s a bit like any divide between different programming languages or different operating systems: users have to pick which technology family to adopt and invest in. It’s hard to switch, later, because of all the investment in tools, built systems, educations, and even socially networks. (People who use some technology build social and professional relationships other people who use the same technology. Thus we have an XML community, an RDF community, etc. Few people are motivated to be in both communities.)

I think we should have better tools for bridging the gap, technologically, so that when data is published in XML, it’s easy for RDF consumers to use it, and when the data is published in RDF, it’s easy for XML consumers to use it.

The leading W3C answer is GRDDL, which I think is pretty good, but could use some love. I’d like to see support for the transforms being in Javascript, which I think is probably the dominant language these days for writing code that’s going to run on someone else’s computer. It certainly has a bigger community than XSLT. I’d probably support Java bytecode, too.

I would also like to see some way to support third-party GRDDL, where the transform is provided by someone not associated with either the data provider or data consumer. Nova Spivack gave a keynote where he talked about this feature of T2. They’re focused on HTML not XML, but the solution is probably the same.

Beyond GRDDL, I think there’s room for a special data format that bridges the gap. I’ve called it “rigid rdf” or “type-tagged xml” in the past: it’s a sub-language of RDF/XML, or a style of writing XML, which can be read by RDF/XML parsers and is also amenable to validation and processing using XML schemas. Basically you take away all choices one has in serializing RDF/XML.

I note the The Cambridge Communiqué is ten years old, this month. It proposed schema annotation as an approach, and that’s not a bad one, either. I haven’t heard of anyone working on it recently, but maybe that will change if the XML community starts to see more need to export RDF.

Amusingly, while I was talking to Gary Katz from MarkLogic about this, he mentioned XSPARQL as a possible solution, and I pointed out Axel Polleres (xsparql project leader) was sitting right next to us. So, they got to talk about it. XSPARQL doesn’t excite me, personally, because I don’t use either SPARQL or XQuery, but objectively, yes, it might solve the problem for some significant userbase.

2. Linked Data Inference

For me, an essential element of a working Linked Data ecosystem is automatic translation of data between vocabularies. If you provide data about the migration of frogs in one vocabulary, and my tools are looking for it in another one, the infrastructure should (in many cases) be able to translate for us. We need this because we can’t possibly agree on one vocabulary (for any given domain) that we’ll all use for all time. Even if we can agree for now, we’ll want this so that we can migrate to another vocabulary some time in the future.

Inference using OWL (and its subsets like RDFS) provides some of this, but I don’t think it’s enough. RIF fills in some more, but the WG did not think much about this use case, and there’s might be some glue missing. Maybe we can get WG Note out of RIF to help this along.

I’d like us to be clear about first principles: when you’re given an RDF graph, and you’re looking for more information that might be useful, you should dereference the predicate IRIs to learn about what kinds of inference you’re entitled to do. And then, given resources and suitable reasoners, you should do it. That is, the use of particular IRIs as predicates implies certain things, as defined by the IRI’s owner. The graph is invoking certain logics by using those IRIs. (Of course you can always infer things that were not implied, but as among humans, those “inferences” are really just guesses you are making. They have quite a different status from true implications.)

If this is put together properly, and the logics are constructed in the right form, I think we’ll get the dynamic, on demand translation I’m looking for. I imagine RIF could be very useful for this, but reasoner plugins written in Javascript of Java bytecode could be a better solution in some cases.

Some of my thinking here is in my workshop keynote slides, but later conversations with various folks, especially Pat Hayes and TimBL, helped it along. There’s more work to do here. I think it’s pretty small, but crucial.

3. Presentation Syntaxes

RDF, OWL, and RIF all have hideous primary exchange syntaxes and some decent not-W3C-recommended alternative serializations. I’m not really sure what can practically be done here that hasn’t been done.

At very least, I’d like to see a nice RDF-friendly presentation syntax for RIF. A bit like N3, I suppose. I did some work on this; maybe I can finish it up, and/or someone else can run with it.

OWL 2 has 3+n syntaxes, where n is the number of RDF syntaxes we have. Exactly one of those syntaxes is required of all consumers, for interchange. I’ll be interested to see how this plays out in the market.

4. Multi-Graph Syntax

Most systems that work with RDF handle multiple graphs at the same time. Sometimes they do this by storing the triples in a quad store, with the fourth entry being a graph identifier. This works pretty well, and SPARQL supports querying such things.

We don’t have a way to exchange multiple graphs in the same document, however. N3 has graph literals (originally called contexts), and there was some work under the term named graphs, which is kind of the opposite approach.

Personally, I don’t yet understand the use case for interchanging multiple graphs in one document, so I’m not sure where to go with this.

Hmmm. I guess RIF could be used for this. You can write RDF triples as RIF frame facts, and the rif:Document format allows multiple rulesets, each with an optional IRI identifier, in the same document. ETA: RIF also gives you an exchange syntax where you can syntactically put literals in the subject and use bnodes as predicates, if you want. But now you’re technically exchanging RIF Frames instead of RDF Triples.

5. RDF Graph Validation

When writing software that operates on RDF data, it’s really nice to know the shape of the data you’ll find. It’s even nicer, if software can check to see if that’s actually what you got. And if reasoners can work to fill in any missing peices.

I don’t exactly understand how important or unimportant this is. It’s closely related to the Duck Typing debate. Whatever mechanisms make duck typing work (eg exception handling, reflection, side-effect-free programming) probably help folks be okay without graph validation. But I think folks trained on C++/Java or XML Schema would be much happier with RDF if it had this

The easiest solution might be using rigid RDF. One could probably also do it with SPARQL, essentially publishing the graph patterns that will match the data in the expected graphs.

The most interesting and weird approach is to use OWL. Of course, OWL is generally used to express knowledge and reason about some application domain, like books, genes, or battleships. But it’s possible to use OWL to express knowledge about RDF graphs about the application domain. In the first case, you say every book has one or more authors, who are humans. In the second case, you say every book-node-in-a-valid-graph has one or more author links to a human-node in the same graph. At least that’s the general idea. I don’t know if this can actually be made to work, and even if it can, it risks confusing new OWL users about one of the subjects they’re already seriously prone to get wrong.

6. Editorial Issues

Finally, I’d like some portions of the 2004 RDF spec rewritten, to better explain what’s really going on and guide people who aren’t heavily involved in the community. This could just be a Second Edition — no need for RDF 2 — because no implementations changes would be involved.

I’d like us to include some practical advice about when/how to use List/Seq/Bag/Alt, and reification, maybe going so far as to deprecate some of them (IMHO, all but List). Maybe bring in some of the best-practice stuff on publishing and n-ary relations.

I understand Pat Hayes would like to explain blank nodes differently, explicitly introducing the notion of “surfaces” (what I would call knowledge bases, probably). Personally, I’d love to go one step farther and get rid of all “graph” terminology, instead just using N-Triples as the underlying formalism, but I might a minority of one on that.

ETA: Of course we should also change “URI-Reference” to “IRI”, and stuff like that.

---

Okay, that’s my list. What’s yours? (For long replies, I suggest doing it on your own blog, and using trackback or posting a link here to that posting.) Discussion on semantic-web@w3.org is fine, too.

Fourth day

Shame on me, but I missed the morning keynote… I was a bit late arriving to the conference site and I got stuck in a conversation at breakfast. Things happen…

The most notable event in the morning, at least for me, was the SPARQL WG panel. All members of the Working Group (me included) were on the panel and the room was full. I mean, full, people were standing in the back. And I regard that as a success by itself, it shows not only the overall importance of SPARQL, but the real interest around the new version, ie, SPARQL 1.1 (in case you have missed it, the first working draft has just been published a few days ago). Lee Feigenbaum (co-chair of the group) gave a quick overview of the new features and then questions came.

The difficulty of the SPARQL 1.1 work is that it has to find a balance between what is realistic to standardize in a relatively short time frame and what could be good to see in a new query language. As a consequence, there are features that the community has discussed but have not made it into the document, or only in a simple format. That came up during the discussion but I had the impression that the audience, by and large, understood this balance. Actually, for some, the set of new features were even too much for an efficient implementation. I have the feeling that  the WG will have to publish a separate conformance document (a bit like OWL 2 has), because there is a certain confusion on whether a conforming SPARQL implementation will have to implement, say, update or inference regimes or not. That clearly came up through the questions. Anyway, remember one email address (yes, it is a bit of a mouthful): public-rdf-dawg-comments@w3.org this is where comments have to be sent on SPARQL 1.1!

I chaired a session on the use track in the afternoon.  The paper of Daniel Elenius et al on reasoning about resources (for military exercises) was interesting to me because it was based on reasoning with relatively large OWL ontologies plus rules. The OWL ’side’ was not very complex (Daniel referred to DLP, today I would say probably OWL 2 RL) but extended with extra rules. What this shows that when RIF will be finished and published, the combination of OWL with RIF may become very important for tons of practical applications. (As an aside, a nice little joke from Daniel: what is the system used by the military today when planning for exercises? The system is called BOGSAT. It stands for ‘Bunch Of Guys Sitting Around a Table’…)

Roland Stuhmer gave a very different style presentation on how user events (clicks, combination of clicks, etc) can be collected, categorized, and integrated into an application, analyzed with some rules for, eg, targeted ads. The system is based on harvesting not only the structure of the Web page, but annotations appearing in the Web page via RDFa. The result is an RDF structure describing the events that can be sent to a server, analyzed locally, distributed, etc. Nice usage of RDFa, but also important to have a Javascript API that can retrieve the RDF triplets from the RDFa structure attached to a specific node. (B.t.w., the old graphics standards of the 80’s and 90’s, called GKS or PHIGS, had notions of combined event structures with different event types. I do not remember all the details any more, but may it be worth looking at those again in a modern setting?)

Personally, the highlight of the day was the presentation of the semantic web challenge finalists. I was member of the jury, which meant that I had to review the submissions in advance and we had two very enjoyable discussions with the rest of the jury on the submissions. We had the first selection the day before, and this time all finalists gave their presentations and demos. And it was a tough task to choose (that is why we had such long discussions:-) because, well, the submissions were great overall. I do not really want to analyze each of the entries; I do not think it would be appropriate for me in this position. But the winner entry for the challenge, namely TrialX, really made a great impression on me. In short, the application is a consumer-centric tool through which patients can find matching clinical trials where they want to participate; it also helps those who organize those trials, etc. It is some sort of a matchmaking tool using all kinds of medical ontologies and vocabularies, public health record data and the like. We should realize the importance of this: here is a great Semantic Web application, winner of the challenge, which is really an application, not only demonstration, already deployed on the Web (soon as an iPhone app, too), and, to be a bit dramatic, may (and possibly has already) save lives. What else to we want as a proof that this technology is not only an academic exercise any more?

Fifth day

Only a partial day for me, as far as the conference goes, because I had to fly out before the end… But I could listen to the last keynote of the conference, ie, that of Nova Spivack.

Not surprisingly, Nova talked about Twine-2, a.k.a. T2. I did not really know what T2 was to be, I only heard that Twine, ie, T1, is moribund. As Nova acknowledged, it is too complicated, it is too hard for users to really figure it out; in fact, most of the users used it for search. Which is not the strongest feature of T1 in the first place.

So T2 is (well, will be) all about semantically backed search. It semantically indexes the Web, with an attempt to extract semantic information from the pages. The user interface would then be some sort of, essentially, faceted interface that would automatically classify the search hit results into different tabs; the user can use these tabs, drill down along other categories, etc. So far nothing radically new, though the user interface Nova showed was indeed very clean and nice. All this is done, internally, via vocabularies/ontologies, using RDF, RDFS, or OWL.

The interesting aspect of T2 (at least as far as I am concerned) is the incorporation of collective knowledge. First of all, T2 will include a system whereby users can add vocabularies that T2 will use in categorization. Users can get back those ontologies in OWL/RDF, they can improve them, etc. The other tool they will provide is a means to help semantically index pages that are, by themselves, not semantically annotated. This can be done via a Firefox extension; users can identify parts of the web pages (I presume, essentially, the DOM nodes) and associate these with classes of specific ontologies. The extension produces an XSLT transformation that can be sent back to the T2 system. Some social mechanism should of course be set up (eg, webmasters annotating their own pages should get a higher priority than third party annotators) but, essentially, it is some sort of a GRDDL transformation by proxy: T2 will have information on how to find transformation to semantically index specific pages without requiring the modification of the pages themselves (in contrast to GRDDL where such transformation is to be referred to from the page itself).

Of course, the system was a bit controversial in this community; indeed, it was not clear whether T2 would make use of the semantic information that do exist in pages already (microformats, RDFa, …) let alone the Linked Open Data information that is already out there. When asked, Nova did not seem to give a clear answer though, to be fair, he did not specifically say no and he also said that the semantic index might be put back to the public in the form of linked data. To be decided. It is also not fully clear whether those proxy-GRDDL transformations would be available for the community at large (hopefully the answer is yes…). It will be interesting to see how it plays out (T2 comes out in beta sometimes early 2010). Certainly a project to keep an eye on.

From a slightly more general point of view it is also interesting to note that two out of the three Semantic Challenge winners are also semantic search engines with different user interfaces (though sig.ma and VisiNav definitely do use the LOD cloud, no question there…). Definitely an area on the move!

I had the time and, frankly, the energy to really listen to only one more paper in the regular track, namely the paper on functions of RDF language elements, by Bernhard Schandl. A nice idea: imagine a traditional spreadsheet, where each cell is a collection of resources from an RDF Graph, or functions that can manipulate those resources (extract information, produce new set of resources, etc). Just like a spreadsheet, if you modify the underlying graph, ie, the resources in a cell, everything is automatically recalculated. Because, just like for a spreadsheet, a function can refer to the result of another function in another cell, one can do fairly complicated transformation and information extraction quite easily. Neat idea, to be tried out from their site.

That is it for ISWC2009. I obviously missed a lot of papers, partly because social life and hallway conversations sometimes had the upper hand, and sometimes simply because there were too many parallel sessions. But it was definitely an enriching week… See you all, hopefully, at ISWC2010, in Shanghai!

Posted in Semantic Web, Work Related Tagged: health care, OWL RL, RDFa, Rules, semantic search, SPARQL, spreadsheet

This year’s ISWC is held in Chantilly, Virginia. In a nice conference building in a beautiful park with autumn colours that, for reasons I do not really know, is always much more striking and amazing in America than in Europe. It is a bit of a pity that it is so far from Washington but, well, you can’t get it all…

First day: tutorials.

(For me, because there were also a bunch of workshops.) In the morning I was at the tutorial on how to consume Linked Open Data, by Juan Sequeda, Jammie Taylor, Patrick Sinclair, and Olaf Hartig; in the afternoon I went to the one on legal and social frameworks for sharing data on the Web, by Leigh Dodds, Jordan Hatcher, Tom Heath, and Kaitlin Thaney.

Juan and his  friends had actually a difficult task, and that became clear right at the start during the intro of Juan: part of the audience did not really know what LOD was all about, whereas there were also others who were, shall we say, old timers on the subject. I think the speakers did a really good job in navigating through these constraints, making short introductions to what LOD is all about but talking about issues and showing examples that were interesting for all of us. Kudos to that. Issues were raised by the audience that were really to-the-point (who should create sameAs,  links, how trustworthy are they, how to choose vocabularies and how they map to one another, etc) and, in his closing slides, Juan actually gave a list of the open  R&D issues in LOD. Worth looking at those (and no reason to repeat the list here…). B.t.w., the slides of the tutorial are on line.

One very interesting technology I heard about that, shame on me, but I did know was a tool based on a traversal based execution scheme for SPARQL called sqin.  Olaf did a presentation on that. What essentially happens is  as follows. At the beginning the default graph of the SPARQL query is empty. However, the system would systematically fetch RDF triples by dereferencing URI-s in the query pattern, adding those to the default graph. The query is matched against it, some variable will match thereby ‘adding’ new URIs to the pattern. And the process starts again, possibly yielding a complete solution (or more) to the original query. At the end of the process, solutions will be found on the Web, even if the system itself does not have any ‘real’ data behind it at the start. Of course, no one can secure that all solutions will be found, and you need some ’seed’ URI-s in the original query pattern, but it nevertheless looks like a very powerful tool to explore, say, the LOD.  Very interesting!

Then there were some examples on how LOD is used. Jammie talked about Freebase, and how Freebase is, in fact, a way for everybody to easily add information to the LOD (after all, Freebase works like a wiki, and all the data is reflected on the LOD).  He also had a very important message that is worth repeating (go to his slides for the rest): it takes very little effort to add a republishing capability to your triples store based application, thereby extending the general LOD. So… do it! This is how the system evolves…

Patrick described a quite geeky system that the BBC folks have developed (hopefully will become public soon): take the BBC’s musical data in RDF (which is available), plus the LOD cloud, plus… an IRC bot. What you get is an IRC channel which will pick up data on music, including the sound tracks, photos, etc, and display it on the machine. I presume you  can give orders and preferences through the IRC. Obviously a geeky stuff not for the masses:-) but shows what you can do…

The afternoon tutorial on the Legal and Social frameworks was of course very different. I think one of the many, but maybe the most important aspect of this tutorial is that… it took place! This may sound a bit strange but it is important for all our community to realize that we will have issues around copyright, licensing, waivers, etc, when it comes to the Web of Data, whether we like these issue or not. Tutorials like this, written notes and information, etc, are essential. Let us face it: most of us do not understand the details of the legal issues. So I was simply listening and trying to absorb what I heard…

I do not want to repeat the details of what I heard here; one thing I learned over the years is that I should leave legal argumentations and descriptions to those who really understand that. Ie, look at the slides. It is worth it. But just to show the complexities: I did not know or fully realize that there are major differences what can or cannot be copyrighted among countries: for example, a phone book cannot be copyrighted in the US or Europe, but can in Australia. That the seemingly simple notion of ‘attribution’ can, in fact, become an endless pit when it comes to data and the queries thereof (eg, if I have a filter in a query that results in data, should I give an attribution to the fact that were, in fact, filtered out?). Etc.

There is also a takeaway message for me (though it may be quite trivial) among the things I learned. Tom showed some practical examples on how can one add, say, licensing information to data by adding some RDF triples. However, for a larger data set the licensing may be different within the dataset. Eg, if you retrieve data from somewhere, and you enrich it with additional metadata, the metadata itself may have a different licensing (it is yours) than the data that you use (which may have its own licence). What this means is that when you organize your data internally, you should think about the licensing information you will add well in advance: organize your URI-s accordingly, for example. If you don’t, and you want to add license at the end, you might find yourself in trouble! Sounds like a simple message, but it is important. (Reminds me of what accessibility people always say: if you take accessibility issues into account right at the beginning when you build up a Web site, it is not complicated; but if you have to add accessibility features after the facts, it may become hell…)

By the way, Leigh has made a kind of an overview of the current ‘blobs’ on the LOD cloud to see whether any kind of licensing information is available or not. He has an overview of the results in his slides. The main fact is: the majority of data sets has no information whatsoever (or, at least, nothing that can be found in about 10 minutes)…

It was a good day. Looking forward to the rest.

Posted in Semantic Web, Work Related Tagged: creative commons, data commons, Linked Data, Linked Data Cloud, science commons, SPARQL

I ran into this two times in a week. I hope it is just a coincidence…

The story is simple. You find some service on the Web which looks nice and helpful. There are various options: you may take a minimal service, which is free of charge, or you can also choose extra services for a fee. It sounds like a decent choice: if the minimal service fits your needs, you are happy, if you need more, you pay something. I presume we all use services like that.

But then… if you take the free option, you may get a mail after 2-3 years’ of  usage saying that sorry, the free service is discontinued next month; you are welcome to upgrade for the paying service, otherwise, well, good bye. As I said I got this type of mail twice in a week: one from a service giving a minimal synchronization of my phone’s calendar with Google’s, the other providing a simple email certificate for signing my mails. On a matter of principle I will not upgrade; I do not find this approach really acceptable.

So… will Gmail, WordPress, or other similar services decide that they have attracted enough customers, they can now start charging? As I said, I hope this was just a coincidence and not some sort of a general direction…

Posted in General, Private, Social aspects, Work Related

OWL 2 has just been published as a Proposed Recommendation (yay!) which means, in laymen’s term, that the technical work is done, and it is up to the membership of W3C to accept it as a full blown Recommendation.

As I already blogged before, I did some implementation work on a specific piece of OWL 2, namely the OWL 2 RL Profile. (I have also blogged about OWL 2 RL and its importance before, nothing to repeat here.) The implementation itself is not really optimized, and it would probably not stand a chance for any large scale deployment (the reader may want to look at the OWL 2 implementation report for other alternatives).  But I can hope that the resulting service can be useful in getting a feel for what OWL 2 RL can give you: by just adding a few triples into the text box you can see what OWL 2 RL means. This is, by the way, an implementation of the OWL 2 RL rule set, which means that it can also accepts triples that are not mandated by the Direct Semantics of OWL 2 (a.k.a. OWL 2 DL). Put it another way, it is an implementation of a small portion of OWL 2 Full.

The core of my implementation turned out to be really easy straightforward: a forward chaining structure directly encoded in Python. I use RDFLib to handle the RDF triples and the triple store. Each triple in the RDF Graph is considered, compared to the premises of the rules; if there is a match then new triples are added to the Graph. (Well, most of the rules contain several triples to match with, and the usual approach is to pick one and explore the Graph deeper check against additional matches. Which one to pick is important, it may affect the overall speed, though.) If, through such a cycle, no additional triples are added to the Graph then we are done, the “deductive closure” of the Graph has been calculated. The rules of OWL 2 RL have been carefully chosen so that no new resources are added to the Graph (only new triples), ie, this process eventually stops.

The rules themselves are usually simple. Although it is possible and probably more efficient to encode the whole process using some sort of a rule engine (I know of implementations based on, eg, Jena’s rules or Jess), one can simply encode the rules using the usual conditional constructs of the programming language. The number of rules is relatively high but nothing that a good screen editor would not manage with copy-paste. There were only a few rules that required a somewhat more careful coding (usually to take care of lists) or many searches through the graph like, for examples, the rule for property chains (see rule prp-spo2 in the rule set). It is also important to note that the higher number of rules does really not affect the efficiency of the final system; if no triple matches a rule then, well, it just does not fire. No side effect of the mere existence of an unused rule.

So is it all easy and rosy? Not quite. First of all, this implementation is of course simplistic in so far as it generates all possible deducted triples that include a number of trivial triples (like ?x owl:sameAs ?x for all possible resources). That means that the resulting graph becomes fairly big even if the (optional) axiomatic triples are not added. If the OWL 2 RL process is bound to a query engine (eg, the new version of SPARQL will, hopefully, give a precise specification of what it means to have OWL 2 RL reasoning on the data set prior to a SPARQL query) then many of these trivial triples could be generated at query time only, thereby avoiding an extra load on the database. Well, that is one place where a proof-of-concept and simple implementation like mine looses against a more professional one:-)

The second issue was the contrast between RDF triples and “generalized” RDF triples, ie, triples where literals can appear in subject positions and bnodes can appear as properties. OWL 2 explicitly says that it works with generalized triples and the OWL 2 RL rule set also shows why that is necessary. Indeed, consider the following set of triples:

ex:X rdfs:subClassOf [
  a owl:Restriction;
  owl:onProperty [ owl:inverseOf ex:p ];
  owl:allValuesFrom ex:A
].

This is a fairly standard “idiom” even for simple ontologies; one wants to restrict, so to say, the subjects instead of the objects using an OWL property restriction. In other words that restriction combined with

ex:x rdf:type ex:X .
ex:y ex:p ex:x .

should yield

ex:y rdf:type ex:A .

Well, this deduction would not occur through the rule set if non-generalized RDF triples were used. Indeed, the inverse of ex:p is a blank node, ie, using it in a triple is not legal; but using that blank node to denote a property is necessary for the full chain of deductions. In other words, to get that deduction to work properly using RDF and rules, the author of the vocabulary would have to give an explicit URI to the inverse of ex:p. Possible, but slightly unnatural. If generalized triples are used, then the OWL 2 RL rules yield the proper result.

It turns out that, in my case, having bnodes as properties was not really an issue, because RDFLib could handle that directly (is that a bug in RDFLib?). But similar, though slightly more complex or even pathological examples can be constructed involving literals in subject positions, and that was a problem because RDFLib refused to handle those triples. What I had to do was to exchange all literals in the graph against a new bnode, perform all the deductions using those, and exchange the bnodes “back” against their original literals at the end. (This mechanism is not my invention; it is actually described by the RDF Semantics document, in the section on Datatype entailment rules.) B.t.w., the triples returned by the system are all “legal” triples, generalized triples play a role during the deduction only (and illegal triples are filtered out at output).

Literals with datatypes were also a source of problems. This is probably where I spent most of my implementation time (I must thank Michael Schneider who, while developing the test cases for OWL 2 RDF Based Semantics, was constantly pushing me to handle those damn datatypes properly…). Indeed, the underlying RDFLib system is fairly lax on checking the typed literals against their definition by the XSD specification (eg, issues like minimum or maximum values were not checked…). As a consequence, I had to re-implement the lexical to value conversion for all datatypes. Once I found out how to do that (I had dive a bit into the internals of RDFLib but, luckily, Python is an interpretative language…) it became a relatively straightforward, repetitive, and slightly time consuming work. Actually, using bnodes instead of “real” literals made it easier to implement datatype subsumptions, too (eg, the fact that, say, an xsd:byte is also a xsd:integer). This became important so that the rules would work properly on property restrictions involving datatypes.

Bottom line: even for a simple implementation literals, mainly literals with datatypes, are the biggest headache. The rest is really easy.  (This is hardly the discovery of the year, but is nevertheless good to remember…)

I was, actually, carried away a bit once I got a hold on how to handle datatypes, so I also implemented a small “extension” to OWL 2 RL by adding datatype restrictions (one of the really nice new features of OWL 2 but which is not mandated for OWL 2 RL). Imagine you have the following vocabulary item:

ex:RE a owl:Restriction ;
    owl:onProperty ex:p ;
    owl:someValuesFrom [
      a rdfs:Datatype ;
      owl:onDatatype xsd:integer ;
      owl:withRestrictions (
          [ xsd:minInclusive "1"^^xsd:integer ]
          [ xsd:maxInclusive "6"^^xsd:integer ]
      )
   ] .

which defines a restriction on the property ex:p so that some its values should be integers in the [1,6] interval. This means that

ex:q ex:p "2"^^xsd:integer.

yields

ex:q rdf:type ex:RE .

And this could be done by a slight extension of OWL 2 RL; no new rules, just adding the datatype restrictions to the datatypes. Nifty…

That is it. I had fun, and maybe it will be useful to others. The package can also be honLib-IH/dist/RDFClosure.tar.gz">honLib-IH/dist/RDFClosure.tar.gz">downloaded and used with RDFLib, by the way…

Posted in Python, Semantic Web, Work Related Tagged: Description logic, Knowledge Representation, OWL, OWL RL, Python, RDF, RDFLib, Resource Description Framework, SPARQL, w3c