Re: how to consume linked data
Kingsley All: OK, it took a few weeks to put my thoughts together, but I think you'll enjoy reading this new post called Linked Data: Interpretants and Interpretation. http://phaneron.rickmurphy.org/?p=36 See below for additional comments. Kingsley Idehen wrote: rick wrote: Danny All: Of course there's been a lot of work on this subject over the years. You can find one nice piece on analogical reasoning here [1]. But for linked data to become useful it's important to refine our understanding of web architecture beginning with the language of resources. I'm currently working up a piece for my blog on this topic. Stay tuned. Amen re. resources! Amen. information and non-information resource terminology is only second to RDF/XML re. historic impediments to comprehending essence of Semantic Web Project vision. I agree that resource terminology remains an impediment even after browsing the TAG discussions and understanding the background of some key W3C recommendations. I hope Interpretants and Interpretation furthers the long standing discussion of resources and URIs in W3C. If by the essence of the semantic web project vision you mean the 2003 Scientific American article, you'll find Interpretants and Interpretation speaks directly to an updated RDF model theory to support the 2003 article. I don't intend the suggestions in Interpretants and Interpretation as a criticism. It took me more than a few years of careful research and a lot of help from the very smart folks who got us where we are today to understand these issues. That being said, I have recommended that the US and UK governments engage in creating a Linked Data roadmap. One essential element of that road map would include a lifecycle for how linked data publishers and linked data consumers could work towards the goals of linked data without the oversight of a directed authority. 1. http://www.jfsowa.com/pubs/analog.htm -- Rick cell: 703-201-9129 web: http://www.rickmurphy.org blog: http://phaneron.rickmurphy.org
Re: how to consume linked data
I find this answer valuable, but unsatisfying. To me this is the fundamental weak spot in the whole chain of semantic web/linked data. I do appreciate the tremendous flexibility, generality, simplicity, novelty, and cool factor in the semantic web/linked data frameworks. But having done everything you can with that, for effective interoperability people doing similar things (i.e., making similar resources) will need to build and label them in known compatible ways. I think it is entirely analogous to folksonomy searching (e.g., Google searches of free text, more or less) vs Controlled vocabulary searching (e.g, using metadata standards with controlled vocabularies). At scale, the former will stay in the lead and be increasingly powerful; but the latter will always be necessary for more deterministic, consistent, and targeted results. Well, at least until computers are Really, Really smart. John On Sep 26, 2009, at 3:08 AM, Olaf Hartig wrote: Hey Danny, On Friday 25 September 2009 22:51:37 Danny Ayers wrote: 2009/9/25 Juan Sequeda juanfeder...@gmail.com: Linked Data is out there. Now it's time to develop smart (personalized) software agents to consume the data and give it back to humans. I don't disagree, but I do think the necessary agents aren't smart, just stupid bots (aka Web services a la Fielding). These stupid bots are able to discover and make use of data from a wide variety of sources on the Web. I'm still convinced this allows applications of an interesting novelty. And let's not forget, these applications enable users to retain full control over the authoritative source of data provided by them. This is a big step. It is more a question of why so little of these applications came up yet. I agree with Kjetil here. Tools are missing that bring developers (who don't know all the technical details) on board. One possible approach to this is: try also using SQUIN (www.squin.org) Thanks, not seen before. ... which is a query service (currently still in pre-alpha) that is based on the functionality of the SemWeb Client Lib. An application simply sends a SPARQL query. This query is executed over the Web of Linked Data using the link traversal query execution approach as implemented in the SemWeb Client Lib. The result is returned to the app which may visualize or process it. Hence, the app developer does not need to bother with traversing RDF links, RDF/XML vs. RDFa, etc. Another important issue in consuming LD is the filtering of data as you mention in your original question. Indeed, we need approaches of filtering automatically during the discovery of data. Unfortunately, for many filter criteria (e.g. reliability, timeliness, trustworthiness) we do not even now very well how we may filter automatically given we have the data. Greetings, Olaf --- John Graybeal Marine Metadata Interoperability Project: http://marinemetadata.org grayb...@marinemetadata.org
Re: how to consume linked data
Hi Kjetil -- You wrote... *I think there is a critical piece of technology that is missing in our arsenal, namely a (free software) programming stack that makes a large group of developers, who are likely to have little prior understanding of semweb, to go yeah, I can do that.* How about being more ambitious? In the above, change a large group of * developers* to a large group of *non-programmers*. That would get you Social Media Meets Linked Data. Here's step in that direction www.reengineeringllc.com/demo_agents/RDFQueryLangComparison1.agent There's also a short paper www.reengineeringllc.com/A_Wiki_for_Business_Rules_in_Open_Vocabulary_Executable_English.pdf and the technology is online at the same site. Cheers, -- Adrian Internet Business Logic A Wiki and SOA Endpoint for Executable Open Vocabulary English over SQL and RDF Online at www.reengineeringllc.comShared use is free Adrian Walker Reengineering On Fri, Sep 25, 2009 at 6:53 AM, Kjetil Kjernsmo kje...@kjernsmo.netwrote: On Friday 25. September 2009 10:15:34 you wrote: sorry if I sound negative, I reckon the semweb is a done deal now, the many-eyeballs arrived. Thanks for asking the right questions, Danny, I believe it is critical for the success that someone does! but - where should we take it? What I'd like to do with it, is to solve problems for people when combining data sets that are cannot be solved by conventional means, i.e. today the number of people who are interested in a particular combination of datasets goes down whereas the cost generally goes up, so it doesn't scale. I think there is a critical piece of technology that is missing in our arsenal, namely a (free software) programming stack that makes a large group of developers, who are likely to have little prior understanding of semweb, to go yeah, I can do that. I think the work done by the Drupal folks is a right step in this direction, for the kind of stuff that people use a CMS for. But I think that we also need a stack, probably built around the MVC pattern, that can be used for more generic purposes. I haven't got anywhere with my ideas on this topic though... Kjetil
Re: how to consume linked data
Hey Danny, On Friday 25 September 2009 22:51:37 Danny Ayers wrote: 2009/9/25 Juan Sequeda juanfeder...@gmail.com: Linked Data is out there. Now it's time to develop smart (personalized) software agents to consume the data and give it back to humans. I don't disagree, but I do think the necessary agents aren't smart, just stupid bots (aka Web services a la Fielding). These stupid bots are able to discover and make use of data from a wide variety of sources on the Web. I'm still convinced this allows applications of an interesting novelty. And let's not forget, these applications enable users to retain full control over the authoritative source of data provided by them. This is a big step. It is more a question of why so little of these applications came up yet. I agree with Kjetil here. Tools are missing that bring developers (who don't know all the technical details) on board. One possible approach to this is: try also using SQUIN (www.squin.org) Thanks, not seen before. ... which is a query service (currently still in pre-alpha) that is based on the functionality of the SemWeb Client Lib. An application simply sends a SPARQL query. This query is executed over the Web of Linked Data using the link traversal query execution approach as implemented in the SemWeb Client Lib. The result is returned to the app which may visualize or process it. Hence, the app developer does not need to bother with traversing RDF links, RDF/XML vs. RDFa, etc. Another important issue in consuming LD is the filtering of data as you mention in your original question. Indeed, we need approaches of filtering automatically during the discovery of data. Unfortunately, for many filter criteria (e.g. reliability, timeliness, trustworthiness) we do not even now very well how we may filter automatically given we have the data. Greetings, Olaf
Re: how to consume linked data
Dan Brickley wrote: This doc-typing idiom never got heavily used in FOAF, beyond the type PersonalProfileDocument, which FOAF defines. Mostly we just linked FOAF files together (initially with seeAlso and IFPs, lately using URIs more explicitly). I think there are many other reasons why characterising typical RDF document patterns makes sense, related to the frustration of dealing with documents when all you know is they have triples in them. We don't have good mechanisms for doing so yet, ie. for characterising these higher level patterns. But various folk are heading in same direction, some using SPARQL, others OWL or XForms, or DC Application Profile definitions Without some hints about what we're pointing at with our links, crawlers don't have much to go on. Merely knowing that the information at the other end of the link is more RDF, or that it describes a thing of a certain type, might not be enough. There are a lot of things you might want to know about a person, or a place, and at many different levels of detail. For apps eg running in a mobile/handheld environment, they can't afford to speculatively download everything.. Interesting... I'm doing work at the moment with CIDOC-CRM (http://cidoc.ics.forth.gr/) and its expression in OWL (http://www8.informatik.uni-erlangen.de/IMMD8/Services/cidoc-crm/versions.html). Something I've noticed is that the extension/refinement mechanism provided by CIDOC-CRM is based on what they call Types (though I think it's more like skos:Concept), so that the core properties tend be be very predictable. There are some areas where I've used new properties to capture finer-grained information, but they tend to be at the margins (e.g. putting numeric values on date-ranges) rather than in the core (e.g. this object was made in this time period). Maybe there's scope for using SKOS in a doc-typing idiom? #g
Re: how to consume linked data
Many thanks for responses, stuff to think about. Yihong got to /root of my question, ...miss the main purpose why we want to have data linked in the first place why are places like itsy, youtube and redtube (yup, pr0n still lives) more compelling, given what we know? people *are* getting the data out, but it seems to me there's a gap between that and stuff that actually improves people's quality of life. sorry if I sound negative, I reckon the semweb is a done deal now, the many-eyeballs arrived. but - where should we take it? -- http://danny.ayers.name
Re: how to consume linked data
On Friday 25. September 2009 10:15:34 you wrote: sorry if I sound negative, I reckon the semweb is a done deal now, the many-eyeballs arrived. Thanks for asking the right questions, Danny, I believe it is critical for the success that someone does! but - where should we take it? What I'd like to do with it, is to solve problems for people when combining data sets that are cannot be solved by conventional means, i.e. today the number of people who are interested in a particular combination of datasets goes down whereas the cost generally goes up, so it doesn't scale. I think there is a critical piece of technology that is missing in our arsenal, namely a (free software) programming stack that makes a large group of developers, who are likely to have little prior understanding of semweb, to go yeah, I can do that. I think the work done by the Drupal folks is a right step in this direction, for the kind of stuff that people use a CMS for. But I think that we also need a stack, probably built around the MVC pattern, that can be used for more generic purposes. I haven't got anywhere with my ideas on this topic though... Kjetil
Re: how to consume linked data
Linked Data is out there. Now it's time to develop smart (personalized) software agents to consume the data and give it back to humans. try also using SQUIN (www.squin.org) Juan Sequeda, Ph.D Student Dept. of Computer Sciences The University of Texas at Austin www.juansequeda.com www.semanticwebaustin.org On Fri, Sep 25, 2009 at 11:56 AM, Leo Sauermann leo.sauerm...@dfki.dewrote: Uh, I thought the answer to danny's question is semwebclient by Olaf Hartig and others. http://www4.wiwiss.fu-berlin.de/bizer/ng4j/semwebclient/ In general, I thought that Olaf Hartig would be the first contact for such things... best Leo It was Danny Ayers who said at the right time 24.09.2009 09:59 the following words: The human reading online texts has a fair idea of what is and what isn't relevant, but how does this work for the Web of data? Should we have tools to just suck in any nearby triples, drop them into a model, assume that there's enough space for the irrelevant stuff, filter later? How do we do (in software) things like directed search without the human agent? I'm sure we can get to the point of - analogy - looking stuff up in Wikipedia picking relevant links, but we don't seem to have the user stories for the bits linked data enables. Or am I just imagination-challenged? Cheers, Danny. -- _ Dr. Leo Sauermann http://www.dfki.de/~sauermannhttp://www.dfki.de/%7Esauermann Deutsches Forschungszentrum fuer Kuenstliche Intelligenz DFKI GmbH Trippstadter Strasse 122 P.O. Box 2080 Fon: +43 6991 gnowsis D-67663 Kaiserslautern Fax: +49 631 20575-102 Germany Mail: leo.sauerm...@dfki.de Geschaeftsfuehrung: Prof.Dr.Dr.h.c.mult. Wolfgang Wahlster (Vorsitzender) Dr. Walter Olthoff Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes Amtsgericht Kaiserslautern, HRB 2313 _
Re: how to consume linked data
2009/9/25 Kjetil Kjernsmo kje...@kjernsmo.net: On Friday 25. September 2009 10:15:34 you wrote: sorry if I sound negative, I reckon the semweb is a done deal now, the many-eyeballs arrived. Thanks for asking the right questions, Danny, I believe it is critical for the success that someone does! Thanks, but I'm not even sure they are the right questions. but - where should we take it? What I'd like to do with it, is to solve problems for people when combining data sets that are cannot be solved by conventional means, i.e. today the number of people who are interested in a particular combination of datasets goes down whereas the cost generally goes up, so it doesn't scale. Yes, but please bear with me now - do we have to wait for another generation arriving on the Web? There must be ways we can kick-start this stuff. I think there is a critical piece of technology that is missing in our arsenal, namely a (free software) programming stack that makes a large group of developers, who are likely to have little prior understanding of semweb, to go yeah, I can do that. Like bengee's ARC2 stack - PHP? I think the work done by the Drupal folks is a right step in this direction, for the kind of stuff that people use a CMS for. But I think that we also need a stack, probably built around the MVC pattern, that can be used for more generic purposes. Absolutely. If we can re-use existing patterns we can get people involved. I haven't got anywhere with my ideas on this topic though... Me neither :) I should insert a Star Trek quote here, but can't think of one. Cheers, Danny. -- http://danny.ayers.name
Re: how to consume linked data
2009/9/25 Juan Sequeda juanfeder...@gmail.com: Linked Data is out there. Now it's time to develop smart (personalized) software agents to consume the data and give it back to humans. I don't disagree, but I do think the necessary agents aren't smart, just stupid bots (aka Web services a la Fielding). try also using SQUIN (www.squin.org) Thanks, not seen before. -- http://danny.ayers.name
Re: how to consume linked data
Olaf, comments? 2009/9/25 Leo Sauermann leo.sauerm...@dfki.de: Uh, I thought the answer to danny's question is semwebclient by Olaf Hartig and others. http://www4.wiwiss.fu-berlin.de/bizer/ng4j/semwebclient/ In general, I thought that Olaf Hartig would be the first contact for such things... best Leo It was Danny Ayers who said at the right time 24.09.2009 09:59 the following words: The human reading online texts has a fair idea of what is and what isn't relevant, but how does this work for the Web of data? Should we have tools to just suck in any nearby triples, drop them into a model, assume that there's enough space for the irrelevant stuff, filter later? How do we do (in software) things like directed search without the human agent? I'm sure we can get to the point of - analogy - looking stuff up in Wikipedia picking relevant links, but we don't seem to have the user stories for the bits linked data enables. Or am I just imagination-challenged? Cheers, Danny. -- _ Dr. Leo Sauermann http://www.dfki.de/~sauermann Deutsches Forschungszentrum fuer Kuenstliche Intelligenz DFKI GmbH Trippstadter Strasse 122 P.O. Box 2080 Fon: +43 6991 gnowsis D-67663 Kaiserslautern Fax: +49 631 20575-102 Germany Mail: leo.sauerm...@dfki.de Geschaeftsfuehrung: Prof.Dr.Dr.h.c.mult. Wolfgang Wahlster (Vorsitzender) Dr. Walter Olthoff Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes Amtsgericht Kaiserslautern, HRB 2313 _ -- http://danny.ayers.name
Re: how to consume linked data
On 25 Sep 2009, at 07:41, Graham Klyne wrote: Interesting... I'm doing work at the moment with CIDOC-CRM (http:// cidoc.ics.forth.gr/) and its expression in OWL (http:// www8.informatik.uni-erlangen.de/IMMD8/Services/cidoc-crm/ versions.html). Have you seen Simon Reinhardt's recent OWL2 version? http://bloody-byte.net/rdf/cidoc-crm/index.html -- Toby A Inkster mailto:m...@tobyinkster.co.uk http://tobyinkster.co.uk
Re: how to consume linked data
Danny All: Of course there's been a lot of work on this subject over the years. You can find one nice piece on analogical reasoning here [1]. But for linked data to become useful it's important to refine our understanding of web architecture beginning with the language of resources. I'm currently working up a piece for my blog on this topic. Stay tuned. That being said, I have recommended that the US and UK governments engage in creating a Linked Data roadmap. One essential element of that road map would include a lifecycle for how linked data publishers and linked data consumers could work towards the goals of linked data without the oversight of a directed authority. 1. http://www.jfsowa.com/pubs/analog.htm -- Rick cell: 703-201-9129 web: http://www.rickmurphy.org blog: http://phaneron.rickmurphy.org Danny Ayers wrote: The human reading online texts has a fair idea of what is and what isn't relevant, but how does this work for the Web of data? Should we have tools to just suck in any nearby triples, drop them into a model, assume that there's enough space for the irrelevant stuff, filter later? How do we do (in software) things like directed search without the human agent? I'm sure we can get to the point of - analogy - looking stuff up in Wikipedia picking relevant links, but we don't seem to have the user stories for the bits linked data enables. Or am I just imagination-challenged? Cheers, Danny.
Re: how to consume linked data
Danny Ayers wrote: The human reading online texts has a fair idea of what is and what isn't relevant, but how does this work for the Web of data? Should we have tools to just suck in any nearby triples, drop them into a model, assume that there's enough space for the irrelevant stuff, filter later? How do we do (in software) things like directed search without the human agent? I'm sure we can get to the point of - analogy - looking stuff up in Wikipedia picking relevant links, but we don't seem to have the user stories for the bits linked data enables. Or am I just imagination-challenged? Cheers, Danny. I think users have to discover, comprehend, and then exploit (consume or extend the reference chain). This is the vital sequence. fwiw, here is how I tell the story to general observers: Today, you put a resource URL in your browser and get either of the following: - Rendered Page - Markup behind the Page Linked Data simply adds the ability to see a resource description (metadata). The description takes honors the Web core architecture by providing links for each component of the description. That's it. All the other smart stuff simply happens behind the scenes and shows up in the resource description. -- Regards, Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen President CEO OpenLink Software Web: http://www.openlinksw.com