- From: Ford, Kevin <kefo@loc.gov>
- Date: Wed, 17 Oct 2012 10:43:52 -0400
- To: "'Dan Scott'" <dscott@laurentian.ca>, "'kcoyle@kcoyle.net'" <kcoyle@kcoyle.net>, "'Jeff (OR) Young'" <jyoung@oclc.org>, "'public-schemabibex@w3.org'" <public-schemabibex@w3.org>
- Message-ID: <B8AF2ACE62D0FD4391BE28FEB91E1FF316CE2E81F7@LCXCLMB03.LCDS.LOC.GOV>
I think Dan's final point/question bears highlighting. The schema.org vocabulary represents a considerable simplification of, and broad view of, that which schema.org designers felt was important to *them*, and it was made intentionally broad and simple as to make it easier to implement and for the big seach engines to parse. That alone argues against narrow ontological arguments and complexity, as Jeff also surmised. Given the uptake of SKOS, and its W3C support, I suspect the schema.org designers/maintainers are more likely to embrace SKOS than an increasingly parsed view of KO information. So, we should keep asking ourselves, what is "reasonably" easy to implement *for the search engines.* Which brings me to another point: during the last call, there was mention of use cases and general questions like " Whom is this extension for? Who is the audience for this?" The answers, so far as I can tell, are "Search engines and search engines." If that is not the case (and I think that is OK), then we should identifiy the audience for this work. If it is plausible that the audience becomes more specific (than Google's, Yahoo's, and Bing's spiders), then we might consider a more granular approach to things, such as KO data, but only after identifying for whom we are making things more granular. The search engines have made a pretty clear statement to the contrary. Yours, Kevin p.s. The schema.org vocabulary "only covers the types of entities for which we (Microsoft, Yahoo!, Google and Yandex), think we can provide some special treatment for, through our search engines, in the near future" - I knew this of course, but reading it again in light of this work resulted in me asking myself "What does that statement mean with respect to the present effort?" Really. From: Dan Scott [mailto:dscott@laurentian.ca] Sent: Tuesday, October 16, 2012 11:56 PM To: kcoyle@kcoyle.net; Jeff (OR) Young; public-schemabibex@w3.org Subject: RE: Next Meeting - Schema Bib Extend W3C Group - 17th October 1) I can't make the call tomorrow, alas, due to prior commitments in that same time slot every week. 2) A concern I have about introducing more conceptual mappings into schema.org is that one of the goals of schema.org microdata, if I understand correctly, was to avoid the "spider food" SEO traps of ye olde META tags. Search engines ended up having to ignore META tags because they were mostly being used to game engines into returning irrelevant results. When one adds microdata around concrete terms (such as a standardized format for opening hours for an organization), there's an assertion being made that can be reasonably easily verified against the marked-up text that's visible in a browser; when one adds assertions about conceptual categories, those might be harder to verify & therefore trust and make use of. Perhaps less so for LCSH / MeSH and the like, because the subject headings will probably be visible in the browser, but I worry a little about something like SKOS if it's not tied to user-visible text. I could be paranoid, or cynical, or dumb (or all three!), but I would expect assertions along the lines of "BT: Nude Celebrity Pics" to start turning up in woeful SEO attempts. 3) I'm over-tired and definitely not at the height of my minimal intellectual powers, but I don't actually understand what Jeff means by a "big bang" approach (partially because I don't know whether I'm supposed to match up the indefinite pronoun "they" with schema.org or the Semantic Web community). In any case, Jeff included some useful links; http://schema.org/docs/datamodel.html contains the following statement: """ The type hierarchy presented on this site is not intended to be a 'global ontology' of the world. It only covers the types of entities for which we (Microsoft, Yahoo!, Google and Yandex), think we can provide some special treatment for, through our search engines, in the near future. """ So, along the lines of Phil Archer's post, "radically simple" appears to be the goal of schema.org... something that even I have been able to understand, put into practice, and see results (via rich snippets and the like). In that vein, then, is it a reasonable goal to take the incredible complexity and granularity of previous linked data / FRBR / FRAD efforts, and extract just the most immediately useful elements for the next iteration of schema.org in the bibliographic realm, with supporting use cases, such that libraries/archives/museums/publishers/book sellers/etc can reasonably implement it and that search engines can realistically make use of it? Thanks, Dan Scott <http://> <http://sch>><http://schema.o>><http://schema.or>><http://schema.org> <http://schema.org/> "<http://schema.org/do>Y<http://schema.org/docs/datamo>o<http://schema.org/docs/datamod>u<http://schema.org/docs/datamode>ng,Jeff (OR)" <jyoung@oclc.org<mailto:jyoung@oclc.org>> 10/16/2012 9:59 PM >>> Karen, You've raise issues that "we" (modeling experts and domain experts alike) need to deconstruct and reconcile to appeal to a broader audience. The W3C community (which I would characterize as a group of "modeling experts") was able to respond almost immediately (within days of the Schema.org announcement) with an RDF mapping: http://schema.rdfs.org/). The Schema.org community has acknowledged this alignment in a few ways: - http://schema.org/docs/datamodel.html - http://schema.org/docs/schemaorg.owl - Hiring Dan Brickley Schema.org may not be as obsessed with ontological/ideological prophylaxis as the Semantic Web (or library) community are, but that's presumably because they "know" (in a sense) that triple-based intuition-based vocabularies are more important (in the broad sense and in these early phases) than modeling-expertise/domain-expertise hygiene. I personally am inclined to believe that this "big bang" approach is in everyone's best interest. This includes emerging razor-sharp industrial-strength "domain models" of the type that could be characterized by the "long tail" diagram in this blog post: http://philarcher.org/diary/2012/danbri/?m=0 Disclaimer: This DOESN'T mean that I intend to shove this view down OCLC's (or the industry's) throat. (The idea that I have this kind of power/influence is so ludicrous I can't even believe I'm saying it. I'm merely here to help.) If the industry needs to believe that SKOS and/or MARCR (or Dublin Core or RDA or whatever) are "superior" to Schema.org then I will do whatever I can (within practical reason) to support it. Nevertheless, I can't promise to believe that it needs to be either/or. Jeff > -----Original Message----- > From: Karen Coyle [mailto:kcoyle@kcoyle.net] > Sent: Tuesday, October 16, 2012 4:55 PM > To: public-schemabibex@w3.org<mailto:public-schemabibex@w3.org> > Subject: Re: Next Meeting - Schema Bib Extend W3C Group - 17th October > > I'm unclear on the proposed use of a SKOS-like addition to schema.org. > My understanding of schema.org is that it is intended to make up for > the fact that most web pages do not have any semantic mark-up, just > HTML. > Any vocabularies in SKOS already do have semantic mark-up. Are you > anticipating that some sites would use schema.org *instead of* SKOS? > Also, are you thinking that schema.org users would use this to define > controlled vocabularies? > > kc > > On 10/16/12 12:33 PM, jean delahousse wrote: > > > > > > 2012/10/16 jean delahousse <delahousse.jean@gmail.com <mailto:delahousse.jean@gmail.com%0b>> > <mailto:delahousse.jean@gmail.com>> > > > > Hi All, > > > > Any subject described in the classes of Schema.org (person, > > organization, creative work, product, intangible...) can be > > referenced in a controlled vocabulary with the specific formalism > of > > a controlled vocabulary and the specific relationships it allows > > (broader, narrower, related and alignment relationships). The aim > of > > the controlled vocabulary description of the concept, is not to > > descripe the subject itself, but to describe it as a "concept" in > a > > controlled vocabulary, describing a given set of subjects, which > > someone wants to use to classify or describe something else. > > > > Administration and publication of a controlled vocabulary implies > > some best practices and rules about concept definition, labels, > life > > of the concept.. which are the same for all class of subjects. > > > > So in my mind the concept description can double a description of > > the subject as a person, a product, an event, in the same > > publication.. This would probably means to be able to describe a > link > > between the skos:concept page and the page about this subject, > for > > example between the concept of "Chicago" in the controlled > > vocabulary, and a page describing Chicago as a place in the same > web > > site. > > > > So the objective here is not to describe subject which are not > > actually described by Shema.org but to give a very oriented > > "taxonomist" view on any possible subject. > > As a matter of fact, my first tought was to add a class "concept" > at > > the upper level under Thing. This class would not have be > exclusive > > from belonging to any other class. > > > > Jean > > > > > > 2012/10/16 jean delahousse <delahousse.jean@gmail.com <mailto:delahousse.jean@gmail.com%0b>> > <mailto:delahousse.jean@gmail.com>> > > > > Richard, > > > > I did publish a brief description a link on the uploaded copy > of > > the document. > > > > Jean > > > > > > 2012/10/16 Richard Wallis <richard.wallis@oclc.org <mailto:richard.wallis@oclc.org%0b>> > <mailto:richard.wallis@oclc.org>> > > > > Jean, > > > > Would you have any objection to your proposal being > > published on the Group Wiki? > > > > If it is OK by you, you could append a brief description > to > > the Areas for Discussion page > > > <http://www.w3.org/community/schemabibex/wiki/Areas_for_Discussion> > > and link to an uploaded copy of the document. > > > > ~Richard. > > > > > > > > > > On 16/10/2012 14:10, "Tami Ezra" > > <Tami.Ezra@exlibrisgroup.com <mailto:Tami.Ezra@exlibrisgroup.com%0b>> > <http://Tami.Ezra@exlibrisgroup.com>> wrote: > > > > Hi, > > > > My name is Tami Ezra and I am a senior business > analyst > > at Ex Libris. > > > > I am interested in the proposal discussed below - > would > > it be possible to get a copy? > > > > Many thanks > > > > Tami > > > > > > > > *From:* Young,Jeff (OR) [mailto:jyoung@oclc.org] > > *Sent:* Tuesday, October 16, 2012 3:24 AM > > *To:* jean delahousse KC; public-schemabibex@w3.org<mailto:public-schemabibex@w3.org> > > <http://public-schemabibex@w3.org> > > *Subject:* RE: Next Meeting - Schema Bib Extend W3C > > Group - 17th October > > > > Jean, > > > > I like where this is heading. In the experimental > > WorldCat.org Linked Data so far (online RDFa and bulk > > N-Triples) I used skos:Concept for these situations. > In > > my dev environment, though, I started the switch to > > schema:Intangible but wasn't entirely happy with it. > > This proposal is much more satisfying. > > > > One issue comes to mind for discussion, though. This > > proposed schema:Concept feels more equivalent to FRBR > > Concept than it does to skos:Concept. The difference > is > > subtle but real, IMO, and has to do with foaf:focus > > (with a range of "Thing" and inverse of > > madsrdf:isIdentifiedByAuthority) being a meaningful > > property for the latter (skos:Concept) but not the > > former (FRBR Concept). VIAF (which doesn't currently > > attempt to identify FRBR Concepts) is probably the > best > > illustration of the issues involved. > > > > I realize that schema:Concept is destined to be a > > compromise, but it would be nice (albeit perhaps not > > necessary) if this group had a clear understanding > and > > articulation of those compromises to minimize > confusion > > in industrial-strength use cases. > > > > Jeff > > > > > > *From:* delahousse.jean@gmail.com<mailto:delahousse.jean@gmail.com> > > <http://delahousse.jean@gmail.com> > > [mailto:delahousse.jean@gmail.com] *On Behalf Of > *jean > > delahousse KC > > *Sent:* Monday, October 15, 2012 10:13 AM > > *To:* public-schemabibex@w3.org<mailto:public-schemabibex@w3.org> > > <http://public-schemabibex@w3.org> > > *Subject:* Re: Next Meeting - Schema Bib Extend W3C > > Group - 17th October > > > > Hi all, > > > > > > > > First I want to thank you for accepting my > application > > to participate to your work group. > > > > > > > > I had been working this summer on an extension of > > Schema.org for controlled vocabularies based on Skos > > ontology. After BnF published Rameau in the LOD but > also > > as web pages, one for each concept, I thought it will > be > > useful to have an extension of Schema.org to make > > concepts defined in controlled vocabularies more > visible > > by search engines. > > > > Concepts are good candidates for TopicPages, and work > as > > hub to access well annotated contents or others Topic > > Pages. They are a valuable asset for content / > knowledge > > access from a search engine. > > > > > > > > Also it happens to find "glossary", "terminology" or > > "lexicon" in a web site. This extension of Schema.org > > will enable to describe those types of publication. > > > > > > > > I took the initiative of this work but immediately > ask > > for support and review work to Antoine Isaac and > Romain > > Weinz. They have been very encouraging and already > > proposed corrections included in this version. > > > > > > > > You'll find attached the proposal for the Skos > > Schema.org extension, we made it as simple and light > as > > possible. > > > > > > > > I propose, if the group agrees, to have a first > > discussion on this proposal inside our group before > to > > publish it for a larger audience.. > > > > > > > > Best regards > > > > > > > > Talk to you on Thursday. > > > > > > > > Jean Delahousse > > > > 2012/10/10 Richard Wallis <richard.wallis@oclc.org <mailto:richard.wallis@oclc.org%0b>> > <http://richard.wallis@oclc.org>> > > > > Hi All, > > > > It is about time we followed up on the excellent > first > > meeting we had. > > > > I have scheduled conference call time for 11:00am EDT > > next Wednesday 17th October for us to start to talk > > through some of the issues and suggestions we > discussed > > last time. > > > > You will find call in details and a provisional > agenda > > on the group wiki here: > > > http://www.w3.org/community/schemabibex/wiki/Meet_20121017 > > > > If you have suggestions for the agenda, either edit > the > > wiki or drop me a line. > > > > Regards, > > Richard. > > > > -- > > Richard Wallis > > Technology Evangelist > > OCLC > > > > > > > > > > > > > > -- > > ------------------------------------------------------------- > --------------------------------------------------------- > > delahousse.jean@gmail.com<mailto:delahousse.jean@gmail.com> <mailto:delahousse.jean@gmail.com> > - > > +33 6 01 22 48 55 <tel:%2B33%206%2001%2022%2048%2055> - > skype: > > jean.delahousse -blog >contenus >données >sémantique > > <http://jean-delahousse.net> - twitter.com/jdelahousse > > <http://twitter.com/jdelahousse> > > > > > > > > > > > > -- > > ----------------------------------------------------------------- > ----------------------------------------------------- > > delahousse.jean@gmail.com<mailto:delahousse.jean@gmail.com> <mailto:delahousse.jean@gmail.com> - > +33 6 > > 01 22 48 55 <tel:%2B33%206%2001%2022%2048%2055> - skype: > > jean.delahousse -blog >contenus >données >sémantique > > <http://jean-delahousse.net> - twitter.com/jdelahousse > > <http://twitter.com/jdelahousse> > > > > > > > > > > > > -- > > --------------------------------------------------------------------- > ------------------------------------------------- > > delahousse.jean@gmail.com<mailto:delahousse.jean@gmail.com> <mailto:delahousse.jean@gmail.com> - +33 6 > 01 > > 22 48 55 - skype: jean.delahousse -blog >contenus >données > >sémantique > > <http://jean-delahousse.net> - twitter.com/jdelahousse > > <http://twitter.com/jdelahousse> > > > > > > -- > Karen Coyle > kcoyle@kcoyle.net<mailto:kcoyle@kcoyle.net> http://kcoyle.net > ph: 1-510-540-7596 > m: 1-510-435-8234 > skype: kcoylenet >
Received on Wednesday, 17 October 2012 14:46:28 UTC