Re: LOD publishing question

Hi all

With due respect to the great work of Sindice and similar projects, seems
to me they belong to Semantic Web history.
As long as we have to go to a specific Semantic Web engine like
Swoogle<http://swoogle.umbc.edu/>
 or Sindice <http://sindice.com/> to search a URI for some-thing, the
Semantic Web will not be a natural part of the Web. Getting URIs of things
as part of a regular search from the major search engine would be a
significant milestone ... (copying from [1] written one year ago)

We are almost there. The Semantic Web search is now in the hands on the big
guys, Google first in line with the Knowledge Graph, and Freebase API. Try
[1], and nothing prevents from feeding Freebase with more URIs.
Conversations on Google Semantic Search community are worth following.

Like it or not, but as Giovanni writes, it's now a billion-dollar story,
where only big guys can afford to play and be efficient.

[1] http://bvatant.blogspot.fr/2013/02/a-small-step-for-google.html
[2]
https://www.googleapis.com/freebase/v1/search?&query=%22Cooper%22&type=film/actor&output=%28%20/common/topic/topic_equivalent_webpage%29
[3] https://plus.google.com/communities/103079475562443214899

Bernard


2014-01-31 Giovanni Tummarello <g.tummarello@gmail.com>:

> Thanks Hugh,
>
> crawling the web accurately is a billion dollar thing nowadays (not my
> words) and all the big guys accurately crawl all the metadata now (though
> dont give any public api).
>
> I still think a more "focused" version of sindice e.g. just on demand etc
> might be useful and have impact but resources are necessarely limited
>
> Announcements with respect to the rest are coming next week :)
> have a good weekend and thanks for the thanks, appreciated.
>
> Gio
>
>
> On Thu, Jan 30, 2014 at 8:13 PM, Hugh Glaser <hugh@glasers.org> wrote:
>
>> Hi Giovanni,
>> Thank you for the update.
>> I am sorry to hear that Sindice is going into a frozen state, and that
>> circumstances are making that happen, but of course pleased that you are
>> able to keep it going at all.
>> I send you and your team my personal thanks for the service you have
>> provided over the last 5 or so years, and wish you all well.
>> Very best
>> Hugh.
>>
>>
>> On 28 Jan 2014, at 14:19, Giovanni Tummarello <g.tummarello@gmail.com>
>> wrote:
>>
>> > With respect to Sindice
>> >
>> > for a number of reasons, the people who originally created it, the
>> former Data Intensive Infrastructure group, are either not working in the
>> original institution hosting it, National University of Ireland Galway,
>> institute formerly known as DERI or have been assigned to other tasks.
>> >
>> > Sindice has been operating for 5+ years, updating its index, (though we
>> were never perfect) and we believe supported a lot of works on the field,
>>  but its now time to move on.  In the meanwhile the project will continue
>> answer queries but without updating its index.
>> >
>> > Apologies for the inconvenience of course, we'll be posting on this
>> soon and update the homepage to reflect the change.
>> >
>> > Giovanni
>> >
>> >
>> >
>> > On Tue, Jan 28, 2014 at 11:27 AM, Hugh Glaser <hugh@glasers.org> wrote:
>> > Good question.
>> > I'll report what I found, rather than advising.
>> >
>> > So I went there when you published that email, looking for stuff to put
>> in my sameas.org site.
>> > I tried exploring, and when I went to Browse I only found a few things,
>> so wasn't encouraged :-)
>> > (And, as an aside, Advanced Search didn't seem to do anything, and the
>> search links at the bottom were not links.)
>> > So I decided that it wasn't really mature enough to make it worth the
>> effort (yet?), even though there should be massive scope for linkage
>> eventually.
>> >
>> > But the real problem was that I couldn't find any Linked Data, or even
>> an RDF store.
>> > The URIs you use are not very Cool URIs, and I tried to see if there
>> was RDF at the end of them by doing Content Negotiation, but there wasn't.
>> > I am thinking of things like
>> http://tundra.csd.sc.edu/rol/view-person.php?id=291
>> >
>> > So I went away :-)
>> >
>> > For people like me, you could put something about how to see the RDF in
>> an About page (or if it is there, make it easier to find). You only get one
>> chance to snare people on the web, after all.
>> > Of course as Alfredo says, for spidering search engines, and it would
>> have helped me too, you need robots.txt (which I couldn't find either),
>> sitemap, sitemap.xml, voiD description.
>> >
>> > Good luck!
>> > Hugh
>> >
>> > On 28 Jan 2014, at 04:12, WILDER, COLIN <WILDERCF@mailbox.sc.edu>
>> wrote:
>> >
>> > > Another question to you very helpful people-
>> > >
>> > > <and apologies again for semi cross-posting>
>> > >
>> > > Our LOD working group is having trouble publishing our data (see
>> email below) in RDF form. Our programmer, a master's student, who is
>> working under the supervision of myself and a computer science professor,
>> has mapped sample data into RDF, has the triplestore on a D2RQ server
>> (software) on our server and has set up a SPARQL end-point on the latter.
>> But he has been unsuccessful so far getting 3 candidate semantic web search
>> engines (Falcons, Swoogle and Sindice) to  be able to find our data when he
>> puts a test query in to them. He has tried communicating with the people
>> who run these, but to little avail. Any suggestions about sources of
>> information, pointers, best practices for this actual process of publishing
>> LOD? Or, if you know of problems with any of those three search engines and
>> would suggest a different candidate, that would be great too.
>> > >
>> > > Thanks again,
>> > >
>> > > Colin Wilder
>> > >
>> > >
>> > > From: WILDER, COLIN [mailto:WILDERCF@mailbox.sc.edu]
>> > > Sent: Thursday, January 16, 2014 11:51 AM
>> > > To: 'public-lod@w3.org'
>> > > Subject: LOD for historical humanities information about people and
>> texts
>> > >
>> > > To the many people who have kindly responded to my recent email:
>> > >
>> > > Thanks for your suggestions and clarifying questions. To explain a
>> bit better, we have a data curation platform called RL, which is a large,
>> complex web-based MySQL database designed for users to be able to simply
>> input, store and share data about social and textual networks with each
>> other, or to share it globally in RL's data commons. The data involved are
>> individual data items, such as info about one person's name, age, a book
>> title, a specific social relationship, etc. The entity types (in the
>> ordinary-language sense of actors and objects, not in the database tabular
>> sense) can be seen athttp://tundra.csd.sc.edu/rol/browse.php. The data
>> commons in RL is basically a subset of user data that users have elected
>> (irrevocably) to share with all other users of the system. NB there is a
>> lot of dummy data in the data commons right now because of testing.
>> > >
>> > > We are designing an expansion of RL's functionality so as to publish
>> data from the data commons as LOD, so I am doing some preliminary work to
>> assess feasibility and fit by matching up our entity types with
>> RDFvocabularies. Here is what I have so far. First are the entity(ies) and
>> relationships, followed by the appropriate vocabularies:
>> > >
>> > > 1.       Persons, social relations: FOAF, BIO. The "Catalogus
>> Professorum Lipsiensis" or CPL(
>> http://svn.aksw.org/papers/2010/ISWC_CP/public.pdf) looks enormously
>> useful for connecting academics (people), their relations and their books.
>>  But, I cannot seem to get any info page or specification page to load,
>> making me worry that it's dead.
>> > > 2.       Membership in organizations: ORG
>> > > 3.       Enrollment in an academic course (e.g. a lecture course):
>> ??? maybe use a RDF container or RDF collection type of resource to list
>> all students enrolled in a certain course?
>> > > 4.       Travel: ??? We are trying to encode trips, in which one or
>> more people leave one place at one time and arrive at another place at
>> another time. This thus links people, places and times.
>> > > 5.       Texts - i.e. old editions of books and manuscripts: Dublin
>> Core, Bibframe. Use FRBR to distinguish sub- and pre-edition levels of
>> manuscripts, works and ideas.
>> > > 6.       Relationship among texts, including intertexts and
>> citations: Bibliographic ontology (Bibo)
>> > > 7.       Collections of texts in historical library catalogs, e.g.
>> from centuries ago: the DC Collection AP. Maybe also the Bibliographic
>> Reference Ontology (BiRO)?
>> > >
>> > > My understanding is that the Linked Open Vocabulary cloud (LOV) is a
>> useful tool for finding relevant ontologies. The Vocabulary of Interlinked
>> Datasets (VoID) seems more like underlying infrastructure - the tool to
>> translate and link data items in a dataset written in one vocabulary to
>> data items in a set written in another.
>> > >
>> > > Any further help or clarifications are much appreciated. Thanks again-
>> > >
>> > > Colin
>> > >
>> > >
>> > > ----------------
>> > > Dr. Colin F. Wilder
>> > > Associate Director
>> > > Center for Digital Humanities (website; projects page)
>> > > Thomas Cooper Library, University of South Carolina
>> > > 1322 Greene St., Columbia, SC 29208
>> > > Phones: office (803) 777-2810 & mobile (603) 831-3998
>> > > Emails: wildercf@mailbox.sc.edu & colinwilder@gmail.com
>> > > open office hours (use week view in upper right)
>> > > frango ut patefaciam
>> >
>> > --
>> > Hugh Glaser
>> >    20 Portchester Rise
>> >    Eastleigh
>> >    SO50 4QS
>> > Mobile: +44 75 9533 4155, Home: +44 23 8061 5652
>> >
>> >
>> >
>> >
>>
>> --
>> Hugh Glaser
>>    20 Portchester Rise
>>    Eastleigh
>>    SO50 4QS
>> Mobile: +44 75 9533 4155, Home: +44 23 8061 5652
>>
>>
>>
>


-- 

*Bernard Vatant*
Vocabularies & Data Engineering
Tel :  + 33 (0)9 71 48 84 59
Skype : bernard.vatant
http://google.com/+BernardVatant
--------------------------------------------------------
*Mondeca*
3 cité Nollez 75018 Paris, France
www.mondeca.com
Follow us on Twitter : @mondecanews <http://twitter.com/#%21/mondecanews>
----------------------------------------------------------