Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Mar 25 12:22

    neumarcx on main

    updated the event time informat… (compare)

  • Mar 25 11:25

    neumarcx on main

    Update README.md (compare)

  • Mar 25 11:24

    neumarcx on main

    just an initial update for the … (compare)

Andy Seaborne
@afs
@klinovp Good point.
Ivan Herman
@iherman
Ie, :_a is the same whether it is within << >> and outside?
It is important to note this if we want to sign an RDF-star graph...
Andy Seaborne
@afs
Signing is structural? = Abstract syntax? Similar to "owl:SameAs two URIs"?
Ivan Herman
@iherman
I am talking about signing on the abstract syntax level
Veronika Heimsbakk
@veleda
Gtg, thanks for a great presentation. Inspiring!
Andy Seaborne
@afs
labels are scoped to the document as before.
Sandro Hawke
@sandhawke
So it's the Accept header on my query, I guess.... I need to include the RDF-star MIME if I want RDF-star results.
Andy Seaborne
@afs
Could be. Implication is server can't send results until all results are seen. e.g. sending back a large graph - need to process whole graph.
Pavel Klinov
@klinovp
the problem with MIME is that it has to be set in the very beginning while it may not be clear till the every end of the query execution whether the query actually returns any new RDF-star terms. At that point some results may have already been sent
Andy Seaborne
@afs
@sandhawke Are there other examples where this has happened?
Similar, not same:: CSS, HTML evolve with same MIME type.
Sandro Hawke
@sandhawke
Not seeing the problem. The server can/should use the rdf-star media type exactly when the client includes the right Accept header -- doesn't matter if there are embedded triples.
Ivan Herman
@iherman
clap clap, thanks
Sandro Hawke
@sandhawke
Yes, excellent presentation!
jay gray
@jaygray0919
clap clap, thanks!
Andy Seaborne
@afs
* / *
Chris
@cw00dw0rd_twitter
Thanks for the great presentation!
Fajar J. Ekaputra
@fekaputra
thanks again for the great presentation and nice discussions!
Sandro Hawke
@sandhawke
Heh -- yeah, well, use / at your own risk? I dunno.
Maxime Lefrançois
@maximelefrancois86
thank you again !
Andy Seaborne
@afs
Yep - no simple answer. If the results are 1B triples, really don't want to check before sending!
Fajar J. Ekaputra
@fekaputra
the zoom link is only for invited participants, right?
I don't find it anywhere
Pavel Klinov
@klinovp
@afs this is the problem we had in Stardog: we wanted all older Stardog clients to be able to parse results of queries even to RDF-star datasets -- as long as those don't contains tuples/triples with triple-terms.
Jon Phipps
@jonphipps
zoom link?
Nick Form
@nickform-ost
Really nice presentation - thanks to the presenters.
Pavel Klinov
@klinovp
@afs and yeah, pre-processing all results on the server-side in order to set the right MIME header is a non-starter
Andy Seaborne
@afs
Also the status code. "406 Not Acceptable" is the ideal : client asked text/turtle, and it there is an RDF-star usage.
Pavel Klinov
@klinovp
right
Andy Seaborne
@afs
A slight change of query and at switches works/does not work. Like ORDER BY-LIMIT
tom lurge
@tomlurge_twitter
But you could implement RDF reification the way you implement RDF-star, couldn't you?
Pavel Klinov
@klinovp
to an extent, yes. but non-atomicity of reified triples is a huge headache.
like imagine a 1B triple data stream where the first triple of the reified representation comes first and the other 2 come last -- they still need to be matched and assembled into a "triple term" (if we talk about the way RDF-star is often supported, i.e. with a new type of RDF term).
neumarcx
@neumarcx
a relay from the zoom chat > From Vassil Momtchev to Everyone: "Here is an evaluation on the performance we did for GraphDB. The dataset is Wikidata using various modelling approaches: https://graphdb.ontotext.com/documentation/free/devhub/rdf-sparql-star.html#how-the-different-approaches-compare. Thanks for the discussion!"
Gregg Kellogg
@gkellogg
From Donny Winston: There is statement reification and there is relationship reification — that is, turning an edge predicate into a class that represents a “hyper edge” that connects to the former triple subject and object through more specialized predicates, and then additional predicates can qualify the relationship. I’m curious when people prefer this approach relative to rdf-star (separate from rdf:Statement reification)
Donny Winston
@dwinston
I did not think about atomicity of triples Pavel. That is fascinating. Is that part of the rdf-star serialization standard, i.e. are embedded triples atomically serialized?
Andy Seaborne
@afs
@dwinston Yes.
Pavel Klinov
@klinovp
@dwinston yeah, everything inside <<>> is a single tning
Andy Seaborne
@afs
c.f. literals - lex+datatype/lang together.
Pavel Klinov
@klinovp
this, by itself, is a huge relief
tom lurge
@tomlurge_twitter
Donny Winston
@dwinston
thanks both!
Pierre-Antoine Champin
@pchampin
funny, I never thought about this use of the annotation syntax (i.e. in a SPARUL DELETE clause) :-)
but that's a good one :)
Donny Winston
@dwinston
@afs thank you for the analogy to literals. makes sense.