IRC log of rdf-wg on 2012-01-11

Timestamps are in UTC.

15:54:04 [RRSAgent]
RRSAgent has joined #rdf-wg
15:54:04 [RRSAgent]
logging to
15:54:06 [trackbot]
RRSAgent, make logs world
15:54:06 [Zakim]
Zakim has joined #rdf-wg
15:54:08 [trackbot]
Zakim, this will be 73394
15:54:08 [Zakim]
ok, trackbot; I see SW_RDFWG()11:00AM scheduled to start in 6 minutes
15:54:09 [trackbot]
Meeting: RDF Working Group Teleconference
15:54:09 [trackbot]
Date: 11 January 2012
15:54:11 [cygri]
cygri has joined #rdf-wg
15:55:34 [Guus]
Guus has joined #rdf-wg
15:56:05 [yvesr]
Zakim, who is on the phone?
15:56:05 [Zakim]
SW_RDFWG()11:00AM has not yet started, yvesr
15:56:06 [Zakim]
On IRC I see Guus, cygri, Zakim, RRSAgent, AZ, MacTed, LeeF, mischat, ivan, SteveH, AndyS1, manu, davidwood, mdmdm_, gavinc, trackbot, yvesr, manu1, NickH, sandro, ericP
15:57:11 [ericP]
i'll be 10 mins late...
15:57:18 [SteveH]
SteveH has left #rdf-wg
15:57:39 [swh]
swh has joined #rdf-wg
15:59:44 [swh]
Zakim, who's on the phone?
15:59:44 [Zakim]
SW_RDFWG()11:00AM has not yet started, swh
15:59:46 [Zakim]
On IRC I see swh, Guus, cygri, Zakim, RRSAgent, AZ, MacTed, LeeF, mischat, ivan, AndyS1, manu, davidwood, mdmdm_, gavinc, trackbot, yvesr, manu1, NickH, sandro, ericP
15:59:48 [cgreer]
cgreer has joined #rdf-wg
15:59:59 [gavinc]
Zakim, start meeting
16:00:00 [Zakim]
I don't understand 'start meeting', gavinc
16:00:06 [pchampin]
pchampin has joined #rdf-wg
16:00:12 [swh]
Zakim, this will be RDF-WG
16:00:12 [Zakim]
I do not see a conference matching that name scheduled within the next hour, swh
16:00:14 [gavinc]
Zakim this is rdfwf
16:00:44 [swh]
Zakim, this will be RDFWG
16:00:44 [Zakim]
ok, swh, I see SW_RDFWG()11:00AM already started
16:00:53 [swh]
Zakim, who's on the phone?
16:00:53 [Zakim]
On the phone I see ??P0, gavinc, ??P2, +1.206.494.aaaa, mhausenblas, cgreer
16:00:55 [cygri]
zakim, mhausenblas is temporarily me
16:00:57 [Zakim]
+cygri; got it
16:00:59 [yvesr]
Zakim, ??P0 is me
16:01:03 [Zakim]
+yvesr; got it
16:01:22 [Arnaud1]
Arnaud1 has joined #rdf-wg
16:01:25 [Zakim]
16:01:32 [AndyS]
zakim, ??P10 is me
16:01:34 [swh]
Zakim, ??P2 is me
16:01:35 [Zakim]
+AndyS; got it
16:01:41 [Zakim]
+swh; got it
16:01:45 [Zakim]
16:01:49 [AndyS]
zakim, who is on the phone?
16:01:50 [ivan]
zakim, dial ivan-voip
16:01:56 [mischat]
zakim, ??P11 is me
16:01:57 [Zakim]
On the phone I see yvesr, gavinc, swh, +1.206.494.aaaa, cygri, cgreer, AndyS, ??P11
16:01:59 [Zakim]
ok, ivan; the call is being made
16:02:01 [Zakim]
16:02:01 [AZ]
zakim, aaaa is me
16:02:09 [Zakim]
+mischat; got it
16:02:13 [Zakim]
+AZ; got it
16:02:16 [mischat]
zakim, mute me
16:02:17 [Zakim]
+ +1.408.996.aabb
16:02:37 [Zakim]
mischat should now be muted
16:02:46 [Arnaud]
zakim, aabb is me
16:03:13 [Zakim]
+Arnaud; got it
16:03:35 [Zakim]
16:04:26 [Zakim]
16:04:46 [davidwood]
Zakim, David_Wood is me
16:04:47 [Zakim]
+davidwood; got it
16:04:53 [Zakim]
16:05:05 [mischat]
davidwood: ww is not here today, i will scribe
16:05:27 [mischat]
davidwood: i will send you an email on that front
16:05:42 [davidwood]
PROPOSED to accept the minutes of the 4 Jan telecon:
16:05:42 [davidwood]
16:05:53 [mischat]
davidwood: any objections to accepting the minutes ?
16:06:00 [JeremyCarroll]
JeremyCarroll has joined #rdf-wg
16:06:02 [mischat]
RESOLVE accept minutes
16:06:05 [davidwood]
Action item review:
16:06:05 [trackbot]
Sorry, couldn't find user - item
16:06:05 [davidwood]
16:06:05 [davidwood]
16:06:40 [mischat]
davidwood: moving on to open actions …
16:07:00 [Zakim]
16:07:09 [mischat]
davidwood: sandro any update on action 82(?)
16:07:16 [sandro]
16:07:16 [trackbot]
ACTION-82 -- Sandro Hawke to draft well-known URI template and propose WG resolution that it is "stable" enough for IETF. -- due 2011-09-14 -- OPEN
16:07:16 [trackbot]
16:07:25 [mischat]
davidwood: any updates on action 98 ?
16:07:33 [mischat]
action-98 ?
16:07:33 [trackbot]
ACTION-98 -- Sandro Hawke to rdf: and rdfs: namespace should resolve to something that meets best practices -- due 2011-12-31 -- OPEN
16:07:33 [trackbot]
16:08:01 [mischat]
davidwood: shouldn't this be something for the w3c systems team
16:08:15 [mischat]
davidwood: should someone else do this action?
16:08:55 [mischat]
sandro: should we be following what the foaf ns does ?
16:08:59 [Zakim]
16:09:30 [Souri]
Souri has joined #rdf-wg
16:09:43 [cygri]
16:09:43 [mischat]
davidwood: should we do it the way SKOS does it ?
16:10:03 [mischat]
JeremyCarroll: is sandro being too picky here ?
16:10:24 [cygri]
16:10:32 [Zakim]
16:10:33 [mischat]
cygri: there is a document best practices for the vocabs
16:10:36 [yvesr]
i think danbri is not overly keen on the way FOAF is published
16:10:39 [MacTed]
Zakim, OpenLink_Software is temporarily me
16:10:39 [Zakim]
+MacTed; got it
16:10:40 [MacTed]
Zakim, mute me
16:10:40 [Zakim]
MacTed should now be muted
16:10:41 [mischat]
cygri: we should follow the above document ^^
16:10:45 [yvesr]
mainly because they're stuck on 0.1 :)
16:11:10 [mischat]
sandro: what is the user experience when users as for HTML
16:11:10 [mischat]
16:11:40 [mischat]
JeremyCarroll: we need 10 lines of HTML, here is the RDF, this is the namespace
16:12:01 [mischat]
sandro: doesn't want to do that project
16:12:03 [Zakim]
16:12:34 [sandro]
16:12:47 [gavinc]
16:12:48 [mischat]
davidwood: right now if we resolve a url like above ^^, as it stands we get no HTML
16:13:10 [mischat]
davidwood: we shouldn't get RDFXML when asking for a human readable document
16:13:52 [gavinc]
eh, _n isn't that bad in javascript ;)
16:14:14 [mischat]
davidwood: so where are we at now …
16:14:18 [sandro]
zakim, who is making noise?
16:14:28 [Zakim]
sandro, listening for 10 seconds I heard sound from the following: cgreer (9%), Arnaud (5%), sandro (34%), davidwood (30%)
16:14:58 [mischat]
agenda: RDFa working group last call
16:15:12 [mischat]
davidwood: manu asks us to review the RDFa documents
16:15:26 [mischat]
davidwood: davidwood will ping Guus about this
16:15:45 [mischat]
davidwood: charles have you reviewed the RDFa doc ?
16:15:53 [Zakim]
16:16:09 [mischat]
charles : happy with the RDFa doc he reviewed
16:16:45 [ivan]
16:16:47 [mischat]
davidwood: we should be reviewing the document in terms of what the RDF WG are interested in document
16:17:01 [cygri]
16:17:03 [mischat]
davidwood: was reviewing with an RDF WG hat on
16:17:11 [davidwood]
ack ivan
16:17:33 [mischat]
ivan: charles please submit under your own name
16:17:34 [gavinc]
+q to ask about CURIEs
16:17:56 [mischat]
ivan: you can tell from the RDFa, that they are staying clear of the named graph issue
16:17:58 [davidwood]
ack gavinc
16:17:58 [Zakim]
gavinc, you wanted to ask about CURIEs
16:18:04 [mischat]
16:18:18 [mischat]
gavinc: has gone through the RDFa curie's section
16:18:37 [mischat]
gavinc: was wondering whether we should comment on the differences between CURIEs and prefixing ?
16:18:47 [mischat]
ivan: which difference are you referring to ?
16:19:24 [cygri]
q+ to ask whether they aren't the same now
16:19:25 [mischat]
gavinc: the set of URIs which can be represented in CURIES is different from the set of IRIs that SPARQL's & RDF prefixes can represent
16:19:39 [mischat]
gavinc: CURIEs don't work with XML
16:19:51 [mischat]
gavinc: CURIE has a broader set than XML names
16:20:04 [mischat]
gavinc: XML names are valid CURIES and prefix names …
16:20:14 [Zakim]
16:20:28 [mischat]
gavinc: we talked about this when talking about Turtle
16:20:41 [ericP]
16:20:57 [mischat]
davidwood: it would be happy if this would be noted in the spec
16:21:07 [mischat]
davidwood: because it is a syntax issue
16:21:11 [AndyS]
CURIE is very open : prefix+local for anything, then says other syntaxes can restrict.
16:21:14 [davidwood]
ack cygri
16:21:14 [Zakim]
cygri, you wanted to ask whether they aren't the same now
16:21:20 [mischat]
cygri: can you give an example please ?
16:21:35 [mischat]
gavinc: not right now
16:21:37 [cygri]
ack me
16:21:56 [mischat]
JeremyCarroll: 2 use-case to motivated CURIE, 1) ending in numbers
16:22:01 [mischat]
as per the IPTC
16:22:24 [mischat]
ivan: would like to see a very detailed example please :)
16:22:31 [cygri]
thanks in advance gavinc!
16:22:36 [mischat]
davidwood: before next week please
16:22:47 [mischat]
ericP: you have 2 hours ;)
16:23:00 [mischat]
moving on …
16:23:27 [mischat]
davidwood: sandro or ivan, what is the best way to get these comments from this WG to the RDFa WG ?
16:23:38 [mischat]
ivan: ideally we should send the comments to their mailing list
16:24:12 [mischat]
ivan: because when they go to CR, it will be easier for the RDFa folks to handle. Please send comments to the RDFa mailing list
16:24:22 [mischat]
davidwood: a link to the public-comments list ?
16:24:38 [mischat]
ivan: please use the rdfa wg's list
16:24:44 [ivan]
16:25:21 [mischat]
ivan: please use ^^
16:25:23 [gavinc]
hey look an example! CURIE: db:resource/Albert_Einstein vs. PNAME db:resource\/Albert_Einstein that's just escaping, will see about others
16:25:49 [mischat]
topic: named graphs
16:26:19 [mischat]
davidwood: sandro wanted Pat's on scoping, Pat sent an email about it
16:26:31 [mischat]
davidwood: Pat would rather not have bnodes in the 4th column
16:26:45 [swh]
+1 to not allowing bNodes in the 4th slot
16:26:45 [mischat]
davidwood: can we make progress based on cygri being here and Pat's email.
16:27:08 [davidwood] "2c: if we allow bnodes in the 4th position, then please lets make a firm decision what their intended scope is going to be, and that they cannot also occur in other positions in the same graph store. But I vote to not allow bnodes in 4th position in any case."
16:27:20 [mischat]
davidwood: Pat's comments re: bnode in 4th slot ^^
16:27:25 [cygri]
q+ to suggest straw poll, let's allow only IRIs in the 4th slot
16:27:37 [mischat]
+1 to not having them either
16:27:54 [mischat]
sandro: the scope for bnode is a document
16:28:40 [Zakim]
16:28:42 [mischat]
sandro: doesn't think that Pat's comment address his use-case from last week
16:28:47 [JeremyCarroll]
16:29:13 [mischat]
cygri: is confused, quote from Pat was about bnodes and not IRI
16:29:22 [mischat]
i parsed that from the conversation too, fwiwi
16:29:26 [Zakim]
16:29:38 [mischat]
cygri: are we considering using bnodes in the 4th slot ?
16:29:45 [Zakim]
16:30:10 [mischat]
cygri: as all the existing syntax, sparql, currently don't support bnodes in the 4th slot
16:30:28 [AZ]
NQuads allows anything in 4th position
16:30:36 [davidwood]
16:30:43 [cygri]
ack me
16:30:43 [Zakim]
cygri, you wanted to suggest straw poll, let's allow only IRIs in the 4th slot
16:30:46 [mischat]
ericP: you can use a variable which matches in a bnode in SPARQL
16:31:05 [cygri]
AZ, fair enough
16:31:17 [mischat]
AndyS: you can use it in SPARQL query, but datasets don't allow for bnodes in the 4th slot
16:31:18 [davidwood]
ack JeremyCarroll
16:31:35 [sandro]
andy: SPARQL datasets dont allow bnodes in the URI part of the pair
16:31:40 [mischat]
JeremyCarroll: re-capping conversation with Pat from 6 years back
16:31:50 [mischat]
JeremyCarroll: wanted the bnodes in the 4th slot, as he is a big fan
16:31:53 [AndyS]
(checking) sandro UC is convenience of not needing to mint a URI
16:32:05 [AndyS]
s/URI/IRI/ <<--- arrg
16:32:11 [ericP]
16:32:22 [mischat]
JeremyCarroll: couldn't see how to get the RDF graph isomorphism with bnodes in 4th slot
16:32:40 [mischat]
JeremyCarroll: this causes problems when software testing
16:32:49 [ericP]
q+ to ask if that's an artifact of the popular algorythm for isomorphisms
16:33:20 [swh]
q+ to talk about use
16:33:30 [mischat]
sandro: doesn't want bnodes in the 4th slot, but we haven't agreed on a design for our use-cases
16:34:08 [ericP]
JeremyCarroll, if i exhaust a mapping of bnodes to bnodes, why would the additional permutations of having a graph named by a bnode be any harder than the other permutations?
16:34:09 [mischat]
sandro: and dismissing bnodes there, is limiting our final design space, i.e. why limit ourselves now, before we have a design, based upon agreed use-cases
16:34:13 [gavinc]
on the other hand, constricting the design space can help force a design?
16:34:29 [mischat]
davidwood: can you walk through the use-case, which you think definitely requires a bnode there
16:35:35 [mischat]
sandro: if you want to state that "dave asserts these triples", would require a IRI, but a bnode would allow us not to mint a new IRI
16:35:52 [AndyS]
q+ to say IF we allow 4th slot bNodes, THEN limiting such bNodes to only 4th slot seems rather odd.
16:36:05 [mischat]
JeremyCarroll: skolemisation is the work around for this
16:36:56 [sandro]
JeremyCarroll: In general using blank nodes is a good way to indicate that we didnt have a good way to agree on a URI for the thing.
16:36:58 [mischat]
JeremyCarroll: a blank-node would allow different people to articulate that they are talking about the same thing, without agreeing upon what the IRI should be minted before hand
16:36:58 [sandro]
16:37:03 [ivan]
16:37:16 [davidwood]
ack ericP
16:37:16 [Zakim]
ericP, you wanted to ask if that's an artifact of the popular algorythm for isomorphisms
16:37:39 [yvesr]
JeremyCarroll, +1 - skolemisation would imply reconciliation a-posteriori, but i think i also understand why it could be a cause fo concerns
16:38:36 [sandro]
JeremyCarroll: If one bnode is also used as a graph name, then isomorphism is more complicated
16:38:47 [swh]
16:39:02 [davidwood]
ack AndyS
16:39:02 [Zakim]
AndyS, you wanted to say IF we allow 4th slot bNodes, THEN limiting such bNodes to only 4th slot seems rather odd.
16:39:36 [mischat]
AndyS: Pat's point about if used in 4th slot, is not clear
16:40:45 [mox601]
mox601 has joined #rdf-wg
16:41:04 [AZ]
q+ to say that what's allowed in the 4th slot probably depends on what it identifies
16:41:10 [JeremyCarroll]
16:41:16 [mischat]
AndyS: there is a balance to be struck, sometimes it is better to mint a URI, we should find if there is a use-case for not wanting to name a set of triples
16:41:25 [mischat]
AndyS: perhaps using the "_".
16:42:16 [swh]
16:42:23 [mischat]
JeremyCarroll: AndyS was suggesting that bnodes used in the 4th column shouldn't be used in the g-snap named by that bnode
16:42:39 [davidwood]
ack ivan
16:42:52 [mischat]
sandro: we shouldn't limit our design space without clear objective/use-cases in mind
16:43:21 [sandro]
sandro: we should build up designs, rather than chopping off options blindly
16:43:51 [sandro]
+1 ivan: it's like the use of [...] in turtle
16:43:56 [swh]
+1 to ivan
16:44:17 [AndyS]
Relative IRIs do that? e.g. <#abc1>
16:44:19 [swh]
maybe .well-known/genid
16:44:21 [JeremyCarroll]
JeremyCarroll: we could restrict bnodes as graph names to ones that are only used in graphs named with an IRI
16:44:23 [mischat]
ivan: do we need a way in the syntax to mint a new IRI for a use, which is scoped to a document. Bnodes are used in turtle, for when users don't care or want to mint a new IRI, something like [ … ] in bnode, which mints a new IRI and not a bnode
16:44:36 [mischat]
personally that is why I use bnodes
16:44:37 [JeremyCarroll]
JeremyCarroll: that wouild meet most of my objections, and maybe Pat's
16:45:02 [JeremyCarroll]
JeremyCarroll: this for me, proves Sandro's point, that we shouldn't chop off the design space a priori
16:45:05 [mischat]
AndyS: if you parse a file with that syntatic sugar, would you get the same IRI generated ?
16:45:22 [pchampin]
16:45:27 [JeremyCarroll]
16:45:58 [pchampin]
16:46:01 [mischat]
ivan: most people use bnodes when they don't want/care to mint a new IRI
16:46:16 [davidwood]
ack AZ
16:46:16 [Zakim]
AZ, you wanted to say that what's allowed in the 4th slot probably depends on what it identifies
16:46:54 [JeremyCarroll]
q+ to suggest a straw poll on either making decision now or postponing til after the rest of the design is made
16:47:07 [gavinc]
+q to propose a VERY concrete use case for Named Graphs
16:47:32 [mischat]
AZ: maybe we will know how to restrict the 4th slot if we know that it identifies. If it is just a label for a graph, it doesn't matter if it is a literal, IRI or a bnode.
16:47:46 [cygri]
16:47:48 [mischat]
AZ: so the question to answer is, "what does the 4th slot identify" ?
16:47:56 [davidwood]
ack swh
16:48:29 [mischat]
swh: doesn't feel convinced that we haven't exhausted all of the use-cases
16:48:42 [mischat]
swh: has been working with quad-stores for 10 years or so
16:49:06 [mischat]
swh: initially we didn't rule out bnodes in the 4th slot, but it has turned out that people don't actual use them
16:49:47 [mischat]
davidwood: feels that we are in a bit of a deadlock here.
16:49:49 [davidwood]
ack JeremyCarroll
16:49:49 [Zakim]
JeremyCarroll, you wanted to suggest a straw poll on either making decision now or postponing til after the rest of the design is made
16:49:53 [mischat]
sandro: we need to revisit the design
16:50:13 [mischat]
JeremyCarroll: 2nd'ing sandro's position re: revisiting the design
16:50:14 [gavinc]
16:50:19 [sandro]
16:51:09 [ivan]
-> Sandro's three design aproaches
16:51:23 [davidwood]
ack cygri
16:51:35 [swh]
it's an existential variable!
16:51:43 [mischat]
davidwood: doesn't think it seems minor given the ramifications for the semantics, for the various syntax, and the implementations
16:52:02 [mischat]
16:52:16 [swh]
JeremyCarroll, cwm isn't the only system to have graph IDs that are bNodes, 3store did too, it just wasn't very popular
16:52:21 [swh]
…with users
16:53:07 [mischat]
cygri: re: AZ's point, we need to figure out what the interpretation of a dataset. Is it true/false? This will help cygri figure out the semantics of a dataset .
16:53:13 [sandro]
yes, absolutely
16:53:18 [AZ]
16:53:31 [cygri]
ack me
16:53:40 [davidwood]
ack AZ
16:53:42 [sandro]
a dataset must have truth conditions, yes. being true or false.
16:54:12 [cygri]
sandro, i disagree. is a dataset containing several versions of a graph true or false?
16:54:15 [mischat]
AZ: the semantics of a dataset hasn't been decided upon yet, AZ proposed one, Pat didn't like it, but he don't have any progress on this front
16:54:36 [cygri]
thanks AZ
16:54:48 [mischat]
AZ: we don't even have the beginnings of what a dataset is yet, this work needs to be performed
16:55:48 [AZ]
s/he don't have any progress/we don't have any progress/
16:56:35 [mischat]
sandro: is talking about the use case re: graphs ^^
16:56:38 [gavinc]
+q for concrete use case
16:57:22 [mischat]
Use case 1 : Several systems want to use the data gathered by one RDF crawler. They don't need simultaneous access to older versions of the data.
16:57:31 [mischat]
Use case 2: Several systems want to use the data gathered by one RDF crawler. They need simultaneous access to older versions of the data.
16:57:58 [mischat]
davidwood: can you find a real-world example for use-case 2
16:58:01 [swh]
we do provenance of that kind, and we don't model it that way
16:58:09 [AndyS]
16:58:21 [gavinc]
Archiving Crawler Concrete!
16:58:23 [mischat]
sandro: people would like to know how and why data has changed
16:58:44 [mischat]
sandro: would allow for provenance data to be modelled in RDF
16:58:54 [mischat]
Use-case 3 : A system wants to convey to another system in RDF that some person agrees with or disagrees with certain RDF triples.
16:59:20 [mischat]
sandro: these 3 use-case could easily be modelled in trig and in nquads
16:59:51 [mischat]
sandro: the syntaxes get used in different ways, and all of the ways can be used to model the use-cases
17:00:32 [mischat]
sandro: enumerated these are called the ways : Trig/REST, Trig/Equality, and Trig/bnode
17:00:33 [sandro]
third approach: eg:sandro eg:endorses { ... the triples I'm endorsing ... }
17:00:59 [gavinc]
17:02:13 [sandro]
and third design on UC1 is: <> rdf:graphState { ... triples recently fetched from there }
17:03:05 [mischat]
davidwood: most discussion was around the 3rd solution, and we haven't had much discussion on this, probably due to the timing of the email
17:03:06 [davidwood]
ack gavinc
17:03:07 [Zakim]
gavinc, you wanted to discuss concrete use case and to
17:03:56 [mischat]
gavinc: we are talking about archiving data on the web, as one of our use-cases, and we have an ISO standard for it at the moment
17:04:38 [JeremyCarroll]
please post link
17:04:40 [gavinc]
17:04:44 [cygri]
i've worked with it
17:04:55 [mischat]
gavinc: in our use-case, without RDF, and without the SW cached in, when people have designed archiving systems for the web, they minted URIs
17:04:56 [cygri]
most off-the-shelf crawlers support it
17:05:12 [mischat]
gavinc: a standard for archiving data from the web ^^
17:05:29 [mischat]
gavinc: so why are we talking about archiving the web, without minting new IRIs
17:06:06 [JeremyCarroll] ISO 28500.
17:06:17 [mischat]
sandro: please put your comments in context
17:06:37 [mischat]
gavinc: use-case 2 is not necessary for needing
17:06:56 [mischat]
s/needing/motivating bnodes in the 4th column/
17:07:38 [cygri]
17:07:45 [mischat]
davidwood: please motivation use-case 3
17:07:51 [mischat]
17:08:00 [cygri]
17:08:13 [LeeF]
Don't people build their own technology for something like this if they want to do it? How does the Tim Clark type group of people do it?
17:08:42 [davidwood]
ack cygri
17:08:42 [mischat]
sandro: doesn't think that anyone is publishing data for use-case 3 because there are no mechanisms for people to make use of the practices described in use-case 3
17:09:03 [davidwood]
LeeF, Tim Clark type group?
17:09:22 [gavinc]
Specific every WARC record must have an IRI
17:09:28 [LeeF]
project formerly known as SWAN - think it became the scientific discourse sub-group of the SW HCLS IG
17:09:33 [LeeF]
but i don't know a lot about what it's been up to
17:09:38 [LeeF]
17:09:54 [davidwood]
WARC specifies a URI, not an IRI
17:10:20 [mischat]
cygri: doesn't believe that the use-case 3 should be top of our agenda
17:11:23 [LeeF]
I'm not that interested in this use case :-)
17:11:53 [swh]
yes, lets ask the question about who's interested
17:12:12 [gavinc]
since you can create a new graph that contains only the subgraph, and endorse that
17:12:22 [mischat]
JeremyCarroll: thinks that we can endorse a graph, but not a subgraph, and doesn't think this is a major issue
17:12:31 [AndyS]
Also - converse is whether it is a requirement to be solved - middle ground of "not blocked"
17:12:44 [mischat]
davidwood: can we have a straw-poll about who is interested in use-case 3 ?
17:12:49 [LeeF]
Talking about graph versus talking about subgraph?
17:12:52 [MacTed]
sorry... link to this?
17:12:52 [JeremyCarroll]
i am interested in uc3 ...
17:12:55 [AndyS]
17:13:15 [davidwood]
17:13:33 [JeremyCarroll]
q+ tp talk about owl test cases
17:13:35 [mischat]
swh: finds it hard to know what use-case 3 is talking about
17:13:43 [JeremyCarroll]
q+ to talk about owl test cases
17:13:48 [davidwood]
MacTed, UC3 in
17:13:53 [MacTed]
17:14:47 [davidwood]
ack JeremyCarroll
17:14:48 [Zakim]
JeremyCarroll, you wanted to talk about owl test cases
17:15:31 [swh]
SELECT ?s ?p ?o WHERE { eg:sandro eg:endorses ?g } GRAPH ?g { ?s ?p ?o }}
17:15:40 [LeeF]
syntax error!
17:15:41 [mischat]
JeremyCarroll: in the owl test case, there are manifest files which stated that one graph entails another graph. JeremyCarroll thinks this is a different concrete use case regarding what sandro is talking about
17:15:45 [swh]
all the triples endorsed by eg:sandro
17:15:54 [swh]
sorry LeeF :)
17:15:57 [mischat]
sandro: thinks that use-case 4 is touching upon what JeremyCarroll mentioned above ^^
17:16:12 [ericP]
+1 to PML use case
17:16:13 [LeeF]
at least you didn't write "SELECT ?s, ?p, ?o" :-D
17:16:14 [MacTed]
+1 interested in expressing endorsement (agreement with, has confidence in, etc.) of <arbitrary g-snap>
17:16:31 [Zakim]
17:16:40 [AZ]
17:16:41 [Zakim]
17:16:41 [AZ]
17:16:42 [Zakim]
17:16:43 [swh]
LeeF, yeah, after years I finally stopped putting the , in there :)
17:16:43 [Zakim]
17:16:43 [Zakim]
17:16:44 [Zakim]
17:16:45 [Zakim]
17:16:47 [Zakim]
17:16:48 [Zakim]
17:16:48 [Zakim]
17:16:50 [Zakim]
17:16:52 [Zakim]
17:16:53 [Zakim]
17:16:57 [cgreer]
cgreer has left #rdf-wg
17:16:58 [Zakim]
17:17:00 [Zakim]
17:17:02 [Zakim]
17:17:21 [Zakim]
17:17:25 [Arnaud]
Arnaud has left #rdf-wg
17:17:36 [AndyS]
Did we make progress today?
17:17:42 [Zakim]
17:17:42 [Zakim]
SW_RDFWG()11:00AM has ended
17:17:43 [Zakim]
Attendees were gavinc, +1.206.494.aaaa, cgreer, cygri, yvesr, AndyS, swh, Ivan, mischat, AZ, +1.408.996.aabb, Arnaud, sandro, davidwood, LeeF, JeremyCarroll, Souri, MacTed, EricP,
17:17:45 [Zakim]
... pchampin
17:17:52 [LeeF]
There's a disconnect somewhere here
17:17:59 [mischat]
do I have to do things now
17:18:08 [mischat]
make scribe logs or something
17:18:13 [LeeF]
Because I think that people are disagreeing over what needs to happen (if anything) in a design to support UC3
17:18:17 [mischat]
been on holiday for a while :)
17:18:26 [cygri]
trackbot, make logs public
17:18:26 [trackbot]
Sorry, cygri, I don't understand 'trackbot, make logs public'. Please refer to for help
17:18:33 [LeeF]
RRSAgent, make logs public
17:18:39 [cygri]
17:18:44 [LeeF]
mischat, think you want this
17:18:59 [mischat]
thanks LeeF
17:19:35 [LeeF]
17:20:53 [mischat]
cool, i will tidy the logs and mail the mailing list later
17:20:56 [mischat]
bye all
17:50:43 [ericP]
gavinc, ears on to start now?
17:50:49 [ericP]
(not scheduled for another 40 mins)
17:57:17 [gavinc]
Yes, I think so
17:57:44 [gavinc]
18:01:04 [ericP]
18:01:07 [ericP]
wait, you'
18:01:22 [ericP]
re not going to try to sell me an alarm system or auto insurace, are you?
18:12:09 [gavinc]
18:29:55 [gavinc]
[7] predicateObjectList ::= verb objectList ( ";" verb objectList )* (";")?
18:30:08 [gavinc]
( VerbPath | VerbSimple ) ObjectList ( ';' ( ( VerbPath | VerbSimple ) ObjectList )? )*
18:30:28 [gavinc]
verb ObjectList ( ';' ( verb ObjectList )? )*
18:30:52 [ericP]
18:32:19 [gavinc]
18:32:32 [ericP]
[ a :Foo ] .
18:33:58 [gavinc]
[ :a :b ; ]
18:35:36 [gavinc]
_: :a :b ;
18:41:01 [AndyS]
AndyS has joined #rdf-wg
18:42:13 [ericP]
18:45:48 [gavinc]
18:46:04 [gavinc]
[ ; ] .
18:53:49 [gavinc]
18:54:08 [gavinc]
prefix ::= NCName
18:54:09 [gavinc]
reference ::= irelative-ref (as defined in [RFC3987])
18:54:11 [gavinc]
curie ::= [ [ prefix ] ':' ] reference
18:54:12 [gavinc]
safe_curie ::= '[' [ [ prefix ] ':' ] reference ']'
18:57:48 [ericP]
19:00:57 [gavinc]
a mapping to use with the '_' prefix, which is used to generate unique identifiers (for example, _:p).
19:01:22 [ericP] ::= ":" | [A-Z] | "_" | [a-z] | [#xC0-#xD6] | …
19:01:56 [ericP] ::= [A-Z] | [a-z] | …
19:02:14 [ericP]
_asdf is a valid XML name
19:03:14 [gavinc]
19:05:22 [gavinc]
19:07:59 [gavinc]
AndyS: Hey there! What do you think of the horrors of _123: ? ;)
19:08:09 [gavinc]
or _a
19:08:11 [gavinc]
or _abc
19:08:20 [gavinc]
or _.: ;_
19:08:35 [AndyS]
For what?
19:08:55 [gavinc]
Those are CURIES but not PNAMES
19:09:27 [AndyS]
_: is a bnode, <_:abcdef12345> is what "certain" systems use for directly addressing bnodes.
19:09:38 [AndyS]
CURIEs allow _:abc as a IRI don't they?
19:09:46 [gavinc]
look closer ;)
19:09:49 [gavinc]
19:09:53 [gavinc]
not _:abc
19:10:33 [AndyS]
I noticed .... as RDF uses _: already, making _abc:def an IRI is hard to justify.
19:10:56 [AndyS]
But CURIES allow _:abc already ... just as an IRI.
19:11:08 [AndyS]
i.e. RDf and CURIEs diverge.
19:11:19 [gavinc]
RDFa of course uses CURIEs
19:11:39 [gavinc]
so would the feedback be _abc: shouldn't be allowed in RDFa
19:11:49 [AndyS]
Ptr to latests CURIE spec?
19:11:55 [gavinc]
19:12:09 [AndyS]
19:13:11 [AndyS]
How does RDFa (full) handle bNode where the bNode is the object of 2+ triples? (The case where you must have lable).
19:13:20 [AndyS]
a case where you must ....
19:13:38 [AndyS]
sec 7.4.5
19:14:08 [AndyS]
OK - they overload (0verwork?) CURIEs.
19:14:25 [gavinc]
yeah, they use them as prefix names
19:14:31 [gavinc]
while using the CURIE syntax
19:15:02 [AndyS]
Can you define the namespace for "_" ?
19:15:24 [gavinc]
19:15:32 [gavinc]
It's magic
19:16:15 [ericP]
'<' ([^<>"{}|^`\]-[#x00-#x20])* '>'
19:19:47 [AndyS]
Don't think adding it to Turtle has much value and it obfusticates. Technically harmless; probably messes some implementations; certainly socially silly. CURIEs are only popular in certain circles. XML allows QNames/NS with _: and _123: RDF overloaded that. so decision goes back a long way.
19:21:25 [AndyS]
ericP - why not use the real regex full for an IRI? It's only 160+ characters long to include everything.
19:24:18 [ericP]
19:25:36 [gavinc]
19:25:40 [gavinc]
19:26:01 [gavinc]
CURIE, it's broken :D
19:26:05 [ericP]
19:26:14 [ericP]
19:30:03 [AndyS]
Not broken so much as (very) different approach. Make very broad, leave to IRI parsing to do all the validation. ____://a@b:20/eric/?question#now.
19:30:24 [gavinc]
19:30:37 [Zakim]
Zakim has left #rdf-wg
19:30:54 [AndyS]
The allowing authority is new to me. Hadn't noticed before.
19:31:20 [gavinc]
Yeah, irelative-ref has a LOT more in it then one might expect
19:32:45 [AndyS]
Which means ... _123://[ipv6]/ ... is a legal CURIE. [] is only legal in IRIs in the host name for IPv6.
19:33:38 [gavinc] has two paths in the grammar :(
19:43:20 [AndyS]
Isn't it "try as IRI first"?
19:43:37 [cygri]
cygri has joined #rdf-wg
19:48:15 [gavinc]
lets see, not in the CURIE part... maybe somewhere else in the document
19:52:48 [mischat]
mischat has joined #rdf-wg
19:55:07 [gavinc]
<div prefix="http:">
19:55:08 [gavinc]
<span about="http:bob">
19:55:10 [gavinc]
<span property="http://asdf" content="The Trouble with Bob">
19:55:11 [gavinc]
19:55:13 [gavinc]
19:56:44 [gavinc]
19:58:01 [gavinc]
and worse
19:58:03 [gavinc]
<div prefix="http:">
19:58:04 [gavinc]
<span about="http:bob">
19:58:06 [gavinc]
<span property="" content="The Trouble with Bob">
19:58:07 [gavinc]
19:58:09 [gavinc]
19:58:10 [gavinc]
@prefix http: <> .
19:58:12 [gavinc]
http:bob <> "The Trouble with Bob" .
19:58:18 [gavinc]
20:06:08 [ericP]
AndyS, yeah, "try as IRI first" could be implemented by a "lex longest token" rule (depending on how much of the language you stuffed into terminal)
20:07:14 [ericP]
i guess "//" is only allowed after the scheme (though i haven't checked to make sure all of the appropriate path candidates include 1 or more chars)
20:37:40 [cygri]
cygri has joined #rdf-wg
21:22:34 [AndyS]
AndyS has joined #rdf-wg
22:45:40 [cygri]
cygri has joined #rdf-wg