IRC log of rdfn on 2010-06-27

Timestamps are in UTC.

00:00:51 [AxelPolleres]
guus: owl:sameAs is semantically not stricter than skos:exactMatch, strange domain/range constraints of skos are problematic
00:01:40 [AxelPolleres]
Jun: context is necessary to determine source of owl:sameAs inference
00:02:55 [AxelPolleres]
stefan: waht to record on the whiteboard?
00:03:27 [AxelPolleres]
ivan: maybe now/later: identity-vocabulary
00:03:39 [AxelPolleres]
jeremy: maybe never
00:03:57 [dbooth]
Topic: Afternoon Wrap-up
00:04:20 [tomlurge]
tomlurge has joined #rdfn
00:05:10 [AxelPolleres]
david: we should have some results tomorrow...
00:05:24 [AxelPolleres]
sandro: we should have micro-charters
00:06:02 [AxelPolleres]
... as results that say what kind of group should do what to work out some features
00:06:52 [AxelPolleres]
mikeUshold: we should prioritize (not enough people to do breakouts on all) issues
00:08:27 [AxelPolleres]
ivan: we should only work on Core tomorrow, not on Infrastructure
00:08:31 [AxelPolleres]
00:08:52 [AxelPolleres]
david: let's better go through one by one
00:09:31 [AxelPolleres]
... named graphs might be the easiest one to start with.
00:10:06 [AxelPolleres]
most people think named graphs is "maybe now", few "maybe later"
00:11:02 [AxelPolleres]
... standardise Turtle?
00:11:06 [AxelPolleres]
most people for it
00:11:35 [AxelPolleres]
... follow your nose (formalise HTTP Get)?
00:12:05 [AxelPolleres]
more maybe nows than maybe later... not as clear as the previous
00:12:19 [AxelPolleres]
... json RDF serialisation?
00:12:46 [AxelPolleres]
14-15 maybe now, frewer maybe later
00:12:54 [AxelPolleres]
... ATOM
00:13:01 [AxelPolleres]
more maybe later than maybe now
00:13:25 [AxelPolleres]
... canonical RDF/XML fixing RDF/XML
00:13:38 [AxelPolleres]
majority maybe never
00:14:27 [AxelPolleres]
... literals as subject?
00:14:34 [AxelPolleres]
majority maybe never
00:15:38 [AxelPolleres]
00:15:58 [AxelPolleres]
majority maybe now (15 now, 8 later, 5 never)
00:16:20 [AxelPolleres]
... consolidate RDF RDFS namespaces?
00:16:27 [AxelPolleres]
majority never
00:16:47 [AxelPolleres]
ivan: alternative would be RDFa profiles.
00:17:58 [AxelPolleres]
axel: could be mentioned as a feature of a Turtle standardisation charter
00:18:43 [AxelPolleres]
RDF URIs vs IRI (vs Qnames???)
00:18:51 [AxelPolleres]
majority never
00:19:14 [AxelPolleres]
... n-ary predicates?
00:19:28 [AxelPolleres]
maybe later
00:19:57 [AxelPolleres]
RDF URIs vs IRI (vs Qnames???)
00:20:14 [AxelPolleres]
change from before maybe now, but needs no discussion
00:20:23 [AxelPolleres]
... binary?
00:20:25 [AxelPolleres]
00:20:35 [AxelPolleres]
revise RDF semantics?
00:20:37 [AxelPolleres]
00:20:43 [AxelPolleres]
00:20:48 [AxelPolleres]
maybe later
00:21:54 [AxelPolleres]
RDFa profiles?
00:22:31 [AxelPolleres]
00:22:53 [AxelPolleres]
Rule based querying?
00:22:58 [AxelPolleres]
00:26:42 [AxelPolleres]
ivan, guus: infrastructure topics could go into another WG such as SWEO, Best practices, etc.
00:27:47 [AxelPolleres]
next feature ...
00:28:02 [AxelPolleres]
identity for non-RDF resources?
00:28:40 [sandro]
For infrastructure issues, the questions is: Does anyone want to fight for one of these to be in a Core group's charter
00:28:50 [AxelPolleres]
(the following features are in question whether we need to discuss that tomorrow)
00:28:57 [AxelPolleres]
identity, not tomorrow
00:29:38 [AxelPolleres]
provenance vocabularies, not tomorrow
00:30:04 [AxelPolleres]
annotations for RDF graphs, tomorrow
00:30:28 [AxelPolleres]
standard APIs, not tomorrow
00:30:46 [AxelPolleres]
Unified query layer, not tomorrow
00:31:38 [AxelPolleres]
context was merged into annotations
00:31:45 [AxelPolleres]
change vocabularies, not tomorrow
00:31:53 [AxelPolleres]
isolation APIs, not tomorrow
00:32:51 [AxelPolleres]
identity vocabularies, not tomorrow
00:33:14 [AxelPolleres]
petermika, guus: think it's important for linked data
00:33:56 [AxelPolleres]
david, ivan: important issue, but not necessarily in the resulting working group
00:34:21 [AxelPolleres]
jjc: argument against, we shouldn't guess what the LOD community will do/use
00:35:23 [AxelPolleres]
richard: owl:sameAs was used because it was the closest available... if we want to fix that, we need to offer something.
00:36:11 [AxelPolleres]
ian: LOD community uses owl:sameAs as a vocabulary.
00:37:03 [AxelPolleres]
Lin: does the use of RDF depend on vocabularies? for defining vocabuaries we need to make sure whether that is essential.
00:37:34 [AxelPolleres]
identity vocabularies: discussed tomorrow
00:39:38 [AxelPolleres]
change vocabulary: later
00:39:50 [AxelPolleres]
weakly deprecation?
00:40:03 [AxelPolleres]
maybe now
00:41:11 [pfps]
Main Entry: deprecate
00:41:13 [pfps]
Part of Speech: verb
00:41:15 [pfps]
Definition: belittle, condemn
00:41:16 [pfps]
Synonyms: cut down to size, depreciate, derogate, detract, disapprove of, discommend, discountenance, disesteem, disfavor, disparage, expostulate, frown, mudsling, not go for, object, pooh-pooh, poor mouth, protest against, put down*, rip*, run down, take dim view of, take down, take exception to
00:41:35 [AxelPolleres]
s/weakly deprecation+bestpractices/
00:41:36 [pfps]
I like belittle or derogate or disesteem
00:41:54 [pfps]
but there are lots of other good candidates
00:42:19 [AxelPolleres]
topic: agenda tomorrow
00:42:34 [AxelPolleres]
stefan: suggest to group in 1) syntax 2) semantics 3) other
00:43:10 [AxelPolleres]
00:43:36 [AxelPolleres]
json, turtle ... breakout group lead by dajobe
00:44:53 [AxelPolleres]
(named graphs go under other, though they have both syntax and semantics bits)
00:45:29 [AxelPolleres]
RDFa profiles fo under syntax as well
00:46:32 [AxelPolleres]
00:49:33 [AxelPolleres]
RDF Semantics doc, RDFS/pfps, Bnodes, follow-your-nose
00:49:41 [AxelPolleres]
00:49:56 [AxelPolleres]
Named graphs vs n-quads, annotations
00:51:15 [AxelPolleres]
(further minutes/notes on tomorrow's agenda will be collected and pasted to the minutes by david Wood ...)
00:55:39 [james]
james has joined #rdfn
00:55:59 [dwood]
dwood has joined #rdfn
00:56:03 [Anchakor]
/me thanks all scribes
00:56:06 [dwood]
00:56:06 [dwood]
1. Syntax (Dave Beckett, TBD)
00:56:06 [dwood]
- Turtle
00:56:07 [dwood]
00:56:08 [dwood]
- RDFa profiles
00:56:09 [dwood]
2. Semantics (Paul Gearon)
00:56:12 [dwood]
- Bnodes
00:56:14 [dwood]
- Revising RDF Semantics
00:56:16 [dwood]
- Weak deprecation
00:56:18 [dwood]
3. Graph Metadata (Elisa Kendall)
00:56:20 [dwood]
- Named Graphs
00:56:22 [dwood]
- Annotations
00:56:24 [dwood]
4. Linked Data (David Booth)
00:56:26 [dwood]
- Identity vocabulary
00:56:28 [dwood]
00:56:47 [_mdean]
_mdean has joined #rdfn
00:56:49 [dwood]
rssagent, draft minutes
00:57:04 [dwood]
rrsagent, draft minutes
00:57:04 [RRSAgent]
I have made the request to generate dwood
00:57:38 [FabGandon]
FabGandon has left #rdfn
00:59:11 [sandro]
RRSAgent, make minutes
00:59:11 [RRSAgent]
I have made the request to generate sandro
00:59:28 [sandro]
rrsagent, make logs public
01:56:12 [iand]
iand has joined #rdfn
03:31:24 [cygri]
cygri has joined #rdfn
03:32:27 [AndyS]
AndyS has joined #rdfn
04:08:44 [dbooth]
dbooth has joined #rdfn
04:56:11 [tomlurge]
tomlurge has joined #rdfn
05:01:06 [cygri]
cygri has joined #rdfn
05:25:12 [pgearon]
pgearon has joined #rdfn
05:42:48 [cygri]
cygri has joined #rdfn
06:24:13 [FabGandon]
FabGandon has joined #rdfn
06:28:20 [FabGandon]
Boards to prepare breakout sessions:
06:42:54 [FabGandon]
FabGandon has left #rdfn
16:16:39 [RRSAgent]
RRSAgent has joined #rdfn
16:16:39 [RRSAgent]
logging to
16:16:44 [pfps]
pfps has joined #rdfn
16:16:45 [mdean]
mdean has joined #rdfn
16:16:45 [sandro]
rrsagent, make log public
16:17:01 [dbooth]
dbooth has joined #rdfn
16:17:02 [sandro]
meeting: RDF Next Steps Workshop
16:17:20 [ekendall]
ekendall has joined #rdfn
16:17:30 [ekw]
ekw has joined #rdfn
16:17:54 [dbooth]
scribe: dbooth
16:18:11 [dbooth]
Topic: Tim's late submission
16:18:42 [dbooth]
Ivan: I think tim wants literals as subjects to allow quoted graphs in subject
16:18:54 [dbooth]
Tim's late submission:
16:19:34 [dbooth]
Ivan: I don't understand what his #2 in "Extending the RDF model to allow N3" means.
16:20:22 [dbooth]
stefan: #1 would be taken care of by the named graph
16:20:34 [dbooth]
... And #2 I don't know what this means.
16:22:41 [dbooth]
stefan: So we should put the literals as subjects on the agenda
16:22:51 [dbooth]
jeremy: He also suggested revisiting rdf/xml syntax
16:24:02 [dbooth]
stefan: but i think that may be covered by deprecating
16:24:12 [dbooth]
(scribe not sure he got stefan's sentiment correct)
16:24:38 [dbooth]
s/deprecating/deprecating certain features
16:25:55 [webr3]
does anybody disagree with any items 1-7 under 'Cleaning up the RDF model'
16:26:52 [dbooth]
Topic: Logistics for today
16:26:52 [DaveB]
16:27:47 [mscottm2]
mscottm2 has joined #rdfn
16:28:11 [dbooth]
Ivan: All discussions in break-out groups should be recorded. Use IRC and make sure RRSAgent is on the channel, and at the end say "rrsagent, make logs public" and "rrsagent, draft minutes".
16:28:47 [mscottm]
mscottm has joined #rdfn
16:28:52 [dbooth]
... We'll also have a plenary after at 3:30 and chairs will report.
16:29:09 [DaveB]
I won't be present for that, so somebody else will have to report for Syntax
16:29:39 [dbooth]
ivan: we also didn't decide yesterday whether to draft a charter. we can collaborate on a draft.
16:30:10 [DaveB]
16:30:28 [dbooth]
... If we start a group we would first draft one w sandro, share it with this group to see if it reflects overall agreement with this workshop, then go to SWIG and others for broader discussion this summer.
16:30:44 [dbooth]
fabien: should we look at a plenary later this year?
16:31:01 [dbooth]
ivan: Not sure it's realistic becuase of the voting time needed by the AC. October is too early.
16:31:16 [pgearon]
pgearon has joined #rdfn
16:31:28 [dbooth]
... realistically we could start a group in january.
16:31:46 [mdean]
mdean has joined #rdfn
16:31:55 [dbooth]
andy: are we just considering an RDf WG or would we also consider best practices.
16:32:17 [dbooth]
dwood: that would fall out of the WG
16:32:38 [dbooth]
ivan: due to the manpower situation, it may all be in one group.
16:33:18 [dbooth]
jeremy: representation by wider group. people here are not paying the full price.
16:34:11 [dbooth]
sandro: goal of this workshop is not to reach consensus on the charter, but to enumerate the pros and cons of a charter, because only w3c members can do that. we tech people produce our suggestions that then go to the AC.
16:34:49 [dbooth]
ivan: true, but if we go to the next step (the community) whatever text we go out there with, it woujld look funny if somebody strongly objected.
16:35:31 [dbooth]
richard: from the list of things on the whiteboard, there isn't much that requires changes to rdf, but adding things. Not sure its an RDF WG that is required.
16:35:33 [DaveB]
the draft charter above is missing several things we discussed - annotations, follow your nose, idenitity model
16:35:42 [dbooth]
ivan: But it's a WG -- i don't care what it's called.
16:36:07 [dbooth]
jeremy: "RDF Core" might fly better to AC.
16:36:32 [dbooth]
Topic: Sandro's draft charter
16:37:37 [jjc]
jjc has joined #rdfn
16:37:58 [dbooth]
sandro: suggestion for what the break=outs should do. Draft a paragraph of what might go into the charter, and outline the pros and cons
16:38:31 [juansequeda]
juansequeda has joined #rdfn
16:39:54 [FabGandon]
RDF Core Charter 2010:
16:40:11 [pgearon]
pgearon has joined #rdfn
16:41:10 [dbooth]
gus: looks rather formal -- like a charter.
16:45:02 [dbooth]
daveb: maybe add a references section
16:46:24 [dbooth]
ivan: breakout chairs: elisa, dbooth, daveb, andy
16:48:28 [AxelPolleres_]
AxelPolleres_ has joined #rdfn
16:50:22 [dbooth]
Breakout groups:
16:50:35 [dbooth]
16:51:20 [ekw_]
ekw_ has joined #rdfn
16:53:57 [pfps]
16:54:47 [DaveB]
DaveB has joined #rdfn
16:55:13 [AndyS]
AndyS has joined #rdfn
16:55:41 [AndyS1]
AndyS1 has joined #rdfn
16:55:54 [DaveB]
DaveB has joined #rdfn
16:55:59 [mscottm]
mscottm has joined #rdfn
16:57:34 [DaveB]
the syntax group has gone to work in #rdfn1
16:58:13 [FabGandon]
RDF Next Step Group is on chanel #rdfn-meta
16:58:42 [AndyS]
AndyS has joined #rdfn
16:59:17 [dbooth]
dbooth has joined #rdfn
16:59:52 [pgearon]
pgearon has joined #rdfn
17:01:11 [cygri]
cygri has joined #rdfn
17:01:26 [ekw]
ekw has joined #rdfn
17:01:54 [pgearon]
it's #rdfn-s
17:03:28 [dbooth]
The linked data break-out will use channel #rdfn-lod
17:05:59 [FabGandon]
Named graphs and metadata logs :
17:06:26 [AndyS]
AndyS has joined #rdfn
17:07:56 [AndyS1]
AndyS1 has joined #rdfn
17:11:38 [AndyS1]
Currently reviewing what we are doing ... looking at TBL syntax issues
17:11:49 [AndyS1]
... building a list of everything on the wiki
17:13:28 [AndyS1]
17:18:26 [AndyS1]
Spliiting to concrete syntax issues and data model issues
17:19:49 [AndyS1]
"Syntax" lists was a collector for homeless items - now categorising into two sections
17:31:03 [AndyS1]
Now Turtle and NG-Turtle/Trig(?)
17:34:49 [AndyS]
AndyS has joined #rdfn
17:44:02 [AndyS1]
AndyS1 has joined #rdfn
17:45:42 [AndyS1]
Idea of a Turtle themes syntax stack
17:46:18 [AndyS1]
(we are updating the wiki as we work - currently generating content, sort out more later)
17:46:23 [AndyS1]
Working on RDF - JSON
17:51:56 [kasei]
kasei has joined #rdfn
17:59:50 [ekw]
ekw has joined #rdfn
18:01:34 [tomlurge]
tomlurge has joined #rdfn
18:02:04 [AxelPolleres]
AxelPolleres has joined #rdfn
18:07:13 [AndyS]
AndyS has joined #rdfn
18:22:28 [kasei]
kasei has joined #rdfn
18:23:42 [AndyS1]
AndyS1 has joined #rdfn
18:24:28 [AndyS2]
AndyS2 has joined #rdfn
18:25:52 [AndyS]
AndyS has joined #rdfn
18:28:03 [AndyS1]
AndyS1 has joined #rdfn
18:28:35 [AndyS2]
AndyS2 has joined #rdfn
18:29:52 [AndyS3]
AndyS3 has joined #rdfn
19:03:54 [AndyS]
AndyS has joined #rdfn
19:41:20 [_mdean]
_mdean has joined #rdfn
19:43:11 [kasei]
kasei has joined #rdfn
19:43:44 [FabGandon]
FabGandon has left #rdfn
20:27:15 [LeeF]
LeeF has joined #rdfn
20:27:52 [LeeF]
RRSAgent, pointer?
20:27:52 [RRSAgent]
20:33:08 [pfps]
pfps has joined #rdfn
20:33:43 [jjc]
jjc has joined #rdfn
20:34:24 [pgearon]
pgearon has joined #rdfn
20:36:21 [cygri]
cygri has joined #rdfn
20:38:45 [ekw]
ekw has joined #rdfn
20:38:54 [AndyS]
AndyS has joined #rdfn
20:40:01 [AxelPolleres]
AxelPolleres has joined #rdfn
20:40:44 [tomlurge]
tomlurge has joined #rdfn
20:41:13 [_mdean]
_mdean has joined #rdfn
20:41:41 [AndyS1]
AndyS1 has joined #rdfn
20:47:24 [jun]
jun has joined #rdfn
20:48:36 [mdean]
mdean has joined #rdfn
20:56:12 [dwood]
dwood has joined #rdfn
21:31:44 [jeremycarroll]
jeremycarroll has joined #rdfn
21:39:47 [pgearon]
pgearon has joined #rdfn
21:39:59 [dwood]
dwood has joined #rdfn
21:40:06 [AndyS]
AndyS has joined #rdfn
21:41:47 [ekw]
ekw has joined #rdfn
21:47:04 [mscottm]
mscottm has joined #rdfn
21:47:07 [pfps]
We're baaack.
21:48:49 [cygri]
cygri has joined #rdfn
21:48:49 [ekendall]
ekendall has joined #rdfn
21:49:59 [FabGandon]
FabGandon has joined #rdfn
21:52:00 [mikeU]
mikeU has joined #rdfn
21:52:38 [mikeU]
Andy presenting syntax
21:52:46 [mikeU]
Principles: do they confuse.
21:52:57 [mikeU]
DW: we are not in position to come up with this list.
21:53:17 [mikeU]
SH: charter should decide si higher level princple. What level of compatibiliy are we guaranteeing or not.
21:53:56 [mikeU]
SH: suprirse by extent to which we kind of wanted to just deprecate RDF/XML, even though a nonstarter in the community
21:54:30 [mikeU]
DB: the notion of weakly deprecated, did we capture this somewhere?
21:54:39 [mikeU]
ANdy: yes.
21:55:06 [mikeU]
Ivan: excpet for RDF ID, these are not RDF/XML things that we want to deprecate. First paragraph almost on semantic level.
21:55:15 [mikeU]
SH: error here, disconnected form data model
21:55:24 [mikeU]
Ivan: good to separate what is RDF/XML features.
21:55:41 [mikeU]
Jeremhy: property attributes muddled.
21:55:50 [mikeU]
SH: i don't find this way of splitting helpful.
21:55:59 [mikeU]
ANdy: I wanted to split them this way.
21:56:06 [mikeU]
SH: huge problem in TUrtle.
21:56:13 [mikeU]
Jeremy: was conscious choice.
21:56:37 [mikeU]
Andy: there are problems, things that codl be done better.
21:56:42 [AndyS1]
AndyS1 has joined #rdfn
21:57:01 [mikeU]
Ivan: what weaklyh deprecated means should be removed from RDF/XML.
21:57:14 [mikeU]
SH: this is combined with next item, lets look there before making changes
21:57:23 [mikeU]
ANdy: reificatin was weakly deprecated last time.
21:57:40 [mikeU]
Jeremy: it was downplayed, did not use the word 'deprecate'.
21:58:15 [mikeU]
SH: editing issues. Should this go into semantic issues group?
21:58:34 [mikeU]
ANdy: can we go ahead and leave it?
21:58:43 [mikeU]
SH: i don't want to mess with it now.
21:58:54 [mikeU]
Ivan: collections syntax.
21:59:15 [mikeU]
ANdy: syntax is very awkward, annoying. How did it ever get to this? Confuses people.
21:59:33 [AxelPolleres]
AxelPolleres has left #rdfn
21:59:41 [mikeU]
ANdy: it will break if any changes to new data.
22:00:06 [mikeU]
ANdy: we want to encourage and discourage certain things to minimize damage.
22:00:09 [AxelPolleres]
AxelPolleres has joined #rdfn
22:00:40 [mikeU]
SH: what do we think of the term 'weakly deprecated'?
22:00:47 [AxelPolleres]
AxelPolleres has changed the topic to: plenary - wrapup of breakout groups (AxelPolleres)
22:00:57 [mikeU]
ANdy: we do expect parsers to handle weaklyh deprecated syntax, and not wave hands and forget about it
22:01:25 [mikeU]
Ivan: what you seem to propose.. we would try now to get together a JSON XG?
22:01:56 [mikeU]
ANdy: yes. IF we wait a year... we lose may ablity to do useful prework.
22:02:27 [mikeU]
ANdy: the styly of getting gitems of prework - is it a useful style?
22:02:57 [mikeU]
SH: i am optimistic that this list will put community on notice that these are important issues, so get thinking about this if it is important to you.
22:03:11 [mikeU]
Ivan: this page will eventually go public (actually, already is).
22:03:32 [mikeU]
SH: some urgenecy to start soon, does not specifically say 'incubator'
22:03:56 [mikeU]
ANdy: RDFa trying to get going soon, certainly before this working group.
22:04:00 [mikeU]
Ivan: why relevant here?
22:04:07 [mikeU]
Andy: not for JSON
22:04:34 [mikeU]
Ivan: only thing is RDFa will be frozen if group comes up with named graph, RDFa will not be able express it.
22:05:00 [mikeU]
SH: we now have many syntaxes. We will move into triple syntax or quad syntax.
22:05:19 [mikeU]
Ivan: RDFa might go into V2 before too long.
22:05:32 [mikeU]
Andy :what is name of Turtle as triple syntax?
22:05:41 [mikeU]
Ivan: if we speak about quads.
22:05:54 [sandro]
sandro: RDF/XML and RDFa will be "triple" syntaxes and Turtle and JSON may turn into "quad" syntaxes.
22:06:32 [mikeU]
Ivan: ok, any questions on the syntax part?
22:09:10 [_mdean]
_mdean has joined #rdfn
22:30:46 [cygri]
cygri has joined #rdfn
22:31:59 [AndyS]
AndyS has joined #rdfn
22:32:39 [_mdean]
_mdean has joined #rdfn
22:33:24 [ekw]
ekw has joined #rdfn
22:35:22 [pgearon]
pgearon has joined #rdfn
22:36:42 [scor]
scor has joined #rdfn
22:36:52 [mikeU]
SH: straw poll... there is huge user community hearing us say they should use RDF/XML. QUESTION: should we put some effort into doing a new RDF/XML (maybe just add named graphs)
22:37:21 [mikeU]
Jeremy: my view of reading syntax breakout. We can add one attribute for named grpahs and be done. This seems plausible.
22:37:31 [mikeU]
SH: having one attribute is as bad as doing everything.
22:37:47 [mikeU]
SH: you will need parallel paths. Is there any other way?
22:37:53 [mikeU]
DB: will this break existing parsers?
22:38:09 [mikeU]
Jeremy: my parser would ignore it and move on.
22:38:14 [mikeU]
SH: will do wrong thing.
22:38:26 [mikeU]
DW: yes, it will not recognize the graph, the wrong thing.
22:38:35 [mikeU]
SH: doing wrong thign is a very bad idea.
22:38:41 [mikeU]
DB: consumer;s choice, how to handle it.
22:38:56 [mikeU]
SH: consumers not know enough to make informed choices.
22:39:36 [mikeU]
?? - updtate version of parser... old parsers not handle named graphs. But id alread not know how to handle a named graph. ALso most people are using Jeremy's implementation anyway.
22:40:03 [mikeU]
ANdy: Jeremy, can you by tonight... update the parser (laughter)
22:40:34 [pgearon]
22:40:38 [mikeU]
RC: we should learn from past. Lesson from old XML story. Many proposals, none went anywyere. Why expect anything different now?
22:41:18 [mikeU]
Ivan: I think there are a large no of applications out there relying on RDF/XML. IT was the standard format, turtle was for geeks. They many nto like it, but they live with it.
22:41:39 [AndyS]
Process point: are we hearing from the other breakout groups or are we crossing an event horizon on XML-related stuff.
22:41:51 [mikeU]
Ivan: for them, the worse may be over Might be worse to ask them to start over again, than to accept small change. Transition path not that clear.
22:41:51 [_mdean]
_mdean has joined #rdfn
22:42:00 [mikeU]
SH: my use case is govt agencies publishing linked data.
22:42:18 [mikeU]
Peter: how many people in known universe...
22:42:31 [mikeU]
SH:pubishing data for 9B people to consume.
22:43:07 [mikeU]
Dave? timing.l.if you deprecate this, then there is the issue that new tools will nto accept RDF/XML. People who continue to publish RDF/XML will not work.
22:43:25 [mikeU]
? noone say depreceate RDF/XML, we are makingit second class citizen
22:43:36 [mikeU]
? this is defacto deprecation.
22:44:41 [dbooth]
dbooth has joined #rdfn
22:44:58 [mikeU]
DW: we oftenhear: lets put the web back into the semantic web. Much of our message has been deviated/diverted because people use XML in government, but we ignor web authors. Now we are talking aboutg TUrtle andJSON. This is more about web authors and less about big data in data centers. I am ok to tweak RDF/XML if keeps it current with named grphas.
22:45:37 [mikeU]
Jeremy: mistake to say RDF/XML is not compatibel with named grahs. Whole point is that a graphs has a name, the URL is it. Not such a big problem.
22:46:01 [mikeU]
STefan D: we maintain compatibility of course, not choice.
22:46:27 [mikeU]
Ivan: this is not the issue. Named graph is an example. Do we go extra mile to put the features in RDF/XML? or we we essentially freeze RDF/XML
22:46:48 [mikeU]
Stefan: if we freeze it, then there is backwards incompatibility.
22:47:13 [mikeU]
Stefan: we may need meta-features to have named graphs supported.
22:47:37 [mikeU]
Ivan: if we accept change in model, do we add syntax into RDF/XML to cover it.
22:47:57 [mikeU]
RC: we should not make the change unless we are prepared to do this.
22:48:17 [mikeU]
SH: do we agree that any changes have to be such that RDF/XML is still viable.
22:49:00 [mikeU]
MikeD: general agreement to update RDF/XML with any new features.
22:49:26 [mikeU]
SH: poll: how many people comfortable with updating RDF/XML? 2/3 of people say yes.
22:49:46 [mikeU]
DB: many people not like RDF/XML but many people want to do new version of RDF/XML.
22:49:53 [mikeU]
DB; would it be backwards compatible?
22:50:00 [mikeU]
SH: i think it shoudl be.
22:50:11 [mikeU]
Dave: back compatibily is key.
22:50:26 [mikeU]
Ivan: I fully agree.
22:50:43 [mikeU]
Dave: if you have only old parser, ti will still work.
22:51:22 [mikeU]
ANdy: you may protect the producers of information but you do not protect the consumers.
22:51:26 [AxelPolleres]
Axel: Any update on RDF/XML it MUST be backwards comaptibility
22:51:30 [Anchakor]
does backwards compatible mean not doing the wrong thing?
22:51:42 [ekw]
ekw has joined #rdfn
22:51:58 [mikeU]
DW: the problems of what happens if you make a new RDF/XML format that looks like it is working, but does not quite work. IT woudl be trivial to make forma that will break parsers.
22:52:36 [mikeU]
DW: URL is the name of the graph. IT is also true, yet againj, that some features of RDF donot serializ into RDF/XML or RDFa.
22:53:11 [mikeU]
DW? what if we guarantee in new world tha tanything we can do in RDF we can do in Turtle. That is wher ewill will likey end up anyway. Lets find middl groud.
22:53:52 [mikeU]
RC: I am scared by idea of revisiting RDF/ XML. This is worst possible things that can come out of this meeting. To me, it is horrible w/ only one things: consistently implemented.
22:54:40 [mikeU]
RC:The ONLY good thing about RDF/XML is that it is truly interoperable, it works eveyr where as std syntax. Let us NOT BREAK IT!
22:55:12 [mikeU]
RC: Lets leave it as it is. Lets move forward to using Turtle and JSON.
22:55:33 [mikeU]
Jeremy: is RDF:atgraph too big a change ?
22:55:37 [mikeU]
RC: not sure.
22:56:11 [mikeU]
Jeremhy: we could have anoher RDF/XML that is identical but for this one thing. New namespace.
22:56:56 [mikeU]
DB: only value in current RDF/XML is that it is a true standard. Otherwise just plain horrible. So why keep it around? Just move forward.
22:57:00 [mikeU]
? (Ontotext) ?
22:57:11 [mikeU]
DB: that means using one parser instead of two.
22:57:41 [dbooth]
s/keep it around/make a new version of it/
22:57:43 [mikeU]
? ONtotext: yes. Jena's parser is most popular. Join with Sesame parser then you hve maybe 80% of all parsing convered.
22:57:58 [mikeU]
to mscottm: thanks!
22:58:24 [dbooth]
s/Just move forward/If turtle is standardized that, then why bother with a new version of RDF-XML?
22:58:25 [mikeU]
SH: i think we are at diminishing returns.
22:59:05 [mikeU]
DW: HTML5 experience makes it look like XHTML is deprecated. God only knows what will happen.
22:59:24 [mikeU]
Ivan: we will see on mailing list when charter goes public what people think. Look forward to this.
23:01:06 [mikeU]
Evan: I don't hate RDF/XML, so implicitly depracating it will send a very bad message.
23:01:40 [mikeU]
Jeremyh: from semantics bfreakout.. we discussed arguments agains literals as subjects is that it breaks RDF/XML.
23:01:55 [mikeU]
SH: so we don't want to break RDF/XML and we don't want to fix it?
23:02:03 [mikeU]
Jeremy: maybe ask RDF atribute.
23:02:13 [mikeU]
SH: need to do technical investigation.
23:02:27 [mikeU]
Jeremy: literals as subjects is totally crap from this point of view.
23:02:35 [mikeU]
DW: I agree, i had not thought of this before.
23:02:59 [mikeU]
23:04:41 [mikeU]
Paul: inference rules. Not complete.
23:04:46 [mikeU]
Jeremy: there is a bug.
23:04:52 [dwood]
dwood has joined #rdfn
23:05:16 [mikeU]
Ivan: there are two things, one there is a bug. Two yo uhave to adopt putting a cap on underscore ? (did not get this).
23:05:56 [mikeU]
Ivan: my question is whether we would propose to change semantics on that point
23:05:58 [mikeU]
Peter: this is for later.
23:06:35 [mikeU]
? fix of adding one extra rule.
23:06:42 [mikeU]
Ivan: bnodes cannot be in predciate positions.
23:07:08 [mikeU]
? this is the one. Extra rule you could add even if you disallow ? in ? position.
23:07:42 [AxelPolleres]
23:07:46 [mikeU]
Paul: bnodes... (reading the wiki notes)
23:08:34 [AxelPolleres]
... this paper describes an alternative to ter Horst's approach of allowing non-RDF intermediate triples, by adding a new "implicit typing" rule to the RDFS inference rules
23:09:18 [mikeU]
Paul: slolemizatoin, do not create URI of rit, but have a namespace. A namespace that is not a URI.
23:09:21 [AxelPolleres]
... anyways, what almost everybody does is the "ter Horst" way, i.e. allowing some generalised RDF for interemediate triples
23:09:28 [mikeU]
DB: this is one possibility, not recommending it per se.
23:10:05 [mikeU]
MikeD: does this mean systems have to remember?
23:10:07 [mikeU]
Paul: yes.
23:10:28 [mikeU]
Paul: if you import same graph twice, you will get two different skolemized constants for same bnode.
23:10:44 [mikeU]
MIkeD: we usually export internal references.
23:10:57 [mikeU]
Paul: we did not try to answer this question.
23:11:25 [mikeU]
MikeD: if they are myh IDs Im not woried. If it is someone else's, could be a problem
23:11:58 [AxelPolleres]
The "skolemising bnodes" issue is relevant for chaining SPARQL queries... If i want to use answers of one sparql query in a subsequent query, I run into exactly that.
23:12:03 [mikeU]
DB: answer to some of tehse questions depends on the choice of skolemization algorithms. Properties are different.
23:13:27 [mikeU]
Paul: if someone givesme a document, and get skolemized ids, if I see document again, I have not idea if it is same one. If two people get same document, the skolemization will not merge.
23:14:10 [ekendall]
ekendall has joined #rdfn
23:14:21 [mikeU]
Paul I will geneate separate set of UUIDs for the blank nodes.
23:14:27 [mikeU]
SH: could be different graphs with same name.
23:14:53 [mikeU]
DB depending on skolemization scheme, you could ensure that the algorithm produces same URIs in both cases.
23:15:35 [mikeU]
Ivan: im afraid this will go against one Richard's principles from yesterday.
23:16:12 [mikeU]
?? I understand this is core issue. Need interoperability.
23:16:48 [mikeU]
Paul: I do nto allow skolemization of bnodes, but am constantly be asked for those features. I refuse to do it, not supported in standards.
23:17:38 [mikeU]
Andy: who is doing this? Need concrete examples.
23:18:05 [mikeU]
Paul: i understand that it is possible to refer to a blank node with a label in Jena. A lot of people are using Jena.
23:18:49 [mikeU]
Ivan: my question is, going back to charter text, is it a feature that we say: it will solve, or it will consider to solve. This is a big difference.
23:19:03 [mikeU]
Jeremy : should be 'consider to solve'.
23:20:10 [mikeU]
Ivan: wkth TUrtle syntax, I can say you must do it. In this case, the WG should have authority to say, sorry, we are too scared to try this.
23:21:11 [mikeU]
Paul: simplifying the RDF semantics. Everyone ignores the semantic standard doc. THis needs to be remedied. Too complex, especially with respect to blank nodes.
23:22:51 [mikeU]
Paul: much of what is there could be truncated / much more concise. E.g. remove model theoretic semantics.
23:23:43 [mikeU]
Paul: safe to change semantics since noone understands them. Reason not to do it is no one cares if there is a semantics document that is followed.
23:24:17 [mikeU]
Ivan: if we toss the model-theoretic semantics, we will have to toss at least one document. This is my problem. Not sure if would also require changing 1 or two other standards.
23:25:11 [mikeU]
Stefan D: confusion about language (especially Peter and I). NO need to remove model-theoretic semantics. Just make it more intuitive to follow. No contradiction there. Removing semantics is not a good idea.
23:26:01 [mikeU]
Ivan: I am scared of the consequences. IT is my job. We will have RDF semantics job. If we want to do things right, we will have to open ?
23:26:32 [mikeU]
Paul: if we leave everythign as it is, we base RIF on semantics hat everyone is ignor9ing.
23:27:17 [mikeU]
RC: my preference is not remove model theoretic semantics, but rather to write separate document that maybe is a note saying: here is what you need to know about RDF semantics. This would include information about rules.
23:27:44 [AndyS]
AndyS has left #rdfn
23:27:49 [mikeU]
RC: this is a bit of a copout, but maybe the best we can do.
23:28:24 [AxelPolleres]
better explaining the semantics is definitly a good idea...
23:29:37 [sandro]
mikeU: People not understand the model theory -- is there anything bad that happens? If so, that should be explained by cygri's document.
23:30:21 [mikeU]
MU: is anything bad happening because people do not understand the model theory? IF so, then the document should explain these things. If not, then in practical terms, we never needed the model theory in the first place.
23:31:21 [mikeU]
Paul: reading off sec 3.2.4 about Archaisms.
23:32:33 [mikeU]
Paul: Literals as subjects: We looked at this, ther are many negative consequences, and insufficient compelling use cases to warrant doing anything now.
23:32:39 [mikeU]
DB: I can tell you a use case now.
23:32:45 [mikeU]
SH: add it to the Wiki.
23:33:05 [mikeU]
DB: it was about modeling semantics of http. I needed to be able to talk about URIs as strings, not as RDF: names
23:33:15 [mikeU]
DB: thre are work arounds, but they are inconvenient.
23:33:52 [mikeU]
Paul; it could mean changes to OWL. People would likely start misusing it. Use string "London" as London.
23:34:01 [mikeU]
ELisa; Saw a lot of this at Dublin Core
23:36:19 [cygri]
scribenick: cygri
23:37:25 [cygri]
Topic: Graph Identification and Metadata (Elisa)
23:37:33 [cygri]
23:37:34 [tomlurge_]
tomlurge_ has joined #rdfn
23:38:02 [cygri]
ekendall: we changed name from "named graphs"
23:39:11 [cygri]
(ekendall presents the table from section 3.3.1 Graph Identification)
23:39:37 [Stefan]
Stefan has joined #rdfn
23:41:43 [cygri]
ekendall: bottom line: we think we should do it but a lot of work to be done
23:41:54 [cygri]
... moving on to 3.3.2 Annotations
23:42:15 [cygri]
... we discussed wether it actually should be in the charter, maybe time permitting
23:44:17 [cygri]
pfps: lots of references are to big documents; can you link to sections?
23:45:40 [cygri]
pfps: especially N3, please point out which section are you alluding to?
23:46:42 [cygri]
(discussion about pointing to a version of jeremycarroll's paper that isn't behind ACM paywall)
23:47:41 [pgearon]
pgearon has joined #rdfn
23:48:00 [cygri]
Topic: Linked Data (David)
23:48:20 [cygri]
dbooth: we broke it out into three parts
23:48:31 [cygri]
... 1. codifying follow your nose
23:48:45 [cygri]
23:50:30 [cygri]
ivan: does this mean putting Rec Stamp on httpRange-14?
23:50:35 [cygri]
cygri: among others
23:50:53 [cygri]
ivan: SPARQL group has the problem of being unable to refer to httpRange-14 because it's just an email message
23:51:21 [cygri]
AxelPolleres: we agreed to handle that in SPARQL by informative reference
23:51:30 [cygri]
ivan: ok i understand the motivation
23:51:41 [cygri]
mdean: what's AWWW?
23:51:50 [cygri]
Architecture of the World Wide Web Volume One
23:52:43 [cygri]
dbooth: 2. how to determine the intended referent of a URI
23:53:05 [ekw]
ekw has joined #rdfn
23:53:06 [cygri]
... document the responsibilities of URI owners, RDF publishers, RDF consumers
23:56:46 [cygri]
jeremycarroll: we had text about this in the original RDF
23:57:09 [cygri]
... after major battle the text got cut
23:57:21 [cygri]
dbooth: jeremy can you supply reference to that?
23:57:50 [cygri]
Guus: this is why we separated 1 and 2
23:58:09 [cygri]
... and this one perhaps should not go into core
23:58:37 [cygri]
ivan: this is interesting and exciting (after suitable number of beers) but not topic of discussion for this group
23:58:52 [cygri]
sandro: perhaps better not to do it
23:59:20 [cygri]
DavidW: in terms of charter, perhaps better not
23:59:51 [cygri]
sandro: wether it's in scope also depends on timeline of WG