See also: IRC log
<trackbot> Date: 29 February 2012
<ivan> Zakim has a leap year bug… it does not recognize the 'this will be XXX' command. But if dialing in with the code, it will start (at least that is what our system guys told me…)
<AndyS> New UC --: <rdf-wg> :teleconDate [ :happendOn "2012-02-29"^^xsd:date ] .
<davidwood> PROPOSED to accept the minutes of the 22 Feb telecon:
<davidwood> http://www.w3.org/2011/rdf-wg/meeting/2012-02-15
<scribe> scribenick: ericP
<PatH> thanks eric
<davidwood> RESOLVED to accept the minutes of the 22 Feb telecon
<davidwood> Action item review:
<trackbot> Sorry, couldn't find user - item
<davidwood> http://www.w3.org/2011/rdf-wg/track/actions/pendingreview
<davidwood> http://www.w3.org/2011/rdf-wg/track/actions/open
antoine: ACTION-149 in
progress
... will add references to related URLs
sandro: i think you should focus on the 5 use cases on @@1
<sandro> http://www.w3.org/2011/rdf-wg/wiki/Why_Graphs
antoine: i tried to focus on easy use cases listed as priority A on the use case page @@2
<AndyS> ?? Sandro says they are his personal choice ??
s/@@1/http://www.w3.org/2011/rdf-wg/wiki/Why_Graphs
s{@@1}{http://www.w3.org/2011/rdf-wg/wiki/Why_Graphs}
<sandro> AndyS, I just didnt want to presume anything from the WG at that point.
<davidwood> Close ACTION-148
<trackbot> ACTION-148 Write email about the "islands" idea closed
<davidwood> Close ACTION-141
<trackbot> ACTION-141 Propose a vocabulary and semantics for adding timestamps to RDF graphs closed
pat: re: ACITON-140, agree that
only XSD 1.1 semantics implications are more valid forms, and
that we refer to xsd:identity, not xsd:equality
... do we want to incorporate facets?
sandro: like from OWL2?
<AlexHall> I thought Pat's action was to review the WG note: http://www.w3.org/TR/swbp-xsch-datatypes/
<FabGandon> http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-xml/index.html
http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/index.html#sec-strings
<AndyS> Does not allow \u in strings?
<cygri> +1 to 4.4, it looks good
<sandro> cygri: Why not name the different escape mechanisms, then use those names when saying which kind is allowed where. eg "Local Name Escaping", etc.
<sandro> ... "Magic Special Character Escaping" is \t stuff. etc. Then say which you can do each place.
<sandro> +1 cygri
<sandro> ericP: Sounds reasonable.
<sandro> ... I'll present two options.
cygri: editorial proposal: name the escaping mechanisms and reference the names in e.g. IRI, local names, strings
<cygri> +1 sandro, to giving *some* explicit name to these escape mechanisms also makes talking about this stuff much easier
PROPOSAL: remove section 4.3 in <http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/index.html#sec-strings> and remove issue tag on 4.4
<sandro> sandro: if you can't come up with good names for the types, then just use type-1-escaping, type-2-escaping, etc. And have a table of which works where.
<sandro> +1
<ivan> +1
<AndyS> +1
<davidwood> +1
<cygri> +1
<MacTed> +1
+1
<PatH> +1
<zwu2> +1
<pchampin> +1
<pfps> +1
<Arnaud> +1
<yvesr> +1
<AlexHall> +1
<Souri> +1
http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/index.html#sec-escapes
<cygri> use both examples.
<pfps> +1 to previous proposal (dropped because my wireless dropped)
<pchampin> I too would keep both examples.
<sandro> sandro: Also, and say it's different from /Foo-bar
<sandro> davidwood: Good idea.
<sandro> ericP: Note is changed; I'll do the naming the different forms of escaping for tomorrow.
PatH: do graph labels mean the
same thing in different data sets?
... these questions affect other decisions
davidwood: big ups to my bro PatH for forcing the WG to adopt philosophical rigor in our terminology
,me , yo
<davidwood> Subtopic: Islands 🌴
<pfps> ?? to David's channelling of Andy
davidwood: in AndyS's island terminology, two graphs may use a uri in the 4th slot to mean different things
AndyS: the island idea was that there's a *social* step of assembling the data, and we're interested in that step
PatH: does it have technical consequences?
<davidwood> Pat said, "What [2] says is that *the same* URI when used in one graph can mean something completely different when used in another graph, and that *this is perfectly correct* and even in fact *consistent*. What this means is that every URI in every graph is interpreted locally to that graph, which in effect makes every URI into a blank node (since this is how blank nodes are interpreted.) This is dissolving the entire Web in a kind of universal solvent." (w
PatH: must the machinery support
it?
... the obvious consequence is that the machinery that we have
shouldn't force false consequences
<PatH> that was andy
<PatH> btw, that mention of bnodes was wrong, I acknowledge.
davidwood: in linked data, we say that using the same IRI in different places means the same thing.
AndyS: i think trying to address
this technical hampers progress because we assume there's a
single technical solution
... saying that "there's good practice and bad practice" is a
weaker way forward
PatH: what "less ambitious"
solution do you have in mind?
... there are datasets out there, people use them in various
ways
AndyS: we have claimed we will
solve some use cases
... maybe that's too ambitious
... we don't want to solve those use cases, we want to give
folks machinery to solve those use cases
... we just have to not mess them up
davidwood: separating tech from social issues is helpful
<sandro> http://www.w3.org/2011/rdf-wg/wiki/Why_Graphs
sandro: in editing Why_Graphs
this AM, these are all things that people want to
standardize
... i don't think you can do any of these 5 things without
defining semantics to trig
... how can i do <shared web crawler> and <archiving
web crawler> in the same world?
<PatH> +1 sandro.
sandro: all use trig as a simple
and obvious solution but they use it with different
semantics
... could scribble into trig some info about how it's being
used
... i think 2 and 3 share a semantics, and 4 just says you need
bnodes shared between graphs
... 3 semantics:
... .. relation between name and graph is log:semantics
... .. relation between name and graph is owl:sameAs
... .. no relation between name and graph
<sandro> <a> owl:sameAs { <s> <p> <o> }
<zwu2> sounds a bit odd :)
ivan: so for owl:sameAs, an RDF graph becomes an OWL individual
MacTed: when i say "Joe" and i
say <http://joe.example/#> i mean
the same joe
... when you say "Joe" you mean Joe Jr.
<sandro> pfps, in OWL2 is an owl:Ontology a kind of individual?
MacTed: likewise when you say <http://joe.example/#> you mean Joe Jr.
<pfps> sandro, you just asked a question that does not really have an informal answer
<sandro> pfps, do you agree with AZ? (No in Direct Semantics, Yes in RDF-based semantics) ?
ericP: when i write a language, i expect the compiler to give me errors
PatH: you can expect to discover
inconsistency
... the role of semantics is to deliver to you the fact that
there's an abberation
MacTed: i have the impression that folks are trying to prevent such abberations from occuring
PatH: i don't think that [preventing abberations] can be done
<AZ> isn't TriG just a syntax?
<pfps> sandro, in OWL 2 Direct Semantics owl:Ontology does not appear
PatH: if an IRI is used as a graph label in a trig doc, and used as a label for another graph in another doc, is it the same thing?
AndyS: i've never seen it meant
to label different things
... if it's contradictory, you've screwed up
sandro: i think the labeling
relation is different
... (generally)
AndyS: if you allow that, you'd allow it to be a partial description
<AZ> +1 ericP
<Souri> I agree with MacTed: aberrations can happen within a single graph or when you merge multiple graphs ... a validation procedure can point it out ... that's all we can do
ericP: i think the common usage
of SPARQL gives the answer here
... and that we should say that the triples associated with a
label mean that labeled graph has at least those triples
zwu2: can we treat quads as
syntactic sugar for reification
... ?
sandro: if we figure out how we want to use quads, we can derive how that looks in triples
PatH: +1 to sandro
... treating it as reification isn't consistent with this loose
merging story, at least not without some work
<sandro> sound problems
<PatH> hiccups?
pchampin: we are talking about what we want to do with a trig file rather than what it means
<PatH> speak veeerrry slowly
<Arnaud> phone over ip or cell phone?
<sandro> pchampin, SORRY but we really can't understand you.
<pchampin> I think I'm gone sorry
pchampin, can you write your point?
<pchampin> pchampin: shouldn't we focus on what a trig file *means*, rather than what we will *do* with it?
davidwood: we've come back to this point several times: specifically, what do we want to do with a quad
<pchampin> pchampin: the question might be harder to answer, but then the answer to the other questions would follow
davidwood: is it fair to say that these five use cases are divergent enough to keep us from making a solution?
<sandro> http://www.w3.org/2011/rdf-wg/wiki/TF-Graphs-Designs
sandro: in the Why_Graphs, if you
use trig in the intuitively obvious way, you end up with
inconsistent uses
... but you can use it consistently in any of the 6 ways used
on http://www.w3.org/2011/rdf-wg/wiki/TF-Graphs-Designs
AZ: trig is just a syntax
<AndyS> The SPARQL dump/restore is priority A in the original list (#1.5)
AZ: using trig solves the problem
at a syntactic level
... we can could agree on the syntax first, and then derive the
semantics
<AndyS> not dropped from my end :-(
<sandro> I think it's fine to use TriG as a default/example format for now.
ericP: is there a clear winner on http://www.w3.org/2011/rdf-wg/wiki/TF-Graphs-Designs ?
sandro: i pushed for #3
... AndyS and steve pushed back because of scale implementation
costs
... i now like 6, but need to work out tech details
<sandro> NOT "scale". Steve's concern was a CONSTANT time cost.
<sandro> (or linear)
PatH: re: picking a syntax and
deriving the semantics, these semantics are
interchangable
... in trig, there are cases where the 4th column is a label
for something like time
... thinking of these things as graphs doesn't fit with all of
the intuitions we want to express
AZ: +1 to PatH
<AndyS> That wasn't my concern. 1) feels too much like complete replumbing to the deployed data web. (2) explaining it to people unless we define graph literals in Turtle.
AZ: with one syntax, we have a
way of express stuff that we other way we can't express
... we could standardize trig and nquads
<sandro> sandro: TriG and N-Quads prejudice against solutions like Graph Objects (N3).
sandro: those prejudice against solutions like Graph Objects (N3)
<PatH> Trig prejudices against 4th field as a time parameter.
<sandro> Andy: as a group, I don't think we're sharing and building on each other's ideas
<zwu2> honestly, if we cannot reach a decision, then we are not ready
or it's too late
or it's hard
<PatH> i feel like I am making progress, for one.
<AndyS> Several solutions are great, if it were a clean sheet of paper.
<zwu2> bye
<PatH> bye
<davidwood> http://www.w3.org/2011/rdf-wg/wiki/Scribes
This is scribe.perl Revision: 1.136 of Date: 2011/05/12 12:01:43 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) WARNING: Bad s/// command: s/@@1/http://www.w3.org/2011/rdf-wg/wiki/Why_Graphs Succeeded: s/OWL/OWL2/ Succeeded: s/Black Hole/Named Graphs/ Found ScribeNick: ericP Inferring Scribes: ericP WARNING: No "Present: ... " found! Possibly Present: AZ AlexHall Andy AndyS Arnaud David_Todd FabGandon Guus Ivan LeeF MacTed OpenLink_Software P14 P2 P30 P32 P5 P8 PROPOSAL PatH Sandro Souri Subtopic antoine bhyland cygri davidwood ericP joined mhausenblas pat pchampin pfps rdf-wg scribenick swh trackbot yvesr zwu2 You can indicate people for the Present list like this: <dbooth> Present: dbooth jonathan mary <dbooth> Present+ amy WARNING: No meeting chair found! You should specify the meeting chair like this: <dbooth> Chair: dbooth Found Date: 29 Feb 2012 Guessing minutes URL: http://www.w3.org/2012/02/29-rdf-wg-minutes.html People with action items:[End of scribe.perl diagnostic output]