RDF Core Working Group IRC logs for 2002-02-26

These are the logs from the RDF Core Working Group IRC chat.

Dave Beckett

[07:25:56] DanC
DanC has joined #rdfcore
[07:46:49] libby
libby has joined #rdfcore
[07:57:23] JosD
JosD has joined #rdfcore
[07:58:33] libby
some photos from yesterday, some of the group: http://swordfish.rdfweb.org/photos/2002/02/25/
[08:03:32] JosD
DanC in full (guitar) preparation of the golden triple song
[08:04:57] gk
gk has joined #rdfcore
[08:05:02] gk
The golden triples:
[08:05:02] gk
[08:05:02] gk
[08:05:02] gk
Chords (Spanish Lady):
[08:05:02] gk
[08:05:03] gk
MP3 of Spanish Lady:
[08:05:05] gk
[08:13:44] bwm
bwm has joined #rdfcore
[08:13:44] JosD
1. DaveB's detailed test case for rdfms-nested-bagID
[08:15:05] JosD
[08:17:04] gk
FYI, I18N discussion: logging in #rdfcore-i18n
[08:23:22] DanC
we tried with the W3C validator, and it seemed to agree; 14 triples.
[08:23:49] DanC
But we're not in a good position to decide; it doesn't seem intuitive to us that the nested stuff wouldn't get tagged with "this was said on tuesday".
[08:23:58] DanC
[08:25:10] JosD
testcas 2 http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Jan/att-0113/02-test002.rdf
[08:29:14] DanC
test2 seems straightforward.
[08:29:16] DanC
[08:33:31] JosD
2. Jeremy's XML base test cases (+ve test cases for error2 and error3)
[08:33:31] DanC
[08:35:00] DanC
LMM: yes, re / question, the text of 6.H isn't quite clear...
[08:39:19] DanC
LMM: no, re #frag , doesn't seem consistent with RFC2396
[08:41:02] DanC
ACTION DanC: get "thismessage:" from RFC2557 in the IANA URI scheme list http://www.iana.org/assignments/uri-schemes
[08:42:49] DanC
LMM: consider defining a "thisdocument:" uri scheme which works similarly
[08:45:05] JosD
(e) When the methods above do not yield an absolute URI, a base URL
[08:47:43] DanC
(re / question is test011.rdf)
[08:48:22] DanC
LMM: for self-document referce, perhaps you're not doing relative URI references.
[08:50:34] DanC
DanC: should we have you review the relevant text?
[08:50:47] DanC
LMM: actually, somebody not familiar with the issue should review it...
[08:50:56] JosD
test011 APPROVED
[08:50:57] DanC
... e.g. implementors, during CR.
[08:59:27] JosD
discussing error001
[09:01:30] DanC
-- err002
[09:01:57] DanC
re-reading RFC2396 carefully, it seems that the path component of mailto:foo is empty, not undefined...
[09:02:13] DanC
so "relfile" relative to mailto:foo@bar is mailto:relfile.
[09:02:40] DanC
so RFC2396 does give an answer in this case. perhaps we needn't make it an error.
[09:03:43] DanC
-- back to same-document refs:
[09:05:14] DanC
in 4.2, there's an "If ... always intended to result in a new request"... that can explain why "#foo" can work with xml:base in RDF.
[09:05:52] DanC
-- back to err002, mailto:foo@bar + relfile
[09:05:59] DanC
LMM: 1.4 seems relevant...
[09:07:43] DanC
suggests that relative URIs are only to be used with schemes that are specified as hierarchical.
[09:08:03] DanC
so while it's not an error to do the syntactic composition, users should be warned that this makes no sense.
[09:08:17] DanC
(meanwhile: there are lots of ways to say nonsense in RDF.)
[09:14:23] DanC
re err002, seems we shouldn't make it an error.
[09:18:10] DanC
-- re the same-document case again:
[09:18:56] DanC
perhaps the algorithm for computing the absolute form has another input, in addition to the normal base and the relative URI: an address to use in the "this document" case.
[09:19:14] DanC
this would require enhancing our test infrastructure to take another input.
[09:38:59] gk
gk has quit
[09:40:05] bwm
bwm has quit
[09:55:11] mdean_
mdean_ has joined #rdfcore
[09:57:24] mdean_
test case for nested-bag-ids is approved
[10:03:50] mdean_
ACTION: jeremy review error2 and error3 test cases
[10:05:11] gk
gk has joined #rdfcore
[10:08:54] mdean_
Jeremy: xml:base should apply to rdf:about="" and "#foo", otherwise we should abandon it
[10:11:14] mdean_
DanC: we need to show that developers can read the spec and implement it
[10:15:20] mdean_
RESOLVED: xmlbase/test001 approved
[10:17:06] mdean_
Jeremy summarized internationalization discussion
[10:17:13] mdean_
very successful, didn't get beat up
[10:17:30] mdean_
most interested in equality as a mathematical operation
[10:17:47] mdean_
charmod normalization has changed slightly
[10:18:09] mdean_
string literals should be in normal form C and shouldn't start with a combining character (to allow concatenation)
[10:19:21] DanC
note to self: testing normalization stuff.
[10:19:58] mdean_
want to have IRI's where URIs
[10:20:25] mdean_
3 options
[10:20:47] mdean_
need to define IRI equality for RDF graphs (in order to define tidiness)
[10:22:00] mdean_
Martin wanted to define equality in such a way that preserves the original input string, since that's what the user intended
[10:22:39] mdean_
talked about xml:lang and literal containing Italiano example from yesterday
[10:22:53] mdean_
fr:chat en:chat problem
[10:22:54] DanC
note to self: does this fuzzy matching of IRIs conflict with the URI opacity axiom? hmm...
[10:23:54] mdean_
they were happy for us to define strict equivality as long as we did not mislead application developers that this was the only type of equality
[10:24:04] mdean_
Brian minuted exact wording
[10:24:58] mdean_
RFC 3066
[10:25:12] mdean_
dialects are considered different languages
[10:25:46] mdean_
no new parseType="Literal" issues
[10:27:02] mdean_
began hearing about what internationalization requirements for NTriples might be, and ruled it out of order (as only a test format)
[10:27:56] mdean_
deployment: if we put all of this normalization into our spec, our implementors are dependent upon the XML layers implementing this -- they haven't yet
[10:28:28] mdean_
question about whether we need a normalizing transcoder
[10:29:13] gk
Only for reading from a non-UTF source.
[10:31:49] mdean_
issue of charmod versions -- synchronizing glaciers
[10:32:01] gk
jjc: Who gets to rec first ... the race of the glaciers
[10:32:51] mdean_
IRI comparison: hexify representations and do binary compare
[10:33:51] mdean_
different handling than namespaces?
[10:34:55] mdean_
charmod spec is uncooked
[10:35:47] mdean_
Jeremy: charmod spec is advice to us on how to meet our internationalization responsibility
[10:36:00] mdean_
... which would be there even without charmod
[10:37:59] mdean_
need for waiver from Director on implementing internationalization, if we're depending on features from XML implementors
[10:38:07] mdean_
view that this is unlikely to happen
[10:38:53] mdean_
suggestion to separate IRI and literal normalization
[10:39:14] mdean_
IRI is not yet an IRC
[10:40:17] JosD
i.e. an RFC
[10:40:30] mdean_
yes -- thanks!
[10:41:50] mdean_
discussion about hex encoding
[10:42:11] mdean_
many IRIs correspond to one URI
[10:42:49] mdean_
RDF works in URI space, not IRI
[10:43:20] mdean_
we don't want to have to normalize hostnames, etc.
[10:45:53] mdean_
Jeremy discussed test case with escaped IDs
[10:45:56] mdean_
not approved yet
[10:46:38] mdean_
[10:47:15] gk
Fort information: there is an XML erratum (http://www.w3.org/XML/xml-V10-2e-errata#E26) says that conversion to hexified form should be deferred until the latest possible moment.
[10:47:52] gk
(This bears only indirectly on our discussion, I think)
[10:49:11] mdean_
DanC: XML folks seem to have moved to IRIs and treat the protocols that use URIs as legacy
[10:50:14] DanC
note to self: this (doing RDF in IRI space, as opposed URI space) seems to have architectural impact.
[10:52:36] mdean_
in Jeremy's example, doing it in IRI space produces 2 nodes, doing it in URI space produces 1 node
[11:19:32] logger
logger has joined #rdfcore
Topic now RDF Core WG meeting, Cannes, France 25-26 Feb (telcon 16:00-18:00 UTC each day) - Agenda http://www.w3.org/2001/sw/RDFCore/20020225-f2f/
Users on #rdfcore: logger gk mdean_ JosD libby DanC AaronSw
* DanC thinks folks now know the issue well enough to form an opinion... let's take a straw poll, please.
[11:20:15] libby
we lost some text in the log
[11:20:17] libby
[11:20:26] libby
mdean_: Jeremy: If we just use a bit, we lose round-tripping
[11:20:26] libby
mdean_: DanC: don't kid people into thinking that RDF is extensible in this way
[11:20:26] libby
mdean_: EricP: think of this as encoding rather than parseType
[11:20:26] mdean_
EricP: what is the cost/benefit of having an extensibility mechanism
[11:20:27] libby
[11:20:55] mdean_
DanC: parsers may reach wrong conclusions
[11:24:20] mdean_
straw poll against using this as an extensibility mechanism
[11:25:05] mdean_
GK: XML has a distinguished status w.r.t. RDF
[11:26:47] gk
... other point is that parseType can still be used for defining alternative pure-syntactic structuired like daml:collection that don't have special representation in the graph (i.e. other than triples)
[11:28:50] DanC
note to self: as to actual n-triples, I don't want to use (0, none, "abc") for "abc" with no lang and no Literal... I want to use "abc". with lang="en", it becomes "abc"/en or some such... for parsetype literal, I'd rather something like xmlparse("c14n form"). hmm... XML with lang... hmm.
[11:32:29] mdean_
possibility of grandfathering support for daml:collection in RDF
[11:32:43] mdean_
RESOLVED: treat parseType as a bit
[11:33:19] mdean_
next topic: regarding xml:lang string vs. URI
[11:33:33] mdean_
Jeremy: this isn't our problem
[11:36:03] gk
[11:36:13] mdean_
DanC: regstries are available, but their URIs keep moving around
[11:36:21] mdean_
[11:36:35] mdean_
DanC: cost of doing this now doesn't seem to be worthwhile
[11:36:49] mdean_
GK: IETF proposal to map lang tags to URIs
[11:37:23] gk
(URI is above)
[11:37:26] mdean_
Jeremy: sounds like an RDF 2 issue
[11:38:22] mdean_
EricP: make lang a string now with intent to move to a URI later
[11:38:42] mdean_
ACTION (Brian): add this to postponed list
[11:39:17] mdean_
RESOLVED: string from XML 1.0 (RFC 3066)
[11:39:58] mdean_
not ready to close charmod-literals
[11:40:57] mdean_
Patrick: test case for daml:collection
[11:41:25] mdean_
ACTION (DanC): test case that daml:collection is an error
[11:42:34] mdean_
ACTION (DanC, DaveB): update NTriples document
[11:42:57] DanC
s/daml:collection is an error/is just like Literal/
[11:43:35] mdean_
issue: ask WebOnt whether they would like us to bless/grandfather daml:collection
[11:44:22] mdean_
ACTION (DaveB): update syntax doc
[11:44:55] mdean_
ACTION (Brian): contact Pat about any required updates to Model Theory
[11:45:31] mdean_
Brian: can we close rdfms-xmllang?
[11:45:38] gk
Ohhh.... but (e.g. for DAML) *not* generating a literal for parseType!=literal may not be an error?
[11:46:55] mdean_
TimBL raised rdfms-xmllang
[11:47:19] mdean_
RESOLVED: close rdfms-xml-lang
[11:47:30] mdean_
[11:48:29] mdean_
regarding rdfms-literal-is-xml-structure
[11:49:35] mdean_
MathML example in M&S section 7.5
[11:50:16] mdean_
already in UTF-8
[11:50:31] mdean_
otherwise we'd need a normalizing transcoder
[11:50:46] mdean_
namespaces that aren't used don't get inherited
[11:51:48] mdean_
exclusive and inclusive canonicalization are different in their handling of QNames
[11:53:17] mdean_
have to explicitly say xmlns="" to turnoff namespaces in parseType="Literal"
[11:54:22] mdean_
will get equivalent canonical XML, not necessarily original -- no way to force representation
[11:55:24] mdean_
what about comments, entities, etc.?
[11:59:14] mdean_
ACTION (Jeremy): review canonicalization spec w.r.t. comments and processing instructions and produce proposal
[12:00:10] gk
gk has left #rdfcore
[12:00:26] gk
gk has joined #rdfcore
[12:00:47] mdean_
Jeremy: concerned about overly-onerous impact on DPH
[12:01:48] mdean_
RESOLVED: closed rdfms-literal-is-xml-structure
[12:02:30] mdean_
canonicalization will address rdfms-literal-namespaces -- wait for Jeremy's proposal before closing
[12:02:31] gk
gk has left #rdfcore
[12:02:40] mdean_
lunch now, back in 1 hour
[12:02:41] libby
cvs commit
[12:02:44] libby
[12:03:18] libby
[12:03:29] libby
libby has quit
[12:40:14] DanC
DanC has quit
[12:40:14] JosD
JosD has quit
[12:58:47] mdean_
mdean_ has quit
[13:09:50] mdean_
mdean_ has joined #RDFCore
[13:11:41] JosD
JosD has joined #rdfcore
[13:11:50] libby
libby has joined #rdfcore
[13:13:25] libby
brian would like to look at rdfs assertion beiefly
[13:13:54] libby
[[asaserting anrdf graph amounts to claiming it is true...]] model theory
[13:15:07] libby
[13:15:10] libby
[13:15:21] libby
the web context is taht you publish it and you meab it
[13:15:31] libby
bran: what about saying this is rubbish?
[13:15:54] gk
gk has joined #rdfcore
[13:16:28] libby
brian: 'assrting' an rdf graph - web context you are asserting it, a mail you are not (for example)
[13:16:28] libby
ps: need to define what asserting means
[13:16:28] libby
jjc: about legal stuff
[13:16:28] libby
danc: social implications
[13:16:59] libby
ep: least maount can do that clear that contextual issuesz are fdor the courts to decide?
[13:17:35] libby
ps there could be other rdf statements saying you are nmot asserting that they are true
[13:18:11] libby
brian: need to cature when asserted and when not: not our job
[13:19:01] libby
danc: publishes somethign traditional way in rdf whioch looks like a promise to supply widgets, and then can deny it,then we havce failed.
[13:19:48] libby
ps: lots diff ways to assert stuff on the web. not our job to say so. klegal issue. we're not qualified
[13:20:00] libby
jos: person need to know has such consequences
[13:20:13] libby
ps: societty not the web
[13:20:25] libby
jjc: 2 cases in uk, web-realted
[13:21:00] libby
...1 low price advertised, didnt get goods; 2 ditto, but w an email ...
[13:21:12] libby
...2nd place email closed the contract, and liable
[13:21:23] libby
brian: we are off-course
[13:21:41] libby
jjc: asserting is a legal issue, so resolved in the courts
[13:22:22] libby
...agree w brain that this closes issue
[13:23:03] libby
danc: prefer to say 'for example publishing it on the web with a 200 response code bla bla bla' maybe better in primer
[13:23:23] libby
...we could be back here again if our readers dont agree
[13:24:08] libby
ps: because gl;obal, wont apply everywhere; also make peopel afraid of legal implicationas of rdf
[13:24:22] libby
...just a means for expressing knowldge
[13:24:31] libby
..danc - a contract mechanism
[13:24:33] libby
danc: yes
[13:24:45] libby
[13:25:19] libby
brian: lots of specs in the legal process...
[13:25:30] libby
...dont believe we have to complete entire chain
[13:25:53] libby
...we have to lay a part in making it so directpor can claim that a contract happened
[13:26:09] libby
gk: violent agreemnt! propose we accept dancs proposal
[13:26:32] libby
...'for example by publication on the web with response code 200'
[13:26:46] libby
md: in absense of legal sigs, problem
[13:26:57] libby
danc, no, is http spec, same as http
[13:27:10] libby
ep: avoid 'true'
[13:27:30] libby
ep: ...some text comparing rdf to any other language....
[13:28:24] libby
danc: agree no more (bit also no less) than html
[13:28:38] libby
jjc, ps like ep's text more than danc's
[13:29:55] bwm
bwm has joined #rdfcore
[13:31:50] libby
ep: assertion made in rdf are analgous to assertions made in any other langiuage. the author of these assertions is responsible for tehse assertions but it remains the responsibility of courts to determine the effects of context and other factors.
[13:31:50] gk
For information: If we're going to talk detailed wording, here are the ones I proposed: "RDF is intended to convey assertions that are meaningful to the extent that they may, in appropriate contexts, be used to express the terms of binding agreements."
[13:32:56] libby
(some tweaking)
[13:33:07] libby
(some arguing)
[13:34:01] libby
brain: the director needs something. we nede to have something to show his, and see if efficient
[13:34:09] libby
..will that work danc?
[13:34:23] libby
danc, yep, but also need to get it right
[13:35:28] libby
arguing about publisher
[13:37:25] libby
danc: treat it like text! I cannot change my position
[13:38:11] libby
dave, jos think too weak, ambiguous
[13:38:25] libby
dac, thinks could change 'may be' to 'is'
[13:39:38] libby
jjc: proposes a change 'is responsible for these assertions. It remains...'
[13:40:52] libby
...'it reamins the responsibility of courts to determine legal responsibioity considering the effects of context and otherr factors'
[13:41:14] libby
dave now happy, jos noit quite happy
[13:41:26] libby
danc yjoimks needs to say somethiogn about web servers
[13:42:18] libby
jjc: does think we shoudl untuil/if director says need something about web servers
[13:42:27] gk
I think we're wasting our time here on two counts: (a) the text is being offered in isolation from the document in which it to appear, (b) from my previous work in legally admissable messaging, legal liability is determined far more by custom and practice than anything we may say in a document.
[13:42:44] libby
brian: action (brian) to send this text to timbl and test whether this meets his needs
[13:44:01] libby
---issue rdfs-not-id-and-resource
[13:44:07] libby
[13:45:02] libby
jjc: was a problem when ID didnt mean reification. but now it does, this is fixed.
[13:45:13] libby
[13:45:19] libby
(shortest time ever)
[13:45:28] libby
actions: daveb to update testcases
[13:46:01] DanC
DanC has joined #rdfcore
[13:46:15] libby
(generate or amend testcase, review existing ones)
[13:46:29] DanC
in case Brian sees this: when you send the text to TimBL, he's gonna want to read it in context; i.e. in the context of the primer or the model theory.
[13:46:37] libby
[13:46:47] libby
(universal happiness)
[13:47:39] libby
danc: time to choose one of teh props
[13:47:48] libby
jjc: liked simple 2 better
[13:48:21] libby
brain: titdy contradicts simple datatypes 2.
[13:48:32] libby
jjc would like 3-4 moins to talk about it.
[13:48:38] libby
jjc talks about simple 2
[13:49:37] libby
jjc: the literal "foo" is is replacable in the MT with a bnode _b <rdf:dlex> "foo"
[13:49:53] libby
...completely replaceable with each other, seen one, seen the other
[13:50:01] libby
danc: infinte regress?
[13:50:06] libby
jjc: can sort it if so
[13:50:46] libby
...from tidiness pov, the bnode becomes typed and so is not tidy
[13:51:28] libby
..jjc thought allows SB idom, interworks with SA idiom, and they are genuinely compatible
[13:52:02] libby
danc: convinced to reopen the issue?
[13:52:38] libby
brian: who supports reopening issue
[13:52:45] libby
3 1/2 - not enough
[13:53:11] libby
jos: proposed similar earlier, coinvinced not workable
[13:53:19] libby
- not reopened
[13:53:53] libby
danc has some questions:
[13:54:36] libby
http://.....#decimal - lots people want to use it as a class name; lots as a property name (xsd:decimal)
[13:55:02] libby
gk: ambivilant, cos have to assin afdditional meaning to it in the rdf context not in schema one
[13:55:47] libby
danc: in your favourite design, doy ou want to use this as a prooperty
[13:55:50] libby
[13:56:36] libby
brain realizes what danc is trying to do....
[13:57:08] libby
brian: pat's latest doc has been amended over night
[13:58:20] libby
...who agrees with it?
[13:58:39] libby
..m,artyn yes; miked reservations; jjc, opoosed
[13:59:46] libby
...lukewarm support
[14:00:23] libby
...made all the key deciisons, minor stuff only
[14:00:23] libby
ps would like path to review some things he's found with it
[14:00:36] libby
jjc will bedragged with the group reluctantly
[14:01:12] libby
brian thinks jeremy is they only one who doesnt have what he wants
[14:01:41] libby
brian: do you want to do the process?
[14:02:00] libby
agrement that if danc fails to persuade we revert to the path doc
[14:02:13] libby
danc: 10 mins
[14:02:45] libby
http://....#decimal (xsd:decimal) is a propert name?
[14:02:56] libby
...sometimes people like this, sometime not
[14:03:20] libby
danc: if so
[14:03:45] libby
.. people like: dc:date range date
[14:04:54] libby
[14:05:22] libby
mary age "10"
[14:05:22] libby
film title "10"
[14:05:22] libby
mary age _x
[14:05:25] libby
filem title _x
[14:05:41] libby
about 1/2 each way as a good thuing
[14:06:55] libby
danc is close to abandoning all hope is his method...
[14:07:16] libby
serveral peple withdraw their 'nos'
[14:07:46] libby
danc gives up in despair
[14:07:58] libby
brain: we go to pat's doc then
[14:08:10] libby
danc: what are the big and what small issues for discussion
[14:08:58] libby
danc: ascethetic arguments are no longer in order
[14:12:16] libby
[scribe missed a bit by jjc, sorry]
[14:12:51] libby
ps: woulod like to talk about changing pats proposal - postponed to later
[14:13:20] libby
brian jjc has pointed out that entailment important to us, and path's rules it aout
[14:13:41] libby
...change your mind anyone, re jjc comments (unfixable by minor changes)
[14:14:24] libby
[14:14:48] libby
danc would not want to move forwar in that case, nor grahma, ps or jjc
[14:15:28] libby
jjc thinks possible to solve problem by halving path's proposal and taking half any half!)
[14:15:34] libby
brian thinks this in scope
[14:17:47] libby
brian: propose simple datatypes doc is the basis for our datatypes solution: http://www.coginst.uwf.edu/users/phayes/simpledatatype23-02-2002.html
[14:18:30] libby
...we will only make changes to fix specific problems, with testcases. A problem is something the WG agrees is a problem
[14:19:37] libby
danc: the lack of xml e.g.s in the doc is a problem. but path doesnt, cant do that
[14:19:44] libby
gk does think a problem
[14:19:51] libby
sorry, doesnt think so
[14:20:04] libby
most people think lack of e.gs in xml is a problem
[14:20:30] libby
ps 'volunteers'
[14:20:43] libby
action PS produce these examples
[14:20:59] libby
brian: is jjcs entailment not being entoialed a problem?
[14:21:02] libby
[14:21:17] libby
jjcs entailment si a problem - WG agrees
[14:22:14] libby
going to do this here and on telecon
[14:22:48] libby
ps: problem that it is too complex correctly
[14:22:56] libby
[14:23:15] libby
jjc: let's appreciate patH's work here
[14:24:25] libby
ps: too cumbersome
[14:24:58] libby
brian: not enough support for that as a problem heree, maybe in another few weeks
[14:25:36] libby
mikeD: take out octal and germal decimal are not good in here
[14:25:51] libby
(also as a point xsd:number shopuld be xsd:decimal)
[14:26:39] libby
..no octal concept in xsd, not commas intead of periods in xsd
[14:27:08] libby
agreemnt that this a problem
[14:27:39] libby
gk: question: are we limited to xsd datatypes for the purposes of evaluating this proposal?
[14:30:40] libby
gk: no rational numbers in sd
[14:30:43] libby
[14:31:06] libby
'what we define a datatypeto be'
[14:31:41] libby
jjc: small changes, e.g. examples form xsd unl;ess compelling reason not to
[14:31:56] libby
[14:32:01] libby
[14:32:49] libby
libby has quit
[14:52:01] AaronSw
<Zakim> SW_RDFCore()10:00AM has now started
[14:52:01] AaronSw
<Zakim> +TP-Iles-C
[14:52:11] AaronSw
<Zakim> +AaronSw
[14:53:07] DanC
testing suggests phone working well... low latency, full duplex.
[14:53:41] AaronSw
[14:53:53] AaronSw
<gk> http://ils.student.utwente.nl/main/r/reilly_paddy/spanish_lady.crd
[14:59:43] AaronSw
<Zakim> +Manola
[15:00:59] AaronSw
<Zakim> +Guest P7 7332 - is perhaps PatH?
[15:01:47] patH
patH has joined #rdfcore
[15:03:19] DanC
for W3C process presentation, the materials are: http://www.w3.org/2001/02pd/rec54.ps
[15:03:29] libby
libby has joined #rdfcore
[15:04:59] DanC
I tried to actually hold a design discussion about datatypes, but it became clear after <10min that we were so far from consensus that we just picked pat's most recent as "as good as it gets".
[15:05:34] gk
gk is now known as gk-scribe
[15:19:21] gk-scribe
(Dan is talking through the process for last call, based on the diagram at: http://www.w3.org/2001/02pd/rec54.png)
[15:19:57] gk-scribe
Thinks we have a good chance o shooting directly for proposed rec (i.e. have sufficient interoperable implementation experience.)
[15:21:00] gk-scribe
The next step (to full recommendation) is not so much about technology development as getting consensus from the AC that this is a worthwhile technology to deploy.
[15:22:54] gk-scribe
So now we need to estimate the schedules (overrun from original charter) going forward from here to proposed rec ... and to full rec???
[15:23:32] gk-scribe
As a working group, we need to agree that this is more that just an editorial clarification of the original RDR M&S, which was the original plan.
[15:23:56] gk-scribe
Aaaron: but we could have made some more interesting decisions....??
[15:24:17] gk-scribe
[But the charter limited our scope there - ed]
[15:24:34] gk-scribe
AGREED: out new documents are not just an editorial clarificaTion.
[15:25:32] gk-scribe
Documents are: primer, syntax, model th, schema, test cases, datatypes(?), relationship to other specs.
[15:25:47] gk-scribe
Which of these are REC track?
[15:26:25] gk-scribe
DanC: all but XML relationship.
[15:26:33] gk-scribe
GK: dont think the primer is.
[15:30:40] gk-scribe
AGREED: the primer will be recommendation track, with dissent from 1 (GK)
[15:31:53] gk-scribe
Discussion of datatyping -- will this be rolled into the schema specfication?
[15:32:50] gk-scribe
FrankM: consequence is that maybe cannot use RDF without using _some_ schema features.
[15:33:45] gk-scribe
jjc: support for idea that Datatyping is published as a NOTE; lack implementation experience, treat design as a hypothesis for a design to be used.
[15:34:34] gk-scribe
Jos: entailment rules in current model theory [...] advantage of having in RDFS is a coherent whole.
[15:34:52] gk-scribe
Aaron: if we find problems, who'se responsible to fix?
[15:35:01] gk-scribe
jjc: tougher if a REC
[15:35:31] gk-scribe
DanC: Have sympathy, but there are real dependencies here, so we really need to have this sorted, not just hypothesis.
[15:36:27] gk-scribe
... seems to be some implementation experience with schema.
[15:36:48] gk-scribe
Aaron: may take RDF schema to CR, and have others skip to PR.
[15:37:08] gk-scribe
DaveB: all documents need to be ready by May -- there are to many interdependencies.
[15:37:23] gk-scribe
[May? did we agree that?]
[15:37:39] gk-scribe
bwm: back to datatypes...
[15:38:05] gk-scribe
jjc: unhappy committing datatypes to REC
[15:38:57] gk-scribe
GK sympathy expressed, but DanC made compelling case.
[15:39:59] gk-scribe
AGREED: Datatyping will be in the schema spec ... maybe with parts in model theory???
[15:40:24] gk-scribe
DanC think we can do last cal my May.
[15:41:00] gk-scribe
i.e. beginning of May (May 7)
[15:41:03] gk-scribe
FrankM: if scope creep can be limited.
[15:41:36] gk-scribe
PatH: remember there's lots of conferences, etc., in April
[15:42:01] gk-scribe
Aaron, DanC: would like finished copies of specs to hand out at WWW2002
[15:42:48] gk-scribe
If working draft= document author not embarassed to put name on (may be incomplete)
[15:42:56] gk-scribe
When can primer be ready???
[15:43:15] gk-scribe
Basically, current material, made presentable.
[15:43:24] gk-scribe
FrankM: a week.
[15:43:58] gk-scribe
ACTION: FrankM, primer WD copy ready by next week.
[15:44:28] gk-scribe
Model theory... last-call working draft??
[15:45:11] gk-scribe
PatH: 1 May 2002, Pat Hayes busy in March, relatively fee in April.
[15:46:10] gk-scribe
EricM: If the primer is the document that will integrate the other parts, it will need a _lot_ of work from all WG members to be ready for last-call.
[15:46:50] gk-scribe
Schema, DanBri is not available -- need a plan B.
[15:47:54] gk-scribe
ABove... *if* DanBri is not available...
[15:48:04] gk-scribe
... and isn't in current call.
[15:48:16] gk-scribe
PatrickS, if people will "look over my shoulder" I'll be plan B editor for Scema.
[15:48:54] gk-scribe
We need a woring draft of schema (as well as Primer) *before* last call.
[15:49:34] gk-scribe
Ask PatrickS to come back with a higher confidence answer in a week.
[15:50:00] gk-scribe
So, for the time being, plan is to have everything to go by 1 May 2002.
[15:51:01] gk-scribe
==== Note target above ====
[15:52:02] gk-scribe
XML relationships NOTE, not same deadline. Words might appear in the primer.
[15:53:58] gk-scribe
[15:54:02] gk-scribe
After May...
[15:54:55] gk-scribe
DanC: It will be too soon to disband the WG, so when should we? conventional wisdom = REC+6 months
[15:55:26] gk-scribe
Will also deliver list of suggestions of what to do next.
[15:57:19] gk-scribe
Suggested we'll hang around for 6 months without necessarily meeting regularly.
[15:57:27] gk-scribe
Seems like general agreement.
[15:59:35] gk-scribe
How long for last call? Aim for 3-week response period, all responses by end of May - need F2F to resolve difficult comments.
[16:00:32] gk-scribe
(If no issues, we have F2F for a big party ;-)
[16:03:06] gk-scribe
Aiming for mid-June (maybe around Sardinia meeting?)
[16:08:14] gk-scribe
AGREED: plan to meet in Europe, dates around the Sardinia conference
[16:09:29] gk-scribe
Bristol is preferred to Pizza, for many.
[16:10:08] DanC
[16:10:31] gk-scribe
ACTION: bwm to specify precise dates and sort hosting details.
[16:11:11] gk-scribe
Schedule another event, later F2F ??
[16:12:05] gk-scribe
... not a WG activity? Workshop as oriming activity for follow-on work. People may want to think iof they're available for more
[16:12:18] gk-scribe
ACTION: bwm, consider arrangements for RDF-2.0 workshop.
[16:13:50] AaronSw
aw, no singing
* AaronSw starts up "Spanish Lady" on Karaoke mode
[16:14:13] gk-scribe
Aaargh! We had it all lined up.
[16:14:30] gk-scribe
End of session.
[16:14:36] gk-scribe
[16:14:43] gk-scribe
gk-scribe is now known as gk
[16:16:11] AaronSw
Whack for the toora loora laddie
[16:16:11] AaronSw
Whack for the toora loora lay.
[16:18:22] AaronSw
Well, we'll have to have at least a get-together at Hawaii.
[16:18:38] DanC
you plan to come to Hawaii? I guess you said that.
[16:19:11] AaronSw
Yeah, I spent my one travel opportunity on it
[16:19:16] AaronSw
I want a F2F I can actually be at, gosh darnit.
[16:22:29] AaronSw
Tho I did find someone who'll pay to fly me out to California for some meetings.
[16:22:51] gk
gk has left #rdfcore
[16:23:31] patH
patH has quit
[16:37:03] bwm
bwm has quit
[16:41:15] AaronSw
We'll have to save the singing for our next telecon.
[16:46:05] mdean_
mdean_ has quit
[17:02:36] libby
libby has quit
[17:52:39] JosD
JosD has quit
[19:42:43] DanC
DanC has quit
[22:41:34] DanC
DanC has joined #rdfcore
[23:17:12] DanC
DanC has quit

Provided by Dave Beckett, Institute for Learning and Research Technology, University of Bristol