W3C RDF Core Working Group IRC Chat Logs for 2002-05-24

13:01:06 Users on #rdfcore: @logger_1

13:17:48 <DaveB> DaveB has changed the topic to: RDF Core http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002May/0086.html

13:28:56 <dajobe_-l> dajobe_-l is now known as DaveB-lap

14:00:38 <bwm> hi Danbri

14:00:58 <danbri> hi

14:01:13 <DanCon> hi

14:01:28 <bwm> Should be getting started - need to interrupt DaveB and JJC

14:01:28 <danbri> no zakim?

14:01:39 <bwm> apparently not - anyone know the magic

14:01:46 <danbri> zakim, who is here?

14:01:47 <Zakim> sorry, danbri, I don't know what conference this is

14:02:00 <danbri> zakim, this is RDF_Core

14:02:02 <Zakim> sorry, danbri, I do not see a conference named 'RDF_Core'

14:02:08 <danbri> zakim, this is RDFCore

14:02:09 <Zakim> sorry, danbri, I do not see a conference named 'RDFCore'

14:02:16 <danbri> er, anyone know the voodoo?

14:02:45 <DanCon> Zakim, list

14:02:46 <Zakim> I see SW_RDF Cor()10:00AM, WAI_EO()8:30AM, QA_QAWG()9:00AM

14:02:51 <DanCon> Zakim, this is rdf

14:02:53 <Zakim> ok, DanCon

14:03:07 <DaveB-lap> DaveB-lap is now known as jang-scri

14:03:13 <bwm> -open

14:03:14 <Jema> RDFCore WG Telecon 2002-05-24 is now open

14:03:14 <Jema> The agenda can be found at http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002May/0086.html

14:03:14 <Jema> Agenda item 1: Volunteer scribe

14:03:15 <danbri> zakim, who is here?

14:03:16 <Zakim> I see ??P16, ??P5, PatH?, DanBri, ??P22, Manola, ??P20

14:03:39 <jang-scri> from bristol: dave b, jjc, bwm, jang

14:03:52 <jang-scri> danbri here, bwm here, em reg

14:04:02 <jang-scri> dave b here, frank b abs, jjc here, danc reg

14:04:15 <jang-scri> rond bs, bill abs, jos here, rael abs

14:04:27 <jang-scri> arno abs, jang here, yoshi abs

14:04:38 <jang-scri> gk here, mi k abs, kwon abs ora abs

14:04:51 <jang-scri> frankm here satoshi abs, steve p here

14:05:03 <Zakim> +MDean

14:05:06 <jang-scri> pierre r abs, anderew s abs

14:05:28 <Zakim> +DanC

14:05:32 <jang-scri> patrick s reg, aaron abs,

14:05:50 <jang-scri> mike d here, guha abs, pat h here

14:05:56 <jang-scri> sergei late(!)

14:06:15 <jang-scri> agenda: comments?

14:06:17 <jang-scri> no aob

14:06:27 <jang-scri> bwm: aob from me

14:06:34 <jang-scri> from mailing list, martin h is noon longer working for profium

14:06:51 <jang-scri> thanks to marthin for all efforts

14:06:57 <jang-scri> on behalf of whole WG

14:07:02 <DanCon> RESOLVED: to thank MartinH for his fine work.

14:07:06 <jang-scri> next telecon: same time next week

14:07:08 <DanCon> er... custom here is:

14:07:15 <jang-scri> comments on minutes?

14:07:16 <DanCon> - decision: to thank MartinH for his fine work.

14:07:16 <Jema> command not recognised

14:07:16 <jang-scri> approved

14:07:26 <jang-scri> completed actions: done

14:07:27 <DanCon> -decision: RESOLVED to thank MartinH for his fine work.

14:07:31 <jang-scri> withdrawn action: ok

14:07:44 <jang-scri> outstanding issues:

14:07:47 <jang-scri> item 9

14:07:55 <jang-scri> DanCon: comment on item 8

14:08:01 <Zakim> +??P25

14:08:11 <jang-scri> faq-html-compliance: daveb adds as aob

14:08:20 <jang-scri> danbri: aob on isdefinedby

14:08:35 <jang-scri> item 9

14:08:40 <jang-scri> -item 9

14:08:41 <Jema> command not recognised

14:08:53 <jang-scri> charmod comments

14:09:05 <jang-scri> (that's item 10)

14:09:24 <jang-scri> jjc: charmod doesn't have an escape clause for ntruiples

14:09:28 * DanCon feels left behind on charmod. oops.

14:09:31 <Zakim> +PatrickS

14:10:09 <jang-scri> dave didn't want to overemphasise ntriples

14:10:42 * jang-scri has a slightly bouncy keyboard, apologies in advance

14:10:54 <jang-scri> jjc proposes: he wants to send off current version

14:10:58 <jang-scri> DaveB: I've looked at it

14:11:07 * DanCon q+

14:11:09 <jang-scri> apart from remooving ntriples references, that's all I have to say about this

14:11:09 * Zakim sees DanCon on the speaker queue

14:11:18 <jang-scri> patH: I've looked at it, it looks ok

14:11:39 <jang-scri> danc:

14:12:06 <jang-scri> does this send us in on the adnre/ vs. andr%xx case?

14:12:21 <jang-scri> jjc: we agree, but we may not currently conform, makes no guarantee to conform

14:12:28 <jang-scri> they _should_ conform at the next revision

14:12:38 <jang-scri> we can ignore this on the basis of time dependencies

14:12:46 <jang-scri> we'd be bound in the next round

14:12:52 <jang-scri> the IRI we disagree with so we're not bound by

14:13:05 <jang-scri> we are saying: charmod MUST NOT depend on IRI

14:13:51 <jang-scri> pointer to jeremy's message in agenda

14:14:01 <danbri> (upcoming for later agenda item) my closure-text msg on isDefinedBy arrived: http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002May/0102.html

14:14:25 <jang-scri>http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002May/0080.html

14:14:35 <jang-scri> propose: comments approved, to forward to i18n wg

14:14:44 <jang-scri> -action jjc/ sumit comments to i18n wg

14:14:45 <Jema> Jema notes action 2002-05-24#1

14:14:49 <jang-scri> next item

14:14:54 <jang-scri> daml:collection

14:15:16 <jang-scri> DanCon: I'm uninspired on this regard, can't see options that I think wil get consensus

14:15:30 <jang-scri> options rundown:

14:15:35 <jang-scri> eg, daml:collection is good syntax

14:15:51 <jang-scri> eg, daml:collection isn't good, but you could write it out longhand

14:15:54 <jang-scri> (webont didn't like that)

14:16:04 <jang-scri> eg, rdf:_n plus maxindex

14:16:13 <jang-scri> as implementor, don't like those (jos didn't like that either)

14:17:18 <jang-scri> daml:colllection - daveb:

14:17:28 <jang-scri> major parser implementors are doing it

14:17:36 <jang-scri> so it's not that much of a problem

14:17:56 <jang-scri> DanCon: that sounds most promising

14:18:11 <jang-scri> jjc: daml:collection has implicit unique properties

14:18:21 <jang-scri> I'd be unhappy with that

14:18:41 <jang-scri> DanCon: subpropertying rdf:first wouldn't work.

14:18:52 <jang-scri> the important thing is that the syntax spec gives you the right triples

14:19:08 <jang-scri> DaveB: there are downsides for explaining how to write this

14:19:32 <jang-scri> DanCon: the main objection seems to be, I might use a non-daml:collection-aware parser and lose

14:19:42 <jang-scri> rond: I may have been the one who raised this

14:20:03 <jang-scri> I'm not going to throw a hissy fit on this

14:20:17 <jang-scri> DanCon: is there sufficient impl experience to say this is a sound design?

14:20:24 <jang-scri> jos: I also like this

14:20:39 <jang-scri> DaveB: we've also said we're not designing new rdf containers.. this is new.

14:20:50 <jang-scri> bwm: webont need this closure capability

14:20:57 <jang-scri> DanCon: they also need syntactic sugar

14:21:07 <jang-scri> bwm... they're major customers of ours

14:21:11 * DanCon thinks jos had something to say

14:21:17 <jang-scri> if we can close our current containers, that is an acceptable way to go

14:21:29 <jang-scri> what I'm hearing is that isn't acceptable. Would webont accept it?

14:21:53 <jang-scri> DanCon: the people who would care about this don't care too much about the rdf issue.

14:22:08 <jang-scri> path: what's the objection to rdf:last or similar

14:22:16 <jang-scri> DanCon: it's simple to use first and rest

14:22:29 <jang-scri> I've tried to write this a few times and I can't figure it out

14:22:38 <jang-scri> I don't want to count the number of things in the collections, certainly

14:22:52 <jang-scri> there's a lot of experience with first/rest, that's generally the way it's done

14:23:15 <jang-scri> jos: I strongly support daml:collection, I can make test case for this

14:23:42 <jang-scri> bwm: so, choice 1, close current containers. jos, danc say implementation experience indicates this won't fly

14:24:11 <jang-scri> DaveB: this is creating an endmarker?

14:24:20 <Zakim> + +1.650.965.aabb

14:24:32 <jang-scri> rdf:_18 rdf:_last won't work

14:24:55 <jang-scri> DaveB: benefits: no new syntax, existing systems can use this

14:24:58 <DanCon> I was thinking of _:myCol rdf:lastIndex rdf:_6.

14:25:04 <jang-scri> jjc: I don't see the semantics of this as well-defined

14:25:28 <jang-scri> to patch it with a change to something else we don't understand, i feel, is incoherent

14:25:40 <jang-scri> path: the current container ssemantics is simple: ie, it's emptry

14:25:43 * danbri not sure how all this works if a partial description of a container covers some, including the last, members of a container.

14:25:53 <jang-scri> DanCon: webont: the requirement is this: that you can tell bob isn't in bag X

14:26:35 <jang-scri> DanCon: the alstindex doesn't satisfy the webont needs

14:26:56 <jang-scri> DaveB: another option is to do nothing. we decided no to change the container syntax,

14:27:00 <jang-scri> webont rejected this

14:27:16 <jang-scri> (guha, pat s, also on call)

14:27:51 <jang-scri> bwm: can someone put together a proposal that closes the current mechanism to take to webont so that they can either accept or reject it?

14:28:13 <jang-scri> path: webont wants us to do their work for it - -why can't webont create it's own containers?

14:28:17 <jang-scri> DanCon: we own the syntax.

14:28:23 <jang-scri> that is, owl uses rdf/xml

14:28:46 <jang-scri> in other words, webont want first/rest, they just don't want to write it out longhand

14:28:59 <jang-scri> bwm: who would object to blessing daml:collection?

14:29:33 <jang-scri> that is, add it to the language and say how it generates the triples

14:30:13 <danbri> 'blessing', 'grandfathering'... we talk in funny metaphors sometimes.

14:30:15 <gk> Why would current RDF users object?

14:30:28 <jang-scri> DanCon: we'd just s/daml/rdf

14:30:50 <jang-scri> there are lots of current files that use daml:collection spelt d-a-m-l

14:30:58 <danbri> DanCon -- we don't need daml.org ns, there's a w3c Note for daml+oil. doesn't that use a w3.org ns?

14:31:12 <jang-scri> it wouldn't be hard to change those to rdf:collection

14:31:32 <jang-scri> miked: I'd prefer to say parsetype=daml:collection then this is what we do.

14:32:00 <jang-scri> path: so do we want first/rest PLUS the old mechanism, with no relation between them>

14:32:14 <jang-scri> miked: we lose benefits if we change namespace

14:32:24 <jang-scri> DanCon: we're not going to keep them in the daml namespace

14:32:36 <jang-scri> DaveB: owl will force namespace changes already

14:32:47 <jang-scri> DanCon: it takes 1 line of perl

14:33:15 <jang-scri> bwm: the other way to follow this route would be to say, it's not part of RDF, preprocess before you hit the rdf mechanisms

14:33:47 <jang-scri> frank M: we need to revisit our premises: if webont want this change made...

14:33:59 <jang-scri> they must've considered the ramifications of the syntax change

14:34:15 <jang-scri> frankm: we're hearing a whole lot of concern about things that must be changed if this is done

14:34:23 <jang-scri> but the thing that started this is the desire for the change

14:34:36 <jang-scri> rond: my understanding is that webont needs functionality onot in the original spec

14:34:42 <jang-scri> s/webont/daml originally

14:34:54 <jang-scri> question: why is it a problem for us to acknowledge existing practice?

14:35:13 <jang-scri> I personally don't have a problem blessing a certain number of items out of someone elese's namepspace.

14:35:20 <jang-scri> it;s not a huge change, but why make people do it?

14:35:30 <jang-scri> jjc: I can't support blessing d.a.m.l:collectgion

14:35:42 <jang-scri> because daml NS is going to have a fairly limited lifespan

14:35:50 <jang-scri> OWL is oging to supercede daml

14:35:55 <gk> (1) changing the surface syntax: is this really an issue - existing processors can still accept non-RDF syntax.

14:36:00 <jang-scri> daml docs will be upgraded to owl soon

14:36:30 <jang-scri> DanCon: if we wanted to normally reference daml.org from our specs we'd be scuppered... unprecedented, would add huge delays

14:36:34 <gk> (2) changing the property namespaces: can daml:first, etc. be subproperties of rdf:first, etc??

14:36:57 <gk> ... then I think the bulk of existing data can still be usable.

14:37:01 <DanCon> the one liner: perl -pe 's/daml:collection/collection/g' *.daml

14:37:12 <jang-scri> bwm: it seems that belessing daml:collection in some form is the chosen solution

14:37:14 <jang-scri> (in some form)

14:37:25 <jang-scri> can someone stick up a hand to volunteer for this?

14:37:26 <gk> Dan, there may be 1000's of files to be changed in this way?

14:37:40 <jang-scri> p-action danc/to write up a proposal for incorporating daml:collection into rdf

14:37:46 <jang-scri> -action danc/to write up a proposal for incorporating daml:collection into rdf

14:37:47 <Jema> Jema notes action 2002-05-24#2

14:37:54 <jang-scri> -action josd/ to produce test case for the same

14:37:55 <Jema> Jema notes action 2002-05-24#3

14:38:08 <jang-scri> item: is defined by (item 9)

14:38:26 <jang-scri> danbri:

14:38:31 <DanCon>http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002May/0102.html

14:39:41 <Zakim> +AaronSw

14:39:47 <jang-scri> discussion of "informal semantics"

14:40:14 <jang-scri> frankm: the issue I have is that while there may be no special MT semantics for these things...

14:40:23 <jang-scri> i have to say in the primer what this means

14:40:38 <AaronSw> sorry i'm late

14:40:43 <bwm> k

14:40:58 <jang-scri> the stuff that we put in our resolutions about the MT not saying things...we need to round this out

14:41:20 <jang-scri> frankm: I'd like similar english to decide rdf:idb and rdf:value

14:41:24 <DanCon> "rdfs:isDefinedBy relates a property/class to a document that specifies the property"

14:41:45 <AaronSw> s/property\/class/resource/

14:41:47 <jang-scri> proposal: to add the above words to danbri's resolution

14:41:55 <DanCon> ok, resource.

14:42:17 <DanCon> "rdfs:isDefinedBy relates a resource (e.g. property or class) to a document that specifies the resource"

14:42:59 <jang-scri> path: I'd love to know what that means

14:43:18 <jang-scri> DanCon: eg, people might extract stuff from specs

14:43:31 <jang-scri> and then point back to those specs

14:43:56 <danbri> q+

14:43:57 * Zakim sees DanCon, Danbri on the speaker queue

14:43:59 <jang-scri> Guha_: concrete example

14:44:11 <jang-scri> let's say we're using uris for states in the us

14:44:20 <danbri> concrete example: http://purl.org/dc/elements/1.1/title --isdefinedBy--> http://purl.org/dc/elements/1.1/

14:44:32 <jang-scri> and there's a geographical DB that says what the states are, how they're bounded

14:44:36 <bwm> ack danbri

14:44:37 <jang-scri> I'd like to connect these two things

14:44:37 * Zakim sees DanCon on the speaker queue

14:45:16 <jang-scri> we want to link, in danbri's example, from a property to the namespace/schema that defines it

14:45:45 <jang-scri> DanCon: Ii thought the original WG wanted to point to either formal or informal definitions

14:45:49 <DanCon> stop saying "to the namespace", danbri. makes me twitch

14:45:50 <bwm> ack dancon

14:45:51 * Zakim sees no one on the speaker queue

14:46:34 <jang-scri> path is worried aobut what a machine processor is supposed to make of this

14:46:48 <gk> PatH: What does a machine processor do about it???

14:46:56 <jang-scri> path: someone might do that intended to say something is meaningful, but actually make a mistake

14:47:06 <jang-scri> DanCon: this sin't about mistakes, it's about informal semantics

14:47:22 <gk> Maybe, it attempts to dereference the URL, and IF it gets an RDF document looks for statements about the resource rel;ated by isdefinedby??

14:47:40 <danbri> see the old RDFS CR doc, http://www.w3.org/TR/2000/CR-rdf-schema-20000327/#s2.3.5 for more text on this property.

14:47:42 <jang-scri> danc gives example: colour with informal semantics (the apple is red) or formal (pantone spec)

14:48:05 <danbri> [

14:48:06 <danbri> The most common anticipated usage is to identify an RDF schema, given a

14:48:06 <danbri> name for one of the properties or classes defined by that schema.

14:48:06 <danbri> Although XML namespace declarations will typically provide the URI where

14:48:06 <danbri> RDF

14:48:06 <danbri> vocabulary resources are defined, there are cases where additional

14:48:08 <danbri> information is required.

14:48:11 <danbri> ]]

14:48:18 <danbri> [[

14:48:19 <danbri> For example, constructs such as

14:48:19 <danbri> <rdfs:subPropertyOf

14:48:19 <danbri> rdf:resource="http://purl.org/dc/elements/1.0/Creator"/>

14:48:21 <danbri> do not

14:48:22 <danbri> indicate the URI of the schema that includes the vocabulary item

14:48:24 <danbri> Creator (i.e., http://purl.org/dc/elements/1.0/).

14:48:26 <danbri> In such cases, the rdfs:isDefinedBy property can be used to explicitly

14:48:28 <danbri> represent that

14:48:31 <danbri> information. This approach will also work when the URIs of the namespace and its

14:48:32 <danbri> components have no obvious relationship, as would be the case if they

14:48:32 <jang-scri> guha prefers "for human consumption" versus "informal"

14:48:34 <danbri> were identified using schemes such as GUIDs or MD-5 hashes.

14:48:37 <danbri> ]]

14:48:40 <jang-scri> guha prefers "for human consumption" versus "informal"

14:48:40 <danbri> q+

14:48:41 * Zakim sees Danbri on the speaker queue

14:48:50 <jang-scri> path: informal is fuzzy, i agree with guha there

14:48:56 <jang-scri> DanCon: wants to delegate to schema editor here

14:49:20 <jang-scri> bwm: propose approve modification of the resolution as discussed here

14:49:40 <jang-scri> and action danbri to produce the words?

14:50:04 <jang-scri> danbri: isdefinedby's understood meaning has drifted from what it was originally intended to be

14:50:14 <jang-scri> bwm: so we withdraw resolution made two weeks ago

14:50:26 <jang-scri> and to action danbri to update the proposal

14:50:50 <DanCon> Guha_, pig-in-a-poke isDefinedBy http://www.google.com/search?q=%22pig+in+a+poke%22

14:50:54 <jang-scri> -action danbri/ (with guha's help) to update the proposal to resolve this

14:50:55 <Jema> Jema notes action 2002-05-24#4

14:51:04 <jang-scri> item 14, webont and dark triples

14:51:14 <AaronSw> more appropriately, http://www.dictionary.com/cgi-bin/dict.pl?term=pig+in+a+poke&r=67

14:51:33 <jang-scri> DanCon: j borden wrote good summary recently

14:51:41 <jang-scri> about motivations and why they need/want this

14:51:58 <jang-scri> he's going to elaborate these examples, so we're pretty close

14:52:57 <jang-scri> Guha_: the construct in rdf already gives dark triples

14:53:13 <jang-scri> offline I've been convinced that since the syntax is so compicated, they're not going to accept it

14:53:23 <DanCon> # DTTF: summary (gasp!) Jonathan Borden (Wed, May 22 2002) http://lists.w3.org/Archives/Public/www-webont-wg/2002May/0168.html

14:53:26 <jang-scri> DanCon: I'm not conviced DTs are not needed

14:53:43 <jang-scri> path: I think they're required, but I think webont can do them themselves

14:54:01 <jang-scri> bwm: summarise - we're waiting on jb's document

14:54:08 <jang-scri> can we move on?

14:54:18 <jang-scri> item, 13, document status + schedules

14:54:33 <jang-scri> syntax: I'm scared about impact on this

14:54:41 <jang-scri> no decision on literals, containers, etc.

14:54:55 <jang-scri> DaveB: without literals, etc. a couple of weeks of work

14:55:06 <jang-scri> pat: converging as long as literals don't change

14:55:20 <jang-scri> if literals become untidy, I've got more work

14:56:01 <jang-scri> I can still redo with untidy literals,

14:56:05 <jang-scri> without that two weeks

14:56:13 <jang-scri> updated MT in two weeks?

14:56:17 <jang-scri> path: yes, if you want it

14:57:07 <jang-scri> path compares changes to "being pecked to death by ducks". this is worth committing to posterity

14:57:54 <jang-scri> bwm: I don't think we've seen something new that we haven't seen beofre (on tidyness)

14:58:04 <jang-scri> path: I'll try to do this quickly - give me a week

14:58:17 <jang-scri> sergei's next after pat.

14:58:29 <jang-scri> we're out of time.

14:58:44 <jang-scri> after hours discussion suggestions...

14:58:49 <jang-scri> someone darkly mutters "tidyness"2

14:59:17 <Zakim> -??P5

14:59:18 <Jema> command not recognised

15:00:07 <jang-scri> discussion over commetns about what the primer/schema documents ought to say

15:00:21 <jang-scri> AaronSw: primer gets smaller if you do this

15:00:43 <jang-scri> frankm: what's in the primer about schemas now that would be appropriately deleted if peolple didn't think theyknow how to write schemas?

15:00:50 <jang-scri> schema bit is 7/45 pp

15:00:58 <jang-scri> most of that is just showing pictures of schemas

15:01:27 <jang-scri> bwm: I've heard a suggestion for after-hours on primer/schema discussion

15:01:30 <jang-scri> meeting closes

15:01:31 <jang-scri> -close

15:01:32 <Jema> The meeting is closed

15:01:33 <Zakim> -??P25

15:01:35 <Zakim> -??P22

15:01:45 <jang-scri> can someone else scribe if desired?

15:01:52 <bwm> k

15:01:53 <Jema> command not recognised

15:01:53 <Jema> command not recognised

15:02:12 <bwm> I've gotta go

15:02:17 <bwm> thanks everyone

15:02:19 <jang-scri> gota go too

15:02:22 <jang-scri> thaknks folks

15:05:09 <AaronSw> danc: no ehaustive enumeration... put it in the schema doc

15:05:27 <AaronSw> frank: when it's there i'll take it out of the primer if i can refer to it

15:07:23 <Zakim> -??P16

15:08:58 <Zakim> - +1.650.965.aabb

15:16:17 <Zakim> -PatrickS

15:16:22 <Zakim> -MDean

15:16:24 <Zakim> -DanC

15:16:25 <Zakim> -??P20

15:16:27 <Zakim> -PatH?

15:16:29 <Zakim> -AaronSw

15:16:31 <Zakim> -Manola

15:16:33 <Zakim> -DanBri

15:16:38 <Zakim> SW_RDF Cor()10:00AM has ended

15:18:31 <DanCon> sigh... I meant to just make an editorial suggestion, and I end up in a heated discussion with the editor(s).

15:18:32 <DanCon> why is life so difficult?