IRC log of rdfcore on 2003-03-14

Timestamps are in UTC.

15:01:39 [RRSAgent]
RRSAgent has joined #rdfcore
15:01:48 [em]
em has changed the topic to: rdfcore 2003-03-14 teleconference
15:01:53 [em]
zakim, this is rdf
15:01:53 [Zakim]
ok, em
15:01:57 [em]
zakim, who is here?
15:01:57 [Zakim]
On the phone I see ??P2, PatH, Manola, EMiller, ??P10
15:01:58 [Zakim]
On IRC I see RRSAgent, Zakim, DaveB, em, gk, bwm, AaronSw, logger
15:02:08 [DaveB]
Zakim, ??P10 is probably ILRT
15:02:08 [Zakim]
+ILRT?; got it
15:02:11 [Zakim]
15:02:13 [DaveB]
Zakim, ILRT has daveb, jang
15:02:13 [Zakim]
+daveb, jang; got it
15:02:29 [em]
zakim, ??P2 is bwm
15:02:29 [Zakim]
+bwm; got it
15:03:33 [bwm]
zakim, who is on the phone?
15:03:33 [Zakim]
On the phone I see bwm, PatH, Manola, EMiller, ILRT?, GrahamKlyne
15:03:34 [Zakim]
ILRT has daveb, jang
15:03:35 [Zakim]
15:03:48 [mdean]
mdean has joined #rdfcore
15:04:36 [DaveB]
meeting starts
15:04:41 [Zakim]
15:04:44 [DaveB]
Zakim, who's on the phone
15:04:44 [Zakim]
I don't understand 'who's on the phone', DaveB
15:04:44 [danbri]
danbri has joined #rdfcore
15:04:56 [danbri]
zakim, EricP is temporarily DanBri
15:04:56 [Zakim]
+DanBri; got it
15:05:04 [DaveB]
Zakim, who is on the phone?
15:05:04 [Zakim]
On the phone I see bwm, PatH, Manola, EMiller, ILRT?, GrahamKlyne, Mike_Dean, DanBri
15:05:07 [Zakim]
ILRT has daveb, jang
15:05:12 [DaveB]
roll call above correct
15:05:13 [DaveB]
15:05:15 [jang_scribe]
jang_scribe has joined #rdfcore
15:05:16 [DaveB]
no aob
15:05:19 [jang_scribe]
15:05:32 [jang_scribe]
tuesday telecon, at risk
15:05:37 [jang_scribe]
bwm can't make it
15:05:47 [gk]
gk can't make it
15:05:49 [jang_scribe]
turnout for this week's wasn'ty that great; pat can't make it next week
15:06:31 [jang_scribe]
em: bwm is a mission-critical resource for this
15:06:47 [jang_scribe]
next tuesday's telecon is cancelled; next telecon this time, next week
15:06:51 [jang_scribe]
two hours ok for everyone?
15:07:02 [jang_scribe]
(gk can't make it, travelling)
15:07:15 [jang_scribe]
em: jjc making it next week?
15:07:41 [jang_scribe]
bwm: ok, 120 minutes next friday
15:07:45 [jang_scribe]
volunteer to scribe?
15:07:47 [jang_scribe]
daveb steps up
15:08:15 [jang_scribe]
moving on, minutes of last friday's telecon...
15:08:20 [jang_scribe]
[pointer anyone?]
15:08:24 [jang_scribe]
15:08:33 [danbri]
15:08:33 [danbri]
5: Minutes of 28 Feb 2003 telecon
15:08:33 [danbri]
15:08:33 [danbri]
15:08:34 [danbri]
15:08:40 [danbri]
15:08:40 [danbri]
6: Minutes of 11 Mar 2003 telecon
15:08:40 [danbri]
15:08:40 [danbri]
15:08:42 [danbri]
15:08:42 [jang_scribe]
minutes of tuesday's telecon?
15:09:02 [jang_scribe]
also approved
15:09:06 [jang_scribe]
completed actions:"
15:09:11 [jang_scribe]
long list, anyone unhappy?
15:09:22 [jang_scribe]
15:09:35 [jang_scribe]
item 8: daveb, xml schema 1.1 requirements
15:10:06 [jang_scribe]
daveb responded immediately, that's done;
15:10:34 [jang_scribe]
the second action continues; hopefully jjc can pick up daveb's comments and add his concerns
15:10:45 [jang_scribe]
uris for as many datatypes as they can
15:10:50 [jang_scribe]
raising the priority for this
15:10:54 [jang_scribe]
webont also keen for this
15:11:17 [jang_scribe]
DaveB: leave this until next week; hopefully we'll have comments by then
15:11:28 [jang_scribe]
if comments complete by middle of next week we should send it then.\
15:12:16 [jang_scribe]
bwm: I'll reraise this next week, if people are keen please look at this asap
15:12:39 [jang_scribe]
item 9: handling lc comments
15:13:02 [jang_scribe]
15:13:10 [jang_scribe]
danbri: I've about a dozen left to deal with
15:13:26 [jang_scribe]
expect reasonable progress by next week
15:14:45 [jang_scribe]
bwm: can anyone pitch in to help danbri with this?
15:15:20 [jang_scribe]
danbri:there are things where people talk about schema but there's a large semantics overlap
15:15:32 [jang_scribe]
path: talk to me via email about it, I can help there
15:16:40 [jang_scribe]
jang_scribe: give me a yell monday, we'll divvy the work up
15:16:57 [jang_scribe]
bwm: what about other docs, are we up to date on these?
15:17:03 [danbri]
zakim, mute danbri
15:17:03 [Zakim]
DanBri should now be muted
15:17:10 [jang_scribe]
that is, has every comment been handled as editorial discretion or in the issues list?
15:17:21 [jang_scribe]
frankm: there's one against the primer raised by eric p
15:17:33 [jang_scribe]
we talked at the plenary about it, not heard officially about this yes
15:17:39 [jang_scribe]
my intent is to handle it editorially
15:17:49 [jang_scribe]
bwm: other docs?
15:18:05 [jang_scribe]
gk: (looks at todo list...)
15:18:35 [jang_scribe]
gk: I've a couple in limbo, should be ok I think
15:18:43 [jang_scribe]
bwm: daveb: spotted chaals' messages?
15:18:56 [jang_scribe]
saying what happened to aboutEach, aboutEachPrefix?
15:19:20 [jang_scribe]
DaveB: just noticed reagle's "wrapping xml literal" too
15:19:27 [jang_scribe]
I'm up to date I think
15:19:51 [jang_scribe]
ACTION jang: take another trawl through -comments for dripped issues
15:20:09 [jang_scribe]
gk: comments from susan lesch on concepts...
15:20:16 [DaveB]
chaals's msg, for syntax, yet todo
15:20:19 [jang_scribe]
I was assuming those are editorial, I'm kind of uncertain
15:20:37 [jang_scribe]
bwm: if editorial, deal with them as editorial
15:20:41 [DaveB]
reagle on rdf-wrapper, not responded (concepts?)
15:20:51 [jang_scribe]
em: that's susan's job: she's the w3c's style guide.
15:21:45 [jang_scribe]
action bwm: series editor to review susan's comments for cross-document consistence
15:22:10 [jang_scribe]
gk: I've another... comments from eric P about canonicalisation, no record of it being tied in with anything else
15:22:28 [jang_scribe]
it's still sitting needing a process ID or something
15:22:55 [gk]
Eric P comment:
15:24:10 [jang_scribe]
ACTION bwm (pp gk, jjc) chase eric P's message, get a process resolution for it
15:24:24 [jang_scribe]
any other incoming lccs?
15:24:27 [jang_scribe]
gk: another one...
15:24:41 [jang_scribe]
aaron commented on the use of 404 not found http responses...
15:24:51 [jang_scribe]
... hang on, I think this fits in with the other social issue stuff
15:25:00 [jang_scribe]
path: yes, think we've dealt with that.
15:25:10 [jang_scribe]
bwm: please ensure aaron gets a response for this.
15:25:13 [danbri]
15:25:15 [jang_scribe]
AaronSw: still here?
15:25:28 [gk]
Aaron, is this response OK:
15:26:01 [jang_scribe]
action: gk, please chase aaron's message ensure he's ok with response
15:26:09 [jang_scribe]
bwm: any other incoming comments?
15:26:12 [DanC]
DanC has joined #rdfcore
15:26:15 [jang_scribe]
then moving on to item 10...
15:26:31 [jang_scribe]
pfps issues 17, 18, 19, 20, 21, 10043, 23819329
15:26:43 [jang_scribe]
acked danc\
15:27:17 [jang_scribe]
pfps was happy with frankm's response on this, said it'd work for the other docs too
15:27:43 [jang_scribe]
proposal from agenda outlined by bwm...
15:27:49 [jang_scribe]
objections? comments?
15:27:51 [jang_scribe]
15:28:03 [jang_scribe]
next, item 11. macgregor-01, macgregor-02
15:28:13 [jang_scribe]
comments on section 4 of concepts, which has gone, I think
15:28:22 [jang_scribe]
gk: possibly a residual issue with -01 here...
15:28:34 [jang_scribe]
it asks how statements may or may not be asserted
15:28:49 [jang_scribe]
I responded early on this and pfps came back and asked if that addressed all the issues.
15:29:16 [DaveB]
macgregor msg:
15:30:58 [jang_scribe]
bwm: did we refer anywhere to distinguishing between asserted and nonasserted forms?
15:31:19 [jang_scribe]
path: it'd be in the spirit of the social meaning discussion/resolution to purge these comments too
15:31:43 [jang_scribe]
gk: ok, I've raised the point against that.
15:31:51 [jang_scribe]
bwm: @asserted@ only appears in section 4 of concepts
15:32:52 [jang_scribe]
gk: do we close this or hang onto it unless it pops up again?
15:33:02 [jang_scribe]
bwm: I'm ok with jjc's proposal to close...
15:33:49 [jang_scribe]
15:34:10 [bwm]
15:34:16 [jang_scribe]
DaveB: grepped in wds for assertion, a few times in lbase, concepts, primer, often in semantics (obviously)
15:35:09 [jang_scribe]
gk: ok with jjc's proposal
15:35:20 [jang_scribe]
DaveB: not hesitating, just saying there's action required.
15:35:36 [jang_scribe]
proposal: to close this, with actions on editors to review for use of the term @asserted@ in light of these comments
15:35:39 [jang_scribe]
any comments...?
15:35:50 [jang_scribe]
[excuse @, american keyboard]
15:36:01 [jang_scribe]
ok, proposed... any dissent?
15:36:12 [jang_scribe]
this is wrt macgregor-01 and macgregor-02...
15:36:44 [jang_scribe]
frankm: strictly speaking macgregor-02 references concepts, he agrees with my proposal to deal with this in the primer
15:37:10 [jang_scribe]
back t the proposal to close:
15:37:24 [jang_scribe]
no dissent, no abstentions?
15:37:29 [jang_scribe]
resolved, actions arising:
15:37:56 [jang_scribe]
action gk: respond wrt macgregor-01, macgregor-02
15:38:10 [jang_scribe]
action: all editors, check for use of term @asserted@ and modify in the light of this comment
15:38:37 [jang_scribe]
frankm: gk should include in his response the information that the other documents are being checked too
15:38:40 [jang_scribe]
moving on...
15:38:52 [jang_scribe]
item 12, reagle-01, reagle-02, skipping in the absence of jjc...?
15:39:02 [jang_scribe]
item 13, williams-01
15:39:09 [jang_scribe]
gk's proposal:
15:39:19 [jang_scribe]
15:41:08 [jang_scribe]
gk: I've put a sketch of words in proposal given,
15:41:23 [jang_scribe]
main proposal is that this is addressed by the terminological approach suggested by pat
15:42:12 [jang_scribe]
DaveB: this issue is @what is a node in an rdf graph?@
15:42:59 [jang_scribe]
bwm: recently decided that an rdf graph IS a set of triples, etc.
15:43:09 [jang_scribe]
some of the text still refers to nodes in a graph
15:43:17 [jang_scribe]
this is asking about that use of terminology
15:43:31 [jang_scribe]
gk's response here is sensible, referring to danc-01
15:43:53 [jang_scribe]
this is an editorial cleanup which we'll use consistently throughout.
15:43:58 [jang_scribe]
proposal, then:
15:44:09 [gk]
15:44:10 [gk]
I propose that this issue is addressed by following the terminological
15:44:10 [gk]
approach suggested by Pat in another message:
15:44:10 [gk]
15:44:11 [gk]
15:44:41 [jang_scribe]
davceb: isn't that already in concepts?
15:44:56 [jang_scribe]
is the response not, @yes this is the case, see these sections of concepts@
15:45:34 [jang_scribe]
DaveB: I'd like to see, @I'm going to fix these sections of concepts@ if what's there already isn't exactly what we're going to finish with
15:45:58 [jang_scribe]
DaveB: this should be in concepts; if concepts is currently wrong then I'd like to see the fixed words in the proposal
15:46:24 [jang_scribe]
gk: proposal is that we make editorial changes in line with pat's suggestion
15:46:48 [jang_scribe]
DaveB: ONE document should give the definition, everything else points to that
15:47:07 [jang_scribe]
if you need to explain more in an email then the email contents belong in concepts
15:48:36 [jang_scribe]
DaveB: so your answer is, @yes, an rdf graph is a set of triples, see concepts section X@
15:49:08 [jang_scribe]
concepts doesn't mention nodes in a graph; (gk: section 6 does)
15:49:16 [jang_scribe]
or parts of a triple
15:49:51 [jang_scribe]
frankm: primer starts off with graphs that are pictures.
15:50:11 [jang_scribe]
path: that's fine. we can say, this is a picture of an rdf graph, and the pictures have nodes in them
15:50:56 [jang_scribe]
bwm: 3.2 in concepts looks a bit iffy, that's the one stuart was complaining about
15:51:53 [jang_scribe]
bwm: the subject of a triple is a uri (or a bnode): that's the terminology we decided to use
15:52:04 [jang_scribe]
15:52:54 [jang_scribe]
path: somewhere prominently we have a section that gives definitions, we can say, we use the term @node@ to apply to subjects and objects, this is a terminological convenience
15:53:48 [jang_scribe]
DaveB: it's handy to have a general term instead of saying, @uriref or literal or typed literal or...@
15:54:12 [jang_scribe]
path: if we have a shorthand terminology it belongs in concepts with the rest of these definitions
15:54:29 [jang_scribe]
DaveB: reads 6.2 @the nodes in an rdf graph consists of...@
15:54:54 [jang_scribe]
path: could add a note to the effect that the node IS the uriref
15:55:43 [jang_scribe]
stuarts wording suggested @labelled with@ which has been excised
15:56:17 [jang_scribe]
bwm: don't think this response is cooked
15:56:31 [jang_scribe]
gk: agree, need to go round again and come back with a firmer proposal
15:56:41 [jang_scribe]
agreed, moving on.
15:56:50 [jang_scribe]
item 14: pfps-03
15:57:08 [jang_scribe]
path: oh, that's trivial, I've taken it on as editorial
15:57:28 [jang_scribe]
the lbase translations weren't accurate, he's right, but I'm putting this off until the semantics settles down
15:57:52 [jang_scribe]
my comment to pfps was @this will be fixed once the rest ofthe document has been signed off@
15:58:15 [jang_scribe]
bwm: I think we need to leave it open until it's fixed, then
15:58:19 [jang_scribe]
path: fine, ok.
15:58:30 [jang_scribe]
item 15:L pfps-04, -05, -06, -07, -10
15:58:46 [jang_scribe]
path: all of those are genuine bugs pfps found which have been corrected in the editor's draft
15:59:12 [jang_scribe]
proposal: accept these comments, close the issues, refer pfps to the updated editor's draft.
15:59:32 [jang_scribe]
bwm: we need someone to review these changes?
16:00:22 [jang_scribe]
path: the issue is if these changes have any unexpected side-effects
16:00:37 [jang_scribe]
bwm: gk, can you review these?
16:00:55 [jang_scribe]
gk: yes, providing the action's clear what it is I have to review and it's not required by next week
16:01:26 [DaveB]
pat's draft
16:01:33 [jang_scribe]
jang_scribe: I'll do it too...
16:02:21 [jang_scribe]
path: there's a chunk of missing machinery dealing with lang tags on dted literals
16:02:30 [jang_scribe]
action: gk - review path's changes to semantics
16:02:38 [jang_scribe]
action jang: review path's changes to semantics
16:03:02 [jang_scribe]
(changes... arising out of pfps 05 06 07 10)
16:03:07 [jang_scribe]
(and 04)
16:03:14 [jang_scribe]
^^^ amend action text
16:03:27 [jang_scribe]
moving on,
16:03:34 [jang_scribe]
item 16: pfps-08
16:03:59 [jang_scribe]
path; that belongs with the others, another tweak
16:04:08 [jang_scribe]
^^^ action add pfps-08
16:04:48 [jang_scribe]
path's dealing with pfps-08 was about xmlliteral
16:05:22 [jang_scribe]
xmlliteral is a bit odd, an owl reasoner could decide eg:x === xmlliteral
16:05:37 [jang_scribe]
then should a dt typed with eg:x behave like it was typed with xmlliteral?
16:05:41 [jang_scribe]
the answer is, no.
16:05:47 [jang_scribe]
pfps thinks that's a mistake, I disagree.
16:06:11 [jang_scribe]
that's the outcome of that issue; the mt has no alterations to make it work like pfps thinks it should in this respect
16:06:18 [jang_scribe]
the text has been changed to clarify this
16:06:50 [jang_scribe]
outcome: owl has to treat xmlliteral as a special case.
16:07:22 [jang_scribe]
bwm: so, then what is the closing resolution?
16:07:47 [jang_scribe]
path: the behaviour described is not an error. the document has been clarified in this respect
16:08:13 [jang_scribe]
so the comment is NOT accepted in spirit
16:08:43 [bwm]
Propose: RDFCore do not accept this comment. The semantics are as intended. The text has been clarified to make this clearer.
16:08:44 [jang_scribe]
[bwm types proposal]
16:09:19 [jang_scribe]
path proposes, daveb seconds, dissent? no abstentions? no
16:09:21 [jang_scribe]
16:09:26 [jang_scribe]
(just like the UN!)
16:10:00 [jang_scribe]
action: bwm - update issue list to opint to path's response to peter onthis
16:10:04 [jang_scribe]
moving on
16:10:11 [jang_scribe]
item 17, qu-01, path...?
16:10:33 [jang_scribe]
should the domain of rdfs:member be rdfs:Container?
16:10:54 [jang_scribe]
the rdfcore decided otherwise.
16:11:28 [jang_scribe]
danc commented, incidentally, that there should be an rdf:member between collections and their member
16:12:03 [jang_scribe]
path: he thought we'd forgotten to say it, we respond, thanks, but nope.
16:12:59 [jang_scribe]
frankm: worthwhile pointing out explicitly somewhere that one of these properties used in a triple somewhere does NOT imply that the subject is a container?
16:13:42 [jang_scribe]
proposal: the use of a container membership property or rdfs:member in a triple should not be taken as a claim that the subject of the triple is a container.
16:13:58 [jang_scribe]
(so this comment is rejected)
16:14:10 [jang_scribe]
proposer: frankm; seconder: gk.
16:14:16 [jang_scribe]
dissent? nope. abstains? nope
16:14:21 [jang_scribe]
so resolved.
16:14:32 [jang_scribe]
path: is there a recorded decision from earlier?
16:14:41 [jang_scribe]
bwm: the term @decided@ in the agenda might have been too strong
16:14:54 [jang_scribe]
[is there an xmodmap for osx?]
16:15:07 [jang_scribe]
actions: path to respond; nothing else need doing
16:15:10 [jang_scribe]
moving on.
16:15:17 [jang_scribe]
item 18, qu-02
16:15:27 [jang_scribe]
rdfs:member a functional property?
16:15:45 [jang_scribe]
rdf doesn't have a functional property. This is in owl's domain.
16:15:49 [jang_scribe]
^^^ path
16:15:54 [danbri]
zakim, unmute danbri
16:15:55 [Zakim]
DanBri should no longer be muted
16:16:16 [jang_scribe]
danbri: there are lots of things we can't effect formally we can express in prose
16:16:59 [jang_scribe]
any other opinions on this, should rdfs:member BE functional?
16:17:14 [jang_scribe]
not member, rdf:_1, etc
16:18:34 [jang_scribe]
frankm: this situation is akin to what we've said about collection
16:18:54 [jang_scribe]
if you use parsetype collection, you get this neat form; if you write it by hand you can make other odd datastructures
16:19:07 [jang_scribe]
frankm: how are you going to check it?
16:19:21 [jang_scribe]
gk: I broadly agree with path; there's a way this is visible with rdf entailment
16:19:32 [bwm]
[[To represent a collection c, create a triple {RDF:type, c, t} where t is one of the three collection types RDF:Seq, RDF:Bag, or RDF:Alt. The remaining triples {RDF:_1, c, r1}, ..., {RDF:_n, c, rn}, ... point to each of the members rn of the collection. For a single collection resource there may be at most one triple whose predicate is any given element of Ord and the elements of Ord must be used in sequence starting with RDF:_1. For resources
16:19:33 [bwm]
that are instances of the RDF:Alt collection type, there must be exactly one triple whose predicate is RDF:_1 and that is the default value for the Alternatives resource (that is, there must always be at least one alternative).]]
16:19:47 [jang_scribe]
this is the thing daml tries to express, so we should try to avoid addressing it.
16:19:57 [jang_scribe]
path: so again, we say, thanks bu tno thanks.
16:20:53 [jang_scribe]
path: we've abandoned all such syntactic constraints on graphs, right?
16:21:15 [danbri]
rdf:Alt is useless clutter imho
16:21:37 [jang_scribe]
path: this currently isn't broken
16:22:18 [jang_scribe]
so path proposaes, gk seconds... anyone against?
16:22:33 [jang_scribe]
propose then, not to accept qu-02.
16:22:34 [jang_scribe]
16:22:57 [jang_scribe]
- it's not necessary; - it might damage behaviour of existing code (mozilla)
16:23:31 [jang_scribe]
proposer: path; seconded: gk.
16:23:39 [jang_scribe]
against? none. abstain? none
16:23:41 [jang_scribe]
16:23:46 [jang_scribe]
actions: path to respond to qu
16:23:53 [jang_scribe]
...wrt qu-02
16:23:59 [jang_scribe]
moving on
16:24:05 [jang_scribe]
item 19, xmlsch-08
16:24:14 [jang_scribe]
DaveB: has a proposal
16:24:38 [jang_scribe]
16:24:47 [jang_scribe]
DaveB: why not use xsi:type?
16:25:24 [jang_scribe]
DaveB: we chose not to for these reasons...
16:25:28 [jang_scribe]
they express concern
16:25:42 [jang_scribe]
we chose not to use xsi:type because THAT would be confusing, it's an overloading of the semantics
16:26:04 [jang_scribe]
eg, there's a difference between uri(refs) and qnames
16:26:52 [jang_scribe]
we accept that there are good reasons; here are a couple more
16:27:15 [jang_scribe]
bwm: ... and if they've any other suggestions we'd be glad to hear them
16:27:29 [em]
16:28:16 [jang_scribe]
em: these guys are really trying hard to work with us, we need to be as appreciative as possible
16:28:27 [jang_scribe]
general consent from path, admiration of their efforts
16:28:49 [jang_scribe]
gk: is it _legal_ to have a literal with an xsi:type in an rdf graph?
16:29:10 [jang_scribe]
DaveB: and xml literal? if it's properly declared, there's no restriction if it's properly declared
16:30:41 [jang_scribe]
questions on the syntactic legality of xsi:types at various points in rdf/xml syntax
16:31:01 [jang_scribe]
DaveB: you couldn't accidentally use it, I'll consider adding that to the response
16:31:18 [jang_scribe]
but listing bad examples in our dopuments, I think, isn't the route to go down
16:31:46 [jang_scribe]
DaveB: so, show why we don't use xsi:type
16:31:53 [jang_scribe]
give an example
16:32:21 [jang_scribe]
we chose not to use xsi:type because THAT would be confusing, it's an overloading of the semantics
16:32:21 [jang_scribe]
<jang_scribe>eg, there's a difference between uri(refs) and qnames
16:32:29 [jang_scribe]
^^^ daveb's suggested resolution
16:32:44 [jang_scribe]
gk: I'm not unhappy; I'm still trying to think if xsoi;type might appear in rdf/xml
16:32:54 [jang_scribe]
DaveB: you can't use it that way in rdf/xml
16:34:59 [jang_scribe]
bwm: do we need towait for gk to think on this more? he's got a niggle about pathological examples
16:35:28 [jang_scribe]
gk: don't think it affects the outcome, i'll try to get an example gk's after
16:35:37 [jang_scribe]
bwm: I think we can decide this now.
16:36:09 [jang_scribe]
proposal: we accept the comment, but in responding, daveb points out the reasons why xsi:type can't be used (their reasons, plus at least one more)
16:36:33 [jang_scribe]
also point at the explanation, give an example of how if someone uses xsi:type by mistake, an rdf/xml parser would pick up the error
16:36:40 [jang_scribe]
DaveB: propose; gk: seconds
16:36:45 [jang_scribe]
against? 0-. abstains? 0
16:36:48 [jang_scribe]
done, resolved.
16:37:23 [jang_scribe]
DaveB: suggeswt jumping to horrocks-01
16:37:29 [jang_scribe]
item 24, horrocks-01
16:37:53 [jang_scribe]
ian wants to add genuine comments to an ontology
16:38:06 [jang_scribe]
that is, you add it it doesn't affect the logical entailment in any way whatsoever
16:38:16 [jang_scribe]
so changing the comment in rdf changes the entailments
16:38:41 [jang_scribe]
the compromise position we arrived at was,
16:38:57 [jang_scribe]
if we @trivialise@ the comments by being true in every entailment
16:39:04 [jang_scribe]
then inference engines can ignore them
16:39:13 [jang_scribe]
so we add that as a semantic condition on rdfs:comment
16:39:29 [jang_scribe]
bwm: example, maybe, of a harmful entailment?
16:39:44 [jang_scribe]
path: changging a comment changes the formal entailment.
16:40:21 [jang_scribe]
jang_scribe: eg, spellchecking comments will break entailed conclusions
16:40:40 [gk]
An example from Pat's message:
16:40:42 [jang_scribe]
path: one option suggested was, why not just get your checker to ignore comments?
16:40:43 [gk]
16:40:46 [gk]
eg by
16:40:46 [gk]
virtue of there only being three comments in the graph, a cardinality
16:40:46 [gk]
constraint applying to a superproperty of rdf:comment and an
16:40:46 [gk]
assertion that rdf:comment was functional could produce an
16:40:46 [gk]
16:40:48 [gk]
16:40:53 [jang_scribe]
the response was, if we ignore it, it should be ignored in the semantics
16:41:56 [jang_scribe]
gk: if I read the comment you made in your message to the group (quoted above)
16:41:59 [danbri]
zakim, q+ to recall RDFS WG design
16:41:59 [Zakim]
I see danbri on the speaker queue
16:42:33 [jang_scribe]
path: it's true that owl can give propositions more complex and wide-ranging effects than it has in rdf
16:43:09 [jang_scribe]
in principle, you could set up an owl definition that becomes inconsistent by virtue of not havoing enough comments in it!
16:43:19 [em]
16:43:21 [em]
16:43:30 [gk]
q+ to ask if the problem is with rdfs:commetn per se, or that lack of a property immune to these effects
16:43:55 [em]
16:44:03 [em]
16:44:28 [jang_scribe]
bwm: this might hurt in owl, @well, don't do that then@
16:44:48 [jang_scribe]
there amy be subproperties of comment, for instance
16:44:51 [jang_scribe]
^^^ path
16:45:00 [jang_scribe]
ian's worry is that such tricks might bite people in unexpected ways
16:45:16 [jang_scribe]
16:45:21 [em]
ack danbri
16:45:22 [Zakim]
danbri, you wanted to recall RDFS WG design
16:45:24 [jang_scribe]
zakim, ack danbri
16:45:24 [Zakim]
I see gk on the speaker queue
16:45:35 [bwm]
ack danbri
16:46:05 [jang_scribe]
the rdf schema design discussions ...
16:46:29 [jang_scribe]
... we anticipated things that came along with @real@ semantics like domain range, and @empty@ things, ie, more like comment
16:47:10 [jang_scribe]
one of the comments I've yet to respond to was a request for better commenting machinery
16:47:25 [jang_scribe]
path: the natural response might be,
16:47:42 [jang_scribe]
we know this should be replaced in the future, and this is what we've currently got
16:48:14 [jang_scribe]
owl has a class of annotation properties
16:48:24 [jang_scribe]
danbri: my next proposal was that owl did that, not us
16:48:29 [jang_scribe]
so, great.
16:48:30 [gk]
16:49:14 [jang_scribe]
gk: is Ian's problem with the existence of rdfs:comment or the lack of a @real@ commenting mechanism?
16:49:23 [jang_scribe]
path: both. he needs a commenting mechanism, but this is too loose.
16:49:40 [danbri]
em, sure
16:49:43 [em]
16:49:43 [jang_scribe]
owl could do that for itself if it wanted to
16:49:48 [em]
16:50:04 [danbri]
q+ to ask whether their problem extends to any unknown properties appearing in a vocab/schema
16:50:09 [jang_scribe]
gk: owl could define for itself an owl:comment with the constraint that A owl:comment b is always true?
16:50:19 [bwm]
ack gk
16:50:19 [Zakim]
gk, you wanted to ask if the problem is with rdfs:commetn per se, or that lack of a property immune to these effects
16:50:22 [jang_scribe]
path: yes, or it could apply additional constraints of the same ilk on rdfs:comment
16:50:25 [bwm]
ack em
16:50:46 [jang_scribe]
em: there's already a lot of confusion about when to use what. owl:Class, rdfs:Class\
16:50:59 [jang_scribe]
there's a communication aspect that needs fl[ue]shing out here
16:51:16 [jang_scribe]
I like danbri's suggestion about owl addition additional information to these particular properties\
16:51:26 [jang_scribe]
but the reduplication of these things as a slight means of constraining it
16:51:39 [jang_scribe]
when trying to go out and explain this to folks seems as a problme
16:51:42 [bwm]
16:51:55 [jang_scribe]
path: I don't see that we'll always be available to avoid two versions of some things, it's inevitable
16:51:59 [bwm]
ack danbri
16:51:59 [Zakim]
danbri, you wanted to ask whether their problem extends to any unknown properties appearing in a vocab/schema
16:52:08 [Zakim]
16:52:10 [jang_scribe]
em: in that case we should be super-cautious when playing that card
16:53:01 [jang_scribe]
path: what bothers ian is that something CALLEd comment with the INTENDED PURPOSE of attaching comments should have real semantic import\
16:53:17 [jang_scribe]
danbri: so is this just a philosophical clash?
16:53:39 [jang_scribe]
path: ian's made a genuine engineering comment, that teams of people collaborating over time
16:53:49 [jang_scribe]
on an ontology, will NEED (absolutely) comments
16:53:57 [jang_scribe]
they're critical for the management of a large effort like that
16:54:14 [jang_scribe]
just like you need them in software engineering
16:54:29 [jang_scribe]
without those comments getting in the way of formal machinery you're using to check the real axioms
16:54:52 [jang_scribe]
it's be a mistake to dismiss this use case, since we expect that people WILL use rdf and DAML and OWL to create large-scale ontologies
16:55:11 [jang_scribe]
bwm: firstly, I'm really not sure what the problem is...
16:55:22 [jang_scribe]
secondly, when I think about writing comments in the small ontologies I deal with,
16:55:35 [jang_scribe]
I think of them not necessarily as comments to the user of a schema, but as showing up in a UI
16:55:41 [jang_scribe]
in ontology tools
16:55:53 [jang_scribe]
path: that's jimh's point of view, that's why he wants them to be retained
16:56:04 [jang_scribe]
to be fair to ian, we don't provide what ian needs
16:56:17 [jang_scribe]
16:56:29 [jang_scribe]
pat, did you see my response on wednesday?
16:56:33 [jang_scribe]
it seems to me that
16:56:40 [jang_scribe]
9a) this is much like the class/instance distinction
16:56:40 [DanC]
I think we should seriously consider an rdfs:noop or rdfs:triviallyTrue property, ala Jeremy's c14n/signature stuff.
16:57:07 [jang_scribe]
why are we precluding people from doing tricks/useful things from comments at the rdf level
16:57:16 [jang_scribe]
as opposed to people building it in in a specialised language
16:57:17 [jang_scribe]
16:57:19 [danbri]
From RDFS WG archives (with Guha's permission):
16:57:20 [danbri]
16:57:21 [DanC]
but to trivialize rdfs:comment is no fair, at this stage of deployment.
16:57:21 [danbri]
Message-ID: <>
16:57:21 [danbri]
Date: Sat, 23 May 1998 14:10:03 -0700
16:57:21 [danbri]
From: (Ramanathan Guha)
16:57:21 [danbri]
16:57:22 [danbri]
16:57:24 [danbri]
Subject: Re: Open Issue #18
16:57:26 [danbri]
Here is one factor to consider before we decide
16:57:28 [danbri]
to postpone this --- pretty much everyone who
16:57:30 [danbri]
uses RDF Schemas will come up with some property
16:57:32 [danbri]
type to hold a human readable description. This
16:57:34 [danbri]
is just going to happen. Wouldn't it be better
16:57:36 [danbri]
if everyone used the same property type (even
16:57:38 [danbri]
if it were not 100% perfect?)
16:57:40 [danbri]
16:57:43 [danbri]
16:57:44 [danbri]
(sorry to interrupt Jang)
16:57:46 [danbri]
16:57:46 [jang_scribe]
(b) we need facilities for people building large ontologies; is rdfs:comment used for that?
16:57:57 [jang_scribe]
miked: rdfs:comment is used usually as a description
16:58:14 [jang_scribe]
frankm: right, a software engineering discipling would need structured kinds of comments
16:58:22 [jang_scribe]
dabnbri: guha's comment...
16:58:34 [DanC]
yes, 'comment' was poorly chosen. but the community groks it (and uses it) as designed, i.e. as description. No fair to change it now.
16:59:05 [jang_scribe]
we built something useful to a lot of people, we need to stand by that
16:59:13 [danbri]
DanC, we're out of time. Please send by email.
16:59:17 [jang_scribe]
path: this isn't getting a very sympathetic hearing.
16:59:56 [jang_scribe]
path: the only option is to say @sorry@ or to do the formal trick I mentioned in the email
17:00:00 [DanC]
roger, danbri.
17:00:06 [jang_scribe]
which means that an empty graph entails every single comment
17:00:43 [jang_scribe]
gk: in proposing to reject it we must'nt be seen as being totally dismissinve of the concerns expressed here
17:00:53 [jang_scribe]
path: yes, I'd come back to the gropu with words for this
17:01:07 [jang_scribe]
bwm: danbri, do you have wording on this?
17:01:25 [DanC]
danbri, are you really out of time, it seems the discussion continues.
17:01:37 [jang_scribe]
postponed until next week
17:01:42 [jang_scribe]
bwm: we'll wordsmith offline
17:01:49 [DanC]
ah. ok.
17:02:07 [jang_scribe]
action: path - produce the words of the decision for this
17:02:15 [jang_scribe]
17:02:17 [jang_scribe]
17:02:19 [danbri]
rrsagent, pointer?
17:02:19 [RRSAgent]
17:02:35 [Zakim]
17:02:37 [Zakim]
17:02:39 [Zakim]
17:02:40 [Zakim]
17:02:44 [Zakim]
17:02:57 [Zakim]
17:03:01 [Zakim]
17:03:03 [Zakim]
SW_RDFCore()10:00AM has ended
17:03:32 [danbri] etc should be public soon. I changed it, waiting for ACLs to mirror.
17:05:37 [danbri] public now.
17:05:39 [danbri]
danbri has left #rdfcore