Semantic Web Interest Group IRC Chat Logs for 2004-08-04

01:18:26 <crschmidt> Do people ever write applications to use PHP/MySQL/Jena?

01:35:42 * mattmcc wasn't aware there were Jena PHP bindings.

01:36:02 <mattmcc> I typically use python & redland.

01:36:49 * crschmidt nods.

01:37:04 <crschmidt> That was what I was asking, I think. I'm not really sure, I'm just passing the question on from a friend, actually :)

01:46:57 <Mutiny> Mutiny is now known as MutinyEvo

01:48:40 <MutinyEvo> MutinyEvo is now known as Mutiny

01:48:49 <chaalsMEL> morning all

01:48:57 <chaalsMEL> any cwm users here?

01:52:47 * chaalsMEL goes back to trial and error. (Although the errors are making it a serious trial :-)

02:49:16 <mattmcc> Hmm. The IBM Semantics Toolkit.

03:24:14 <sandro_-> sandro_- is now known as sandro_

05:20:20 <ear1grey> mattmcc: did you try downloading it?

06:20:14 <[GNU]> goog morning all

06:20:27 <chaalsMEL> morning

06:20:36 <[GNU]> anyone did some work on representing cantracts using rdf ?

06:20:43 <[GNU]> contracts...

06:20:52 <chaalsMEL> Can you confirm or deny that cwm variables are scoped to the rule they appear in?

06:21:01 <chaalsMEL> contracts: Yep. I forget who though :-(

06:21:12 <chaalsMEL> There are a number of people interested in it...

06:21:25 <[GNU]> ... if its not blogged its lost ... :)

06:22:00 * chaalsMEL hasn't really got into the blogging thing yet. Still working on web pages and nice tools for writing them.

06:22:24 <[GNU]> i guess blogging is just the new style wording for the same thing

06:24:42 <[GNU]> there is a daml integration ontology, but what it defines as a contract is nothing more than a nice mega-abstract-contract

06:24:58 <[GNU]> nothing describing roles and conditions and stuff...

06:26:49 <[GNU]> and... defining a owl class Contract as a subclass of wordnet/Contract is not smart i guess, isntit?

07:02:16 <chaalsMEL> Well, I suspect not.

07:02:30 <chaalsMEL> What you really want to do is describe conditions

07:03:03 <chaalsMEL> look at things like web service descriptions. The agreement to produce some result if you get some input is the basis of a contract.

07:03:48 <chaalsMEL> (although you then want to describe the inputs and outputs - maybe Cyc or some everything-and-the-kitchen-sink ontology might have enough things in it to start describing some example contracts...)

07:13:43 <mmealling>http://masten-space.com/index2.php

07:13:45 <dc_rdfig> A: http://masten-space.com/index2.php from mmealling

07:14:28 <mmealling> crud, I never can remember the chump syntax

07:14:45 <mmealling> A| Masten Space Systems, Inc.

07:14:53 <mmealling> well, that wasn't it...

07:14:59 <chaalsMEL> nope, you want A:| for title and A: for comments

07:15:03 <mmealling> A:| Masten Space Systems, Inc.

07:15:04 <dc_rdfig> Titled item A.

07:15:10 <chaalsMEL> dc_rdfig help?

07:15:18 <chaalsMEL> dc_rdfig, help

07:15:18 <dc_rdfig> Post a URL by saying it on a line on its own

07:15:19 <dc_rdfig> To post an item without a URL, say BLURB:This is the title

07:15:20 <dc_rdfig> I will reply with a label, for example A

07:15:21 <dc_rdfig> You can then append comments by saying A:This is a comment

07:15:22 <dc_rdfig> To title a link, use a pipe as the first character of the comment

07:15:23 <dc_rdfig> Eg. A:|This is the title

07:15:24 <dc_rdfig> To see the last 5 links posted, say dc_rdfig:view

07:15:25 <dc_rdfig> For more features, say dc_rdfig:morehelp

07:15:27 <mmealling> A: Something that Michael Mealling is doing in his spare time that has nothing to do with RDF or the Internet at all

07:15:27 <dc_rdfig> Added comment A1.

07:17:49 <chaalsMEL> A:Looks like more fun than [trying to merge EARL versions|http://www.w3.org/2001/sw/Europe/200407/earl] ...

07:17:50 <dc_rdfig> Added comment A2.

11:29:19 <Emmy> Good morning all :)

13:41:20 <libby_> libby_ is now known as libby

14:03:20 <swh_> swh_ is now known as swh

15:29:17 <libby>http://foafcamp.asemantics.org/tshirt/

15:29:19 <dc_rdfig> B: http://foafcamp.asemantics.org/tshirt/ from libby

15:29:33 <libby> B:|trying out some designs for foafcamp tshirts

15:29:34 <dc_rdfig> Titled item B.

15:30:00 <libby> B:also: what colour should the shirt be? I think orange.

15:30:00 <dc_rdfig> Added comment B1.

15:30:19 <libby> B:what do you think?

15:30:20 <dc_rdfig> Added comment B2.

15:31:32 <DanC> ooh! "T-Mobile USA Flips the Lid Open on the T-Mobile Sidekick II"

15:32:04 <DanC> . http://www.danger.com/press_release.php?item=040804

15:33:20 <DanC> "Along with the launch of the T-Mobile Sidekick II, T-Mobile and Danger plan to introduce software that will allow Sidekick customers to wirelessly synchronize their desktop contacts and calendar information with their T-Mobile Sidekick."

15:47:34 * libby remembers she was writing a blog entry about calendaring

15:47:43 * libby easily distractable

15:50:34 <Talliesin> heh, I can only remember that I was doing something. The details still escape me.

15:52:24 <libby> heh

15:52:39 <libby> B:[http://www.campusprintservice.nl/textiel.html possible tshirt colours]

15:52:40 <dc_rdfig> Added comment B3.

18:15:46 <libby> B3:[http://www.campusprintservice.nl/textiel.html|possible tshirt colours]

18:16:22 <dc_rdfig> Replaced comment B3.

18:16:29 <libby> ah, just slow

19:25:49 <DanC> anybody gottaminute re calendaring task force? danbri? libby?

19:29:19 * libby waves

19:36:02 <DanC> hi

19:36:15 <libby> heya

19:37:03 <DanC> in a recent meeting, somebody was asking "we have a task force on calendaring; what are they up to?" and I feel some obligation to answer... though the truth is sorta "well, they were going along pretty well until I broke the timezone stuff"

19:38:06 <libby> heh

19:38:42 <libby> I think it was an improvement, the way you 'broke' it, but is was a pretty big change I guess

19:38:57 <libby> I don;t know how to resolve it though

19:39:08 <DanC> well, what's objectively broken is that the schema and the test data don't agree.

19:39:13 <libby> if we could do some sort of versioning maybe...

19:39:17 <libby> yep true

19:39:48 <DanC> I'm still pretty sure maintaining multiple versions isn't worthwhile

19:40:45 <libby> I agree, and I don;t know how to do that type of versioning anyway

19:41:03 <DanC> rather than multiple versions of an iCalendar schema, I'd rather do one direct-mapping iCalendar schema, and one sorta general purpose talking-about-when-things-happen ontology.

19:41:53 <libby> yeah, icalendar is pretty specific

19:43:11 <libby> so what can we do to (1) bring the test data into line with teh schema and (2) get people like Msaahide back on board?

19:43:18 <DanC> I'm conflicted: the way to get a really valuable schema is to interact with lots of real-world data. But this is sort of a skunkworks project for me; I'm hesitant to recruit folks with LOTS of data cuz I can't really be sure I'll be around to deal with their issues.

19:43:26 * danbri catches up

19:43:44 <danbri> yes, a direct ical plus a more RDF-friendly 'when' vocab seems useful

19:44:11 <libby> I would also like a simple subset that works with most PIMs

19:44:19 <libby> main complaint is that icalendar is too complex

19:44:50 <DanC> some of the most complex parts are quite valuable in practice, to me; e.g. timezones.

19:44:53 <danbri> is it complex cos the problem is complex?

19:45:28 <DanC> though i don't really need to be able to make up new timezones just for my calendar; I can live with the selections from the tz database nicely.

19:46:24 <libby> +1

19:46:53 <libby> no the complexity that people encounter is hjust: which propwerties do I actually need to use to get this to work in ical or evolution wehatever

19:47:02 <libby> sorry about spelling!

19:47:40 <libby> that's an icalendar problem really, but we could perhaps help

19:47:47 <danbri> how stable are the timezones? does the tz database change in scary ways? additions? deletions?

19:47:51 <libby> plus I'd make rdfical simpler

19:48:02 <libby> s/I'd/It'd/

19:48:12 <libby> I don;t think so. not often

19:48:21 <DanC> how stable... very relevant to my other TODO item this week: a persistence ontology.

19:48:58 <DanC> I get the impression the tz database gets deployed into open source operating systems at a rate of about once every 18 months.

19:49:16 <libby> that's something we could find out

19:49:28 <deltab> most changes are incremental additions for DST start and end, I think

19:49:32 <DanC> and I think the individual timezones change at a rate inversely proportional to the number of people living in them.

19:50:09 <DanC> s/rate/probability/

19:50:52 <DanC> noodling out loud about goals/direction for a bit...

19:52:18 <DanC> for me, the role of the calendar ontology is (a) to manage my airline and telcon reservations/plans/invitations, but meanwhile, (b) to collaborate with peers on best-practice modelling issues, like InterpretationProperties.

19:53:04 <DanC> an idea/plan: get RdfCalendarDocumentation "finished" to the point where it's worth a read by the BPDWG; present it to them, and see what happens.

19:53:48 <DanC> "finished" includes: a schema derived from the RFC. (issue: DL-happy, or diff-happy?)

19:54:03 <DanC> .. includes: toIcal.py and fromIcal.py

19:54:14 <libby> I'd go with BPWG....I'd like people to be able to use it. there's a lot of potential users out there, with the right docs

19:54:23 <DanC> ... includes: test data that demonstrate, to a large extent, that toIcal.py, fromIcal.py, and the schema agree.

19:54:35 <golbeck_> golbeck_ is now known as golbeck

19:54:55 <libby> +1

19:55:32 <DanC> as I mentioned in email, libby, I'd like RdfCalendarDocumentation to go not property-by-property, but rather: sketch the simple mapping, and then dive down into why life isn't that simple, mostly follwing the structure of the fromIcal.py code.

19:55:51 <libby> where do you stand on simplifying the schema. or perhaps providing advice on which props are most used in PIMs?

19:56:06 <libby> yep makes sense, I dont think it's that good as is.

19:56:25 <libby> ah k, you sorta answered that one

19:56:46 <DanC> there are some _very_ good things about the current version of RdfCalendarDocumentation ... the links from properties to relevant decisions is nifty.

19:56:56 <libby> maybe I should ask to do this as part of BPWG, insteda of apps and tools

19:57:20 * DanC thought "apps and tools" was part of BPWG; clarify, pls?

19:57:24 <libby> yeah, I like the papertrail, but it all needs to be much more readable

19:57:46 <libby> I mean, I'm leading apps and tools so my time as part of BPWG goes on that

19:58:06 <libby> which is just documenttaion stuff really, what's out there

19:58:20 <libby> and not a huge amount of that either :/

20:00:18 <DanC> I think I've got enough feedback on my plan/idea to mail it to www-rdf-calendar and start working on it. there's still a substantial risk that I'll flake out, but if I get enough prods in the right direction, I can see us getting there.

20:00:34 <libby> :)

20:00:36 <libby> cool

20:00:51 <DanC> logger, pointer?

20:00:51 <DanC> See http://www.ilrt.bris.ac.uk/discovery/chatlogs/rdfig/2004-08-04#T20-00-51

20:01:23 <DanC> meanwhile... another idea/goal:

20:02:04 <DanC> planetRDF is cool. What if all the mentions of conferences and what-not in various blogs were sufficiently formal that planetrdf could aggregate them into a calendar?

20:02:41 <danbri> +1 good goal

20:02:46 <DanC> cross-referenced to photos. and places.

20:03:16 <aharth> danc: i am working on a mapping from dbworld mailing list->calendar

20:03:19 <DanC> I know ndw has lots of formal calendar data, and an increasing body of photo metadata.

20:03:38 <DanC> .g dbworld

20:03:40 <danbri> some event-individuation principles would be needed (and useful...) or same thing could show several times

20:03:41 <phenny> dbworld: http://www.cs.wisc.edu/dbworld/

20:04:05 <DanC> aharth, is that the dbworld you're talking about?

20:04:07 <libby> that'd be v cool danc. +1 rew identifying events too....

20:04:26 <aharth> yes, they have defined e-mail headers with metadata about call for papers

20:04:42 <aharth> should be easy to convert into some vocabulary (maybe ical, not sure yet)

20:06:07 <DanC> libby, did you automate the stuff in http://esw.w3.org/topic/RdfCalendarDocumentation at all?

20:06:18 <libby> ical is good enough for most conference stuff I think

20:06:24 <libby> nope...

20:07:16 <DanC> wow.

20:07:23 <aharth> if i get the time to understand ical :/

20:07:36 <libby> I trued to announce the foaf sumer events as rdf ical and icalendar, but messed up the RDF ical a bit

20:07:41 <libby> tried

20:07:43 <libby> http://lists.w3.org/Archives/Public/www-rdf-interest/2004Jul/0081.html

20:07:49 <libby> (see attachments)

20:08:12 <libby> a generator with an obvious payoff (like display in planetRDF) would be used, I think

20:08:24 <libby> Masahide has a rdfical generator

20:08:41 <danbri> DanC, can you remind me (what you know of) the situation re publishing something like http://esw.w3.org/topic/RdfCalendarDocumentation as a W3C Note when it contains so much IETF material?

20:08:49 <aharth> thanks libby for the example

20:09:05 <libby> this bit:

20:09:06 <DanC> let's see, danbri...

20:09:07 <libby> <dtend rdf:parseType='Resource'>

20:09:07 <libby> <date>2004-09-03</date>

20:09:07 <libby> </dtend>

20:09:23 <danbri> [[

20:09:24 <danbri> This document and translations of it may be copied and furnished to

20:09:24 <danbri> others, and derivative works that comment on or otherwise explain it

20:09:24 <danbri> or assist in its implementation may be prepared, copied, published

20:09:24 <danbri> and distributed, in whole or in part, without restriction of any

20:09:24 <danbri> kind, provided that the above copyright notice and this paragraph are

20:09:26 <danbri> included on all such copies and derivative works. However, this

20:09:28 <danbri> document itself may not be modified in any way, such as by removing

20:09:30 <danbri> the copyright notice or references to the Internet Society or other

20:09:32 <danbri> Internet organizations, except as needed for the purpose of

20:09:38 <danbri> developing Internet standards in which case the procedures for

20:09:40 <danbri> copyrights defined in the Internet Standards process must be

20:09:42 <DanC> as I recall, we need to paste that a notice from the IETF into our derivative works. otherwise, we're fine.

20:09:44 <danbri> followed, or as required to translate it into languages other than

20:09:46 <danbri> English.

20:09:48 <danbri> ]]

20:09:54 <libby> is the old way of doing it aharth. the other error is that in icalendar dates are exclusive so the end should be 3rd not 2nd

20:09:57 <danbri> -- http://www.ietf.org/rfc/rfc2445.txt

20:10:53 <DanC> but I intend to quote quite a bit less.

20:11:35 <libby> well one problem was not being able to link to parts of RFC 2445, but you fixed that danC

20:12:41 <libby>http://www.kanzaki.com/docs/sw/rdfical-a-matic.html

20:12:42 <DanC> yes

20:12:43 <dc_rdfig> C: http://www.kanzaki.com/docs/sw/rdfical-a-matic.html from libby

20:12:52 * DanC is quite happy with http://www.w3.org/2002/12/cal/rfc2445

20:12:56 <libby> C:Masahide Kanzaki's RDF ical generator

20:12:57 <dc_rdfig> Added comment C1.

20:13:07 <libby> C:|Masahide Kanzaki's RDF ical generator

20:13:07 <dc_rdfig> Titled item C.

20:13:35 <libby> C1:see also his [http://www.kanzaki.com/docs/sw/rdf-calendar.html|RDF ical documentation]

20:13:35 <dc_rdfig> Replaced comment C1.

20:14:21 <aharth> ok, i'll use masahide's rdfical format as template

20:14:26 <libby> C:and his [http://www.kanzaki.com/courier/ical2rdf|online RDF ical converter]

20:14:27 <dc_rdfig> Added comment C2.

20:14:54 <libby> it differs from the current schema and tests I think aharth....

20:15:10 <libby> ...but should be easyish to convert I guess

20:15:26 <aharth> hmm, yeah, i need something to get started

20:15:33 <libby> yep

20:16:03 * aharth has to go, cold beer is waiting in a pub :)

20:16:19 * danja yearns

20:17:11 * DanC likes the idea of a cold one... checks fridge...

20:17:56 <libby> mmmm, beer

20:18:29 <aharth> ok, see you around... cheers!

20:18:54 <golbeck>http://www.amazon.com/exec/obidos/ASIN/059600687X/jennifergolbeck

20:18:54 <dc_rdfig> D: http://www.amazon.com/exec/obidos/ASIN/059600687X/jennifergolbeck from golbeck

20:19:15 <golbeck> D| Foafbot, web services bot, and trust bot appear described in IRC Hacks by Paul Mutton (aka Jibbler)

20:19:19 <golbeck> D:| Foafbot, web services bot, and trust bot appear described in IRC Hacks by Paul Mutton (aka Jibbler)

20:19:20 <dc_rdfig> Titled item D.

20:19:22 <DanC> BLURB: A couple RDF Calendar plans/ideas

20:19:22 <dc_rdfig> E: A couple RDF Calendar plans/ideas from DanC

20:19:28 <DanC> logger, chump E

20:19:28 <DanC> E:See [http://www.ilrt.bris.ac.uk/discovery/chatlogs/rdfig/2004-08-04#T20-19-28|discussion]

20:19:29 <dc_rdfig> Added comment E1.

20:19:30 <ericP> danc, what have you done in terms of expressing APPEL in n3?

20:19:57 <DanC> BLURB: Constraints, Capabilities, Rules, and Policies

20:19:58 <dc_rdfig> F: Constraints, Capabilities, Rules, and Policies from DanC

20:20:12 <golbeck> D: Foafbot has a very general semantic web introduction, and there are about 5 hacks (out of 100 in the book) that deal with semantic web technologies and IRC

20:20:12 <dc_rdfig> Added comment D1.

20:20:51 <golbeck> D: in case the title didn't give it away, the book is part of the "Hacks" series by O'Reilly

20:20:51 <dc_rdfig> Added comment D2.

20:21:03 <DanC> F:[Dec 2001 P3P/APPEL/N3 work|http://lists.w3.org/Archives/Public/www-p3p-dev/2001Dec/0002.html]

20:21:04 <dc_rdfig> Added comment F1.

20:21:18 <ericP> tx

20:21:38 <DanC> I think there's a test case somewhere in there; let's see if we can reproduce it in the modern context, shall we, ericp?

20:22:22 <golbeck> D: Trustbot and WSbot, both hanging out on this channel, are written up with code in the book.

20:22:23 <dc_rdfig> Added comment D3.

20:22:31 <ericP> danc, i'll follow your lead

20:23:04 * DanC looks in http://www.w3.org/2000/10/swap/p3p/

20:23:37 <DanC> F:[swap/p3p stuff|http://www.w3.org/2000/10/swap/p3p/]

20:23:38 <dc_rdfig> Added comment F2.

20:24:13 * DanC tries 'make'

20:24:14 <DanC> phpht.

20:24:16 <DanC> bin/sh: line 1: java: command not found

20:24:33 <DanC> ah; it's trying to use XT. better to use xsltproc...

20:27:12 * DanC commits Makefile,v 1.3 2004/08/04 20:27:02

20:27:20 <DanC> ok, now make works. you try it?

20:27:52 <DanC> logger, chump F:

20:27:52 <DanC> F:See [http://www.ilrt.bris.ac.uk/discovery/chatlogs/rdfig/2004-08-04#T20-27-52|discussion]

20:27:52 <dc_rdfig> Added comment F3.

20:28:29 <ericP> i'm followin the dependencies now

20:28:29 <ericP> ...

20:29:36 <chaalsMEL> morning...

20:30:15 <DanC> golbeck, I've got Real Work to do; no fair distracting me with cool stuff like that!

20:30:16 <DanC> :)_

20:30:30 <golbeck> :)

20:30:43 <golbeck> sorry, danc. i accept NO responsibility for your attention span!

20:31:18 <chaalsMEL> <span class="attention" />

20:31:25 * golbeck groans

20:32:08 * ericP thinks hard about "make: execvp: ex32.xml: Permission denied"

20:34:03 <ericP> danc, did you have to do anything besides "include $(PD)/XSLTPROC.mak" instead of XSLT.mak?

20:34:29 <DanC> yes, I changed several $(XSLT)s to $(XSLTPROC)s and changed the order of the args

20:34:47 <ericP> roger

20:34:50 <DanC> and I'm not using any XSLTPROC.mak

20:36:44 <ericP> do you have an "p3pbase.rdf" ?

20:37:09 <DanC> umm..

20:37:30 <DanC> yes; dated today. so it got built somehow, I think

20:37:39 <ericP> warning: failed to load external entity "p3pbase.rdf"

20:37:39 <ericP> unable to parse p3pbase.rdf

20:37:52 <DanC> #seems Brian already did this...

20:37:53 <DanC> p3pbase.rdf: grokDataSchema.xsl ,cache-p3pbase.xml

20:37:53 <DanC> $(XSLTPROC) -o $@ grokDataSchema.xsl ,cache-p3pbase.xml

20:38:45 <DanC> I need a line or two before "warning: failed to load..."; what step is it in?

20:39:09 * DanC wonders if we should be using vnc or screen or some such

20:39:16 <ericP> xsltproc grokPolicy.xsl ex32.xml forShoppers.rdf

20:39:16 <ericP> xsltproc grokDataSchema.xsl ,cache-p3pbase.xml p3pbase.rdf

20:39:22 <ericP> <?xml version="1.0"?> \n\n\n\n\n\

20:39:27 <ericP> (for a long time)

20:39:38 <ericP> warning: failed to load external entity "p3pbase.rdf"

20:39:38 <ericP> unable to parse p3pbase.rdf

20:39:38 <ericP> make: *** [p3pbase.rdf] Error 6

20:40:20 <DanC> did you cvs update the makefile? the xsltproc call needs a -o

20:41:16 <ericP> Concluding DEFINITELY _g6->_work _g_L43C18->_g57 p->forShoppers

20:41:17 <ericP> ...

20:41:26 <ericP> Grand total 32 facts found

20:41:26 <ericP> Begining output.

20:41:34 <ericP> ends there, but looks like it worked

20:42:17 <DanC> yes, you win.

20:42:27 <DanC> now... to explain which end is up...

20:43:37 * DanC attempts to read ericp's brain to see how much of this he already knows

20:44:14 * ericP attempts to read ericp's brain

20:44:15 <DanC> F:ref [APPEL spec|http://www.w3.org/TR/P3P-preferences/]

20:44:15 <dc_rdfig> Added comment F4.

20:44:21 <ericP> kind of opaque

20:44:45 <DanC> F:esp Figure B.1: "Almost Anonymous" Ruleset

20:44:46 <dc_rdfig> Added comment F5.

20:45:33 <ericP> i futzed with p3p a while ago when i was getting into grammar-based XSLT generation

20:45:37 <ericP>http://www.w3.org/2003/07/p3p2html/

20:45:38 <dc_rdfig> G: http://www.w3.org/2003/07/p3p2html/ from ericP

20:45:50 <DanC> the way APPEL works is: a web site gives a policy; the policy is evaluated using some APPEL rules. The output is stuff like "yes, prompt the user" and "here's some text s/he should see..."

20:45:53 <mmealling> :decides to start thinking really hard about the intersection of RDF, things like Ebay, and supply chain management.....

20:46:16 * DanC waves to mmealling

20:46:21 <ericP> g:| P3P to HTML

20:47:59 <ericP> "The output is stuff like "yes, prompt the user" and "here's some text s/he should see..."" -- that's the output of swap/p3p ?

20:48:48 <DanC> here's what I think is going thru mmealling's head: "RDF? what good is that? useless. Hmm... there's that RDF thing again; maybe I better take a look. Oh, now I see. Ah! This will solve all our problems! Oh... hmm... maybe not. Well, I can see how it'll help a little. And shaving 10% off a big problem helps. OK, so RDF isn't the solution to all the world's problems, but I can see how it can help with supply chain management. Oh... and I

20:48:49 <DanC> wish Apple had used RDF rather than this plist garbage."

20:50:01 <DanC> yes, that's the output of swap/p3p, since swap/p3p sorta mimics APPEL

20:50:25 <mmealling> hehe.... sort of.....it turns out my little embedded systems company just couldn't generate the returns I needed, so I'm back to square one....

20:50:39 * sandroMIT surprised by GNOME killing x-chat when I kill the window in which I typed "xchat &".

20:50:50 <sandroMIT> Does square one involve having a job?

20:51:06 <ericP> rebirth?

20:52:36 <mmealling> sandro, yes... I'm still with VeriSign....

20:53:13 <mmealling> although I'm also doing this: masten-space.com/index2.php in my 'spare' time....

20:53:16 <sandroMIT> That's probably good. :-)

20:53:27 * DanC tries to read his own code in p3pAArules.n3

20:54:13 <DanC> I wonder if I got :x and :o confused

20:55:03 * ericP loves make's regexp support -- /:(x|o)

20:56:11 <ericP> what's the "log:racine [ log:semantics [ log:includes" idiom mean?

20:56:24 <DanC> this stuff is horrendously complex... there's a mapping of P3P policies into RDF, and I sorta manually mapped APPEL into RDF... and P3P has a whole set of structures... data schemas and whatnot.

20:56:33 <DanC> that I can answer

20:56:44 <DanC> think { <foo#bar> log:racine <foo> }.

20:57:23 <DanC> so that's saying: take the policy, thwack off the stuff after the #, fetch and parse as RDF; then, if the result log:includes {}, do XYZ.

20:57:38 <DanC> rather: conclude XYZ

20:58:49 <DanC> capice?

20:59:42 <ericP> ":o p3pr:categories [ a _:InterestingCategory ]." is inventing a new node for this thing with categroies?

21:00:39 <DanC> er... no, that's in a rule premise, right?

21:00:53 <DanC> I can't figure out what I meant by :o

21:01:00 <ericP> oh eyah

21:01:26 <ericP> looks like a disconnected graph

21:02:08 <DanC> I can't figure out why that rule fires.

21:02:11 * DanC noodles...

21:04:02 <ericP> can we pick this up in fifeteen minutes?

21:04:30 <ericP> rather i? perhaps you can figure out this conundrum

21:05:14 <ericP> brb in 15

21:06:49 <DanC> I think I figured it out. http://www.w3.org/2000/10/swap/p3p/p3pAArules.n3 1.3

21:07:13 <DanC> (and I tweaked the makefile to use --data. Makefile,v 1.4)

21:19:19 <ericP> back

21:19:49 <DanC> hi

21:19:59 <GiovanniT> hi from me as well

21:20:25 <ericP> howdy

21:20:27 <ericP> "Concluding DEFINITELY _g16->_work p->forShoppers"

21:20:50 <ericP> Concluding DEFINITELY _g18->_work _g_L72C18->_g72 p->forShoppers

21:22:14 <ericP> real vocab example is in forShoppers.rdf

21:22:50 <DanC> yes, forShoppers.rdf is a P3P policy converted to RDF

21:23:41 <DanC> does the rule we were staring at make sense now? p3pr:data [ p3pr:categories :o ] ... :o a _:InterestingCategory.

21:24:35 * ericP grabs a napkin to sketch the picture

21:24:47 <ericP> (of the antecedent)

21:27:41 <ericP> mostly

21:28:00 <ericP> still trying to get the effects of log:racine into my brain

21:28:03 <ericP> will read back...

21:29:20 <DanC> log:racine relates <forShoppers.rdf#forShoppers> to <forShoppers.rdf>. I think it's unnecessary; I think log:semantics automatically disregards #foo

21:30:45 <ericP> and then you look for statement.data.categories.(one of physical, online, uniqueid, financial, other-category)

21:32:18 <DanC> yes

21:33:21 <ericP> aha! <forShoppers.rdf#forShoppers> a p3pr:Policy.

21:33:48 <DanC> oops; I didn't realize that would have been a useful clue

21:33:53 <ericP> no prob

21:34:03 <ericP> it helps to make a connection from time to time

21:36:24 <DanC> ok, now that I think we've worked thru the essential details, the big picture is: P3P policies are data. RDF is pretty good for data. Lots of different kinds of data. It's clear what happens when you mix RDF data from multiple sources.

21:36:46 <DanC> ... and policies are going to usefully use data from lots of sources; so let's write them using RDF rules.

21:41:07 <ericP> "policies are going to usefully use data from lots of sources" probably needs to be backed up

21:41:22 * sh1mmer waves to ericP and danbri

21:41:24 <sh1mmer> er

21:41:26 <sh1mmer> danC

21:41:39 <ericP> providing that scenario will, i think, argue well for RDF

21:42:10 <DanC> hmm... I have a bunch of timezone data in RDF; you're more than welcome to make up an example that mixes that in.

21:42:19 <ericP> XSLT could be used to solve the problem you just solved, but it would be more brittle

21:42:24 <DanC> or you could mix in some foaf foo.

21:42:44 <DanC> "but it would be more brittle" is the hypothesis, yes.

21:42:52 <ericP> i expect the audience to be a bit conservative on mixing data.

21:42:55 <DanC> it would be great if you could argue that effectively in your position paper.

21:43:38 <DanC> hmm... well, I gather the WS constraints+capabilities crowd is bumping up to the mixing issues in real life. But PLH/Hugo would know better.

21:44:16 <ericP> generally, security/policy people don't subscribe to the "look around" approach to determining policies.

21:44:23 <ericP> yeah, i'll talk to them

21:44:31 <ericP> do you remember when this is due?

21:44:44 <ericP> plh is on vacation for another 1.3 weeks, i think

21:44:50 <DanC> I remember minuting the due date...

21:46:12 <ericP> not in 5.1.0.14.2.20040803135632.01a85b48@127.0.0.1

21:46:16 <ericP> looking back...

21:47:48 <DanC> F:[workshop CFP|http://www.w3.org/2004/06/ws-cc-cfp.html] says papers are due August 27

21:47:49 <dc_rdfig> Added comment F6.

21:48:46 <ericP> ok, shouldn't be a prob then

21:48:54 <ericP> i;ll poke hugo

21:50:57 <ericP> he's at a WSDL f2f now

21:53:45 <DanC> so I guess you're briefed on my APPEL work... I was hoping to try a similar trick with ws-policy

21:53:53 <DanC> have you slurped up ws-policy yet?

21:56:09 <ericP> not totally

21:56:17 <ericP> bits of it from converstations

21:56:56 <ericP> how bad does P3P hurt because it's not written in RDF?

21:57:26 <ericP> has anyone (apart from RDF geeks like us) noticed?

21:58:17 <ericP> i can imagine a somewhat pedestrian argument just from the perspective of leaning on composability

21:59:17 <ericP> if you have a policy split amoungst a few files, RDF can make the combination a trivial process

21:59:20 <ericP> this has been somewhat interesting to the PICS folks

22:00:11 <ericP> doesn't deal with extensibility of data or evolving/refining semantics

22:00:50 <ericP> i guess that was your argument anyways, but if feels like i'm solving a more banal problem.

22:05:21 <DanC> no one application hurts because it's written in RDF, just like no one pre-web application hurt because it wasn't part of the web.

22:05:28 <DanC> s/written/not written/

22:06:20 <DanC> cf "Frameworks and Applications" section of http://www.w3.org/2000/08/w3c-synd/

22:07:14 <DanC> i.e. http://www.w3.org/2000/08/w3c-synd/#Frameworks

22:07:34 <ericP> it's possible that ws-policy has been more complicated than necessary because re-using existing tags isn't as easy in XML land as reusing properties in RDF

22:07:47 <DanC> yes, that's my hunch

22:08:12 <ericP> ie, if P3P were RDF, it would have been a no-brainer for ws-policy to re-use the relevent bits

22:08:37 <DanC> umm... don't overemphasize the relevance of P3P here.

22:08:59 <DanC> I'm pretty sure there are plenty of mixing issues in the "constraints and capabilities" world without mixing in P3P...

22:09:13 <DanC> ... in fact, I gather ws-policy pretty much re-invents RDF to address them.

22:09:44 <ericP> roger

22:09:54 <DanC> Some days, I think: If ws-policy is gonna get deployed widely, and it can express what RDF/XML expresses, we might as well switch.

22:10:23 <DanC> But I've tried that idea out on a few people, and they said that ws-policy doesn't ground all its terms in URI-space. which is a non-starter.

22:10:58 <ericP> i guess it's time to read ws-policy

22:11:17 <DanC> yes, you do that, would you? that way, I can avoid it for another while. ;-)

22:13:34 * ericP reads http://www-106.ibm.com/developerworks/library/ws-polfram/

22:15:55 <ericP> interesting spec, it uses XPath to identify the parts of the grammar in the description

22:18:22 <DanC> F:[WSPolicy|http://www-106.ibm.com/developerworks/library/ws-polfram/]

22:18:23 <dc_rdfig> Added comment F7.

22:18:47 <DanC> F:I'm hoping the [Frameworks and Applications|http://www.w3.org/2000/08/w3c-synd/#Frameworks] point will resonate.

22:18:47 <dc_rdfig> Added comment F8.

22:19:41 <DanC> where does http://www.w3.org/2003/07/p3p2html/ come from? it's not signed and it has no links to context

22:26:26 <danbri> ACLs suggest eric

22:26:30 <danbri> as does the style

22:26:31 <golbeck_> golbeck_ is now known as golbeck

22:42:36 <sandroMIT> Funky -- I didn't know XHTML could be valid without namespaces.

22:42:47 <sandroMIT> (re p3p2html)

22:56:52 <sh1mmer> hey sandro

22:57:02 <DanC> me neither. in fact, I sure hope it can't.

23:01:15 <sh1mmer> sandroMIT did you write anything up about what did Sean call em... *thinks* something constellations

23:02:00 <sh1mmer> "unambiguous property contellation"

23:04:57 <DanC> sh1mmer, sounds like a concept we treated in our diff paper... http://www.w3.org/DesignIssues/Diff

23:05:14 <sandroMIT> No, I've never written it up, sh1mmer .

23:05:37 <sh1mmer> danC Sean and I were talking about it

23:05:39 <sandroMIT> compound keys.

23:05:43 <DanC> "Recursively, v is functionally ground if it is a name (URI or literal) or a bnode with a functional term label."

23:05:59 <sh1mmer> I was thinking about doing some work on it for my disertation

23:06:07 <sh1mmer> sandroMIT right.

23:06:40 <DanC> compound... we only treated single-column keys, though we treated chains of them.

23:06:54 <sh1mmer> I was thinking off investigating strategise to find the best compound on a given graph

23:07:10 <DanC> there's a "keys for free in description logics" paper that's lodged in my brain near the "compound keys" synapse.

23:07:11 <sh1mmer> or to identify the possible compound keys that could be used

23:08:33 <sh1mmer> I just want to make sure I am not repeating work that has already been done

23:26:23 <DanC> hmm... sounds like joe lambda's cousin or something... http://www.techquila.com/blog/archives/000039.html