Semantic Web Interest Group IRC Chat Logs for 2003-06-25

01:16:46 <bitsko> when was RDF first proposed, in the name and manner we use it today?

01:22:50 <sbp`> the first draft was in '97: http://www.w3.org/TR/WD-rdf-syntax-971002/

01:47:21 * nb is back (gone 02:46:56)

02:11:42 * nb is away: zzz

02:57:13 <golbeck_> golbeck_ is now known as golbeck

04:12:08 <sbp`-> sbp`- is now known as sbp`

04:16:19 <sbp`-> sbp`- is now known as sbp`

07:20:41 <mdupont> mdupont is now known as mdupont-work

08:02:58 * nb is back (gone 05:51:19)

08:55:18 <arnarl> hi

10:44:33 * nb is away: lunch

11:30:13 * nb is back (gone 00:45:43)

13:26:39 <MorbusIff> MorbusIff is now known as Morbus

13:31:31 <_joshua> blogland is busy trying to reinvent RDF yet again

13:32:36 <dajobe> could be

13:32:50 <dajobe> you mean ruby's pie?

13:33:10 <_joshua> yes

13:33:13 <_joshua> do you agree?

13:33:21 <_joshua>http://intertwingly.net/wiki/pie/

13:33:22 <libby> eh?

13:33:22 <dc_rdfig> A: http://intertwingly.net/wiki/pie/ from _joshua

13:33:53 <_joshua> A:| The Echo Project: standardizing a system to syndicate, archive, and edit weblogs

13:33:53 <dc_rdfig> Titled item A.

13:34:14 <dajobe> or, moving descriptions of web content around using the web

13:34:18 <_joshua> A: looks like they're reinventing RDF.

13:34:19 <dc_rdfig> Added comment A1.

13:34:55 <_joshua> A: http://intertwingly.net/wiki/pie/SyntaxConsiderations

13:34:56 <dc_rdfig> Added comment A2.

13:36:24 <edd> A: I'm [not so sure|http://usefulinc.com/edd/blog/2003/6/25#13:44] they're reinventing RDF. The trouble with RDF for this application is that it works fine for metadata, but falls down a bit where the actual content itself is concerned. There've been some tortuous attempts to carry around HTML markup around inside RSS.

13:36:25 <dc_rdfig> Added comment A3.

13:37:17 <edd> A:It will be interesting to see how it all develops. There are many obstacles, most of them down to process and politics.

13:37:18 <dc_rdfig> Added comment A4.

13:38:43 <Morbus> A:I'm, personally, staying out of it.

13:38:44 <dc_rdfig> Added comment A5.

13:39:23 <edd> well advised, Morbus

13:39:39 <Morbus> edd: more so because I can't stand web-based discussions.

13:39:50 <Morbus> edd: it's a heckuva lot harder for me, compared to email threads.

13:42:21 <dajobe> and it's harder to go along and edit other people's comments out

13:43:11 <edd> Yes, danny ayers already seems to have suffered from that, from my brief reading

13:43:42 <dajobe> I was refering to the owner of the wiki above, who did that to a third person

13:43:55 <Morbus> edd: what happened to danny?

13:44:10 <Morbus> ah.

13:44:25 <dajobe> ugh, what have I done. Dragging blogosphere sillyness here.

13:44:31 <Morbus> heh, heh.

13:44:39 <edd> from http://intertwingly.net/wiki/pie/SyntaxConsiderations, "[DannyAyers] (Refactor *not* ok for now as someone just interpreted that as 'deleteme')"

13:44:52 <Morbus> cute.

13:45:23 <edd> i fear Pie's just an excuse for everyone to repeat their entrenched positions.

13:45:42 <edd> Join that with attempting to find consensus, and you have something that doesn't go anywhere fast.

13:46:18 <edd> For instance, I'm surprised that whether or not to use XML was even an issue at hand.

13:46:59 <dajobe> aaron still trying to foist ASCII solutions

13:47:13 <dajobe> hmm, Sam will be at OSCON

13:51:55 <Morbus> heh, edd: I just suggested "or YAML" on Aaron's comment at the bottom of the page ;)

13:52:17 <deltab> YAML simple?

13:52:52 * Morbus shrugs.

13:52:54 <Morbus> it's ascii.

13:53:06 <deltab> that seems to me like calling SOAP lightweight

13:53:14 <dajobe> or RSS2 simple

13:54:36 * Morbus thinlines lips.

13:55:13 <_joshua> Hmm. I agree. RDF doesn't encapsulate content very well.

13:55:34 <sethl> why not use parseType="Literal"?

13:55:37 <deltab> I don't think it's meant to

13:55:42 <_joshua> Right.

13:55:52 <deltab> it's not Resource Encapsulation Framework

13:55:55 <sethl> true

13:56:07 <_joshua> that's an ongoing irritation for me; that I can't use other already-defined DTDs and whatnot

13:56:21 <_joshua> it'd be useful if i could tote around a bit of GML for example and add metadata to it or whatever

13:56:43 <edd> Morbus: i assume that was a joke :)

13:56:44 <dajobe> maybe you should lookat Namespace Routing Language - see the xmlhack.com article

13:57:00 <_joshua> Morbus: This is the Everybody Except For Dave Winer Cabal

13:57:24 <_joshua> Dajobe: interesting

13:57:42 <Morbus> yeah, they mentioned winer on the page RoadMap page for a split second.

13:57:51 <Morbus> then took it off. then put him back on as a community leader. than removed it.

13:58:08 <edd> I love how nobody learns anything from history, ever.

13:58:17 <_joshua> I guess it doesn't have to be literal XML; collapsed into CDATA or whatever

13:58:24 <Morbus> I dunno, man. If you weren't sitting there watching, you'd never have seen it, and I don't consider that discussion. I consider that reinventing history, like Winer does with old blog posts that stir the pot.

13:58:34 <Morbus> that's why I prefer email discussions.

13:58:55 <_joshua> Yes

13:59:14 <Morbus> but, I guess that's the whole point of a Wiki.

13:59:15 <_joshua> an indelible record is a compelling thing.

13:59:19 <Morbus> so I'm really complaining about Wiki's.

13:59:24 <Morbus> but even some Wiki's provide CVS like diff's.

13:59:32 <Morbus> I couldn't find that feature on PIE though.

13:59:36 <sethl> twiki does

13:59:38 <_joshua> Not only that, but Wikis are very hard to read

13:59:48 <dajobe> yes

14:00:00 <_joshua> good as an information store, poor as documentation or narrative

14:00:16 <_joshua> "where do I start anyway" etc

14:00:25 <sethl> the main prob. we found w/ wikis is that there is no central information architect. That is, with everyone just adding stuff on every page, finding the /right/ info becomes difficult

14:00:40 <_joshua> Morbus, but put the problems of wikis aside for the moment

14:00:51 <sethl> our team kept saying "It's on wiki!" and the other members would say "yeah, but where?"

14:01:00 <_joshua> heh

14:01:16 <Morbus> heh, heh.

14:01:37 <dajobe> how does the weblog world come to consensus?

14:01:38 <sethl> sure, you can search for it, but browsing is powerful, and very difficult on a wiki

14:01:42 <dajobe> when the loud people shout down the rest?

14:01:57 <dajobe> or enough people implement things (trackback, rss discovery etc.)

14:02:04 <_joshua> indeed

14:02:24 <dajobe> I mentioned oscon in a couple of weeks, wondering if enough people will be there

14:02:52 <dajobe> but not open to those who can't, of course

14:03:06 <sethl> are you attending?

14:03:11 <dajobe> yes

14:03:27 <sethl> nice, I've heard good things about past conferences

14:04:01 <sandro_> There's text in RDF Semantics talking about rdf:first/etc which is absolutely incompatible with semantics-are-defined-by-URI, I think. What to do about that......

14:04:06 <sandro_> sandro_ is now known as sandro

14:05:51 <sandro> I think the only way to make the semantics web work given how "RDF Semantics" imagines things is by saying all RDF must be accompanied by a set of extensions/ontologies which are being committed to by the author.

14:06:34 <sandro> At least if you want to use any URIs beyond those defined by W3C, or use them in some other way (as RDF Semantics says you can).

14:06:43 <sethl> do you mean, without those ontologies, it's impossible to define the semantics of the RDF document?

14:07:47 <dajobe> all very interesting, but we were discussing Pie/Echo...

14:07:54 <sandro> I mean if I say in RDF "<http://.../sethl> rdf:type <http://.../SmartGuy>" but don't tell you which extensions/ontologies I'm using, you can't possible know what I meant.

14:08:30 <sandro> That was at least two minutes ago Dajobe, jeez.

14:21:15 <libby> BLURB: RDF calendar meeting at 1600UTC today

14:21:16 <dc_rdfig> B: RDF calendar meeting at 1600UTC today from libby

14:21:38 <libby> B:see [http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Jun/0008.html|agenda] (see thread)

14:21:39 <dc_rdfig> Added comment B1.

14:22:10 <libby> B:[http://www.timeanddate.com/worldclock/fixedtime.html?day=25&month=6&year=2003&hour=16&min=0&sec=0&p1=0|2003-06-25, 1600UTC]

14:22:10 <dc_rdfig> Added comment B2.

14:35:20 <_joshua> anyone have handy urls for one of danc's calendaring exploits?

14:35:37 <_joshua> i want to send someone an example with using cwm, n3 etc for this

14:36:19 <sandro> yes

14:36:49 <sandro> Here: http://www.w3.org/2000/10/swap/pim/travel

14:37:43 <_joshua> impressive. that was the one i was thinking abot

14:37:52 <deltab> try the Calendar Chump at http://swordfish.rdfweb.org/calendar/links/

14:38:53 <_joshua> thanks

14:47:33 * tim-cape_ prints off RDF Semantics to see what Sandro means

14:47:44 <DanC> ugh; another evolution crash. Is the mozilla mailer out yet? the analog to firebird?

14:48:05 <edd> DanC: evo crashes? which ver? it's ages since i had an evo crash :(

14:48:10 <danbri> thunderbird?

14:48:35 <_joshua> I use an exceedingly low-tech PIM. I have a bunch of text files and grep.

14:48:39 <DanC> yeah; thunderbird. anybody using it?

14:48:47 <danbri> not me. went back to Mutt.

14:49:02 <DanC> grep is a wicked PIM tool. *very* hard to beat.

14:49:13 <_joshua> hey DanC: I got the camera thingy working. wanna try?

14:49:26 <DanC> not just now, _joshua...

14:49:34 * sandro prefers i-search to grep, usually. But i-search pales next to google, sometimes.

14:49:38 <_joshua> also freeform text search. "cat dad" to get my dad's info

14:50:14 <_joshua> DanC: Ok. It mostly works actually, but the interface is just ugly. i'd need your sending email address to set you up

14:50:41 <sandro> Tim, the editor's draft of RDF Semantics is: http://www.w3.org/2001/sw/RDFCore/TR/WD-rdf-mt-20030117/

14:51:07 <dajobe> no it isn't

14:51:27 <_joshua> I think I actually went with all your suggestions, of the stuff I implemented.

14:51:30 <sandro> According to http://www.w3.org/2001/sw/RDFCore/ it is.

14:51:44 <_joshua> Now we need to implement FOAF emissions, so you can have WearableFOAFGizmo ro whatever

14:51:53 <sandro> Who should I trust, Dave or the Web? :-) :-)

14:52:03 <DanC> edd, evolution Version: 1.4.0-3

14:52:16 <dajobe> that document's date shows it is old

14:53:16 <edd> DanC: Strange. Sorry to hear it's so bad for you.

14:53:17 * DanC finds an 1.4.0-4 available

14:53:44 <dajobe> 1.4.0-4 never gets past the splash screen for me, just crashes

14:56:21 <lilo> [Global Notice] Hi all. One of our more stable main rotation servers has apparently decided to chime in with a problem, please bear with us.

15:01:41 <DanC> CRAP. -4 build of evolution still exhibits the drag-n-drop crash bug I reported.

15:03:09 <sandro> dajobe, "18 May 2003" doesnt seem that old to me. Not compared to "23 January 2003" at least.

15:18:16 * DanC boots up for rdf-calendar meeting; finds a boatload of mail in my "unknown senders" inbox

15:28:39 * tim-cape_ understands calendaring meeting starts on this channel in about 30 minutes

15:29:51 <libby> yep

15:29:54 <libby> .time

15:29:54 <datum> Wed, 25 Jun 2003 15:29:54 GMT

15:42:19 <DanC> B:[http://rdfig.xmlhack.com/2003/06/11/2003-06-11.html#1055344744.381612 weblog of previous meeting, 11 June]

15:42:20 <dc_rdfig> Added comment B3.

15:42:34 <DanC> B3:[http://rdfig.xmlhack.com/2003/06/11/2003-06-11.html#1055344744.381612|weblog of previous meeting, 11 June]

15:42:34 <dc_rdfig> Replaced comment B3.

15:43:55 <DanC> B:nearby: [http://lists.w3.org/Archives/Public/www-rdf-calendar/|www-rdf-calendar list archive], [http://esw.w3.org/topic/RdfCalendar|RdfCalendar wiki topic], [http://www.w3.org/2002/12/cal/|RDF Calendar Workspace]

15:43:55 <dc_rdfig> Added comment B4.

15:44:33 * DanC works on his actions...

15:47:52 <libby> BLURB: calendar meet agenda item C: outstanding actions

15:47:52 <dc_rdfig> C: calendar meet agenda item C: outstanding actions from libby

15:48:23 * DanC prefers not lumping outstanding actions together

15:48:49 <DanC> chump each of the agenda from which they came, please

15:49:12 <libby> I was thinking the chump urls, I have a list of those

15:49:18 <DanC> C:|handling prodid (calendar agendum C)

15:49:18 <dc_rdfig> Titled item C.

15:49:49 <DanC> BLURB:representing attendance (calendar agendum D)

15:49:49 <dc_rdfig> D: representing attendance (calendar agendum D) from DanC

15:49:56 <DanC> D:news on DanC's action?

15:49:56 <dc_rdfig> Added comment D1.

15:50:40 <libby> C:see [initial weblong action item|http://rdfig.xmlhack.com/2003/02/26/2003-02-26.html#1046279854.884486]

15:50:41 <dc_rdfig> Added comment C1.

15:51:35 <DanC> D: from [http://rdfig.xmlhack.com/2003/04/23/2003-04-23.html#1051113769.199079|23Apr]

15:51:35 <dc_rdfig> Added comment D2.

15:51:47 <libby> beat me to it

15:51:55 <DanC> I don't recall the context for these "ACTION libby look at skical optimeset and report back; ACTION libby send a mail to rdf-interest on interprettatyon properties and datatypes..."

15:52:09 <libby> I've got them, hang on

15:52:13 * DanC follows link to previous meeting

15:52:19 <DanC> ok, I'll work on D.

15:54:11 <bsmith> bsmith is now known as ghuo

15:54:17 <DanC> I did some work on an email schema recently, for WebOnt issue tracking. anybody want it on today's agenda?

15:55:05 <libby> BLURB: repeating events - optimset in skical, repeating in icalendar, with respect to opening hours in particular

15:55:05 <dc_rdfig> E: repeating events - optimset in skical, repeating in icalendar, with respect to opening hours in particular from libby

15:55:10 <danbri> email schemas come up regularly, so i'd be interested. it's very date-oriented data...

15:57:30 <libby> em: see [http://ilrt.org/discovery/chatlogs/rdfig/2003-05-14.html#T16-58-34|may 14 chatlog] [action for libby on skical optimset|http://rdfig.xmlhack.com/2003/05/14/2003-05-14.html#1052921472.659227], [previous action about chefmoz to icalendar|http://lists.w3.org/Archives/Public/www-rdf-calendar/2003May/0005.html]

15:57:38 <libby> E:see [http://ilrt.org/discovery/chatlogs/rdfig/2003-05-14.html#T16-58-34|may 14 chatlog] [action for libby on skical optimset|http://rdfig.xmlhack.com/2003/05/14/2003-05-14.html#1052921472.659227], [previous action about chefmoz to icalendar|http://lists.w3.org/Archives/Public/www-rdf-calendar/2003May/0005.html]

15:57:38 <dc_rdfig> Added comment E1.

15:57:51 * libby would be interested to in the email schema

15:58:36 <libby> BLURB:InterpretationProperties, esp as applied to timezones

15:58:36 <dc_rdfig> F: InterpretationProperties, esp as applied to timezones from libby

15:59:00 <libby> F:see [14th may action|InterpretationProperties, esp as applied to timezones]

15:59:00 <dc_rdfig> Added comment F1.

16:00:03 <libby> E:| repeating events - optimset in skical, repeating in icalendar, with respect to opening hours in particular (agenda item E)

16:00:04 <dc_rdfig> Titled item E.

16:00:16 <libby> F:|InterpretationProperties, esp as applied to timezones (agenda item F)

16:00:16 <dc_rdfig> Titled item F.

16:00:21 <libby> .time

16:00:21 <datum> Wed, 25 Jun 2003 16:00:20 GMT

16:00:22 <mortenf> libby, that F1 comment seems wrong?

16:00:38 <libby> really?

16:00:46 <libby> ------RDF calendar meeting-----

16:01:07 <DanC> logger_2, pointer?

16:01:07 <DanC> See http://ilrt.org/discovery/chatlogs/rdfig/2003-06-25#T16-01-07

16:01:09 <mortenf> a link to "14th may action"

16:01:10 <libby> mortenf, I had an action to email rdf-interest about datatypes; I havn;t done it yet though

16:01:27 <DanC> B:[http://ilrt.org/discovery/chatlogs/rdfig/2003-06-25#T16-01-0|chatlog]

16:01:27 <dc_rdfig> Added comment B5.

16:01:28 <libby> oope yep

16:01:29 <mortenf> datatypes?

16:01:50 <mortenf> ah, yeah, got it, that's ok.

16:02:14 * DanC wonders if round-trip testing is or should be on the agenda

16:02:19 <libby> F1:see [14th may action|http://rdfig.xmlhack.com/2003/05/14/2003-05-14.html#1052921345.552964] InterpretationProperties, esp as applied to timezones]

16:02:19 <dc_rdfig> Replaced comment F1.

16:02:30 <mortenf> methinks there's no news on that?

16:02:32 <libby> yep it shoudl be danc, I'm behind on my chumping

16:02:53 <DanC> D:a start: [http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Jun/0011.html|representing attendance, plans to attend]

16:02:54 <dc_rdfig> Added comment D3.

16:03:46 <libby> BLURB: Round-tripping RDF ical tests, toIcal.py (agenda item G)

16:03:47 <dc_rdfig> G: Round-tripping RDF ical tests, toIcal.py (agenda item G) from libby

16:04:00 <DanC> B:[http://www.w3.org/People/Connolly/|Dan Connolly] aka DanC attending

16:04:00 <dc_rdfig> Added comment B6.

16:04:15 <libby> ok, hello everyone who is here for the calendar meet, apologies to those not here for that - we'll be about 90 inutes

16:04:26 <libby> see http://rdfig.xmlhack.com for the agenda

16:04:39 <mortenf> B:[http://purl.org/net/morten/|Morten Frederiksen] attending

16:04:39 <libby> please add your name to B if attending like DanC just did

16:04:40 <dc_rdfig> Added comment B7.

16:05:03 <libby> B:[libby Miller|http://ilrt.org/discovery/libby/] attending

16:05:03 <dc_rdfig> Added comment B8.

16:05:06 <DanC> the agenda is negotiable, right libby?

16:05:16 <libby> as faras I'm cocerned, yes

16:05:16 <danbri> B:[http://www.w3.org/People/DanBri/|Dan Brickley] attending

16:05:16 <dc_rdfig> Added comment B9.

16:05:25 <libby> if anyone woudl like to add anything, that's great

16:06:01 <libby> does anyone have any other suggestions for the agenda?

16:06:29 * mortenf is happy (if currently over-linked)

16:06:42 <DanC> I did some work on an email schema; tangentially related. I'm happy to present it if anybody else wants it on the agenda

16:07:07 <libby> yes please danc

16:07:17 <danbri> +1

16:07:25 <DanC> BLURB:an email vocabulary/schema and tools (agendum H)

16:07:25 <dc_rdfig> H: an email vocabulary/schema and tools (agendum H) from DanC

16:07:30 <mortenf> DanC, is http://www.w3.org/2000/10/swap/pim/travel your latest on calendaring?

16:07:45 <DanC> er... pretty much, mortenf

16:07:50 <danbri> esp re the dates, whether the 'rdf lets you mix your data super easy' story came true for you, or if theres a huge mismatch in how dates are representated

16:08:05 <libby> shall we get started with C?: handling prodid (calendar agendum C)

16:08:12 <DanC> hmm...

16:08:36 * DanC might like to start with an agendum about how much more energy we have for this effort

16:08:47 <libby> I'd second that

16:08:51 <DanC> or perhaps start with the round-tripping agendum

16:09:25 * DanC nominates libby to chair

16:09:40 <libby> I'm happy with either of those

16:09:48 <libby> lets go with the energy thing

16:10:09 * libby doesnt have much more for it at the m,ometn. I know danc doesn't

16:10:19 * danbri yawns, stretches, do we have to...? ;)

16:10:27 <libby> ...because busy with webont etc

16:10:42 <DanC> let's discuss it in the context of G.

16:10:53 <mortenf> at the moment or in general?

16:11:26 <libby> ok, want talk about G, Danc?

16:11:30 <DanC> I feel guilty showing up for these things empty handed on all my actions.

16:11:45 <DanC> and I don't like that.

16:12:27 <DanC> in theory, a threshold for "done" is that we have a test suite that we're all happy with.

16:12:29 <danbri> OK I see two things an RDF Interest Group task force can do. Both useful; but easily confused with each other. (a) try to get some specific scoped thing done (eg. compare rdql/squish query engine behaviour; represent ical in rdf) (b) provide a more scoped topic-based forum for SW developer to meet and swap notes on some corner of semweb work. The latter can be ongoing for as long as the IG and shared interest exists; the former is more sui

16:12:29 <danbri> deadlines, but suffers from not having WG-like levels of energy and organisation thrown at it.

16:12:30 <DanC> (re G)

16:12:30 <mortenf> well, i think we anticipated this when meeting last?

16:12:49 <DanC> anticipated which?

16:12:56 <mortenf> the non-actions...

16:13:43 <mortenf> as in: i realize you're busy atm, no need to be sorry. :)

16:13:47 <danbri> 'task force' suggests a task... that we're trying to succeed at. Not just a topical-chat.

16:13:49 <DanC> we have some test data, but it's not really useful in the XP sense. we don't have a regression test infrastructure so that you can run all the tests to see whether you've broken anything

16:13:55 <danbri> yep, we had advance warning you were busy.

16:14:27 <DanC> I'd like to see the test suite mature, but I can't see preempting any of my existing obligations to get that done.

16:14:34 <libby> what would such an infrastrcure require danc?

16:14:53 <DanC> Meanwhile, this calendar stuff is getting deployed in production at W3C (I wish the users were here to discuss it) so I'd really like the tests to be there.

16:15:12 <libby> do you think we need more tests? different tests?

16:15:13 <DanC> it would require some way to run the tests and have the machine tell you whether they passed.

16:15:22 <danbri> perhaps we could collectively agree that this is a slow-burning field of activity, given current effort levels. And not take slow progress as sign of doom/failure...?

16:15:26 <DanC> we don't have *any* automated tests. just test input data

16:15:52 * danbri is only recently integrating IETF-format icalendar tools into his life (thanks to the Mac iCal application)

16:16:06 * libby not sure what is required for automated tests

16:16:11 <danbri> ...so until lately, rdf-cal and roundtripping has been too much an academic interest for me

16:16:39 <DanC> libby, for cwm, we have a regression test suite. all cwm developers agree that all the tests have to pass every time you check in your code.

16:16:57 <DanC> you type 'make tested' or some such, and if a test fails, the output makes that clear

16:16:58 <danbri> for an ontology, what's the equivalent?

16:16:59 <libby> so need input/output.....

16:17:23 <danbri> ...any changes to the vocab still need to support full roundtripping of test dataset into and out of RDF

16:17:32 <danbri> (but that assumes a particular conversion tool implementing the mapping)

16:18:01 <DanC> we have some .ics files and a tool to convert them to .rdf ; we even have the beginnings of expected .rdf results. We just don't have a mechanism to compare the actual results with the expected results and complain if they're different

16:18:02 <libby> at the moment we have 3 special tests which generate the schema, and some others

16:18:24 <DanC> we have no testing of the .rdf->.ics mapping

16:18:42 <libby> ah ok, danc. I use jena for that. cwm can do that too right? (compare actual and expected)

16:18:50 <DanC> cwm could perhaps do it

16:18:53 <libby> re ics->rdf I mean

16:19:00 <mortenf> what is used for the rdf tests re cwm?

16:19:10 <danbri> similar methodology to the rdf query and parser testing? ie needs graph-compare facilities...?

16:19:19 <DanC> diff, mortenf

16:19:23 <mortenf> heh.

16:19:31 <libby> that's probbaly the simplest way danbri

16:19:35 <libby> simple in some sense anyway

16:19:41 <DanC> graph-compare is a better tool than diff. dunno if diff would suffice, though

16:20:09 <libby> diff would be a bit annoying because of whitespace but would probbaly do

16:20:30 <DanC> whitespace, insignificant ordering, etc. yes.

16:20:58 <DanC> the question of diff vs. graph compare is also a difference in the specification of the target. I think graph compare is the more relevant target.

16:21:01 <libby> yep the ordering mightnot be too tricky because we both process one line at a time

16:21:11 <libby> I agree

16:21:47 <DanC> ok, so we all agree We Should have regression tests. (I'm not sure how we should test the schema, btw... using swap/util/validate maybe?). Now what?

16:21:47 <mortenf> methinks an NTriple "diff" shouldn't be to hard to do, but that still leaves .ics comparisons

16:21:54 <libby> danc, if there is a need for the testing, and you are too busy, perhaps someone using theis stuff at w3c culd be persuaded to come to these meetings, do tests etc?

16:22:15 <DanC> well, I tried, libby.

16:22:28 <libby> darn :)

16:22:37 <DanC> (not very hard, but if I had time to try hard...)

16:23:18 * tim-cape_ tunes in after being distracted, starts to catch up

16:24:10 <danbri> I agree such testing tools would be wonderful, and the next step forwards...

16:24:19 <libby> we could maybe do with moving to the next topic soon

16:24:29 <mortenf> i could have a go at that ntriple "diff", if it would be of use?

16:24:41 <DanC> I partly justify my time on this under the SWAD-EU banner... is there some upcoming SWAD-EU milestone where we/you could look at the priority of this w.r.t. other SWAD-EU stuff?

16:24:49 <mortenf> (and if noone says it's too hard)

16:24:52 <danbri> It sounds to me that (a) we all remain interested (b) we're all overcommitted (c) slow progress isn't evidence of low-interest...

16:24:54 <libby> what language tool do you use mortenf?

16:25:10 <mortenf> i'd prbably go with Perl, but i'd like to learn Python, so...

16:25:20 <dajobe> mortenf: sandro tried this too a while back IIRC, also big problems when there are lots of bnodes

16:25:28 <mortenf> hmm...

16:25:47 <danbri> I have a sub-project for the summer I want to get going that is cal-related, with Kate and CarolineM (project manager and admin), re extracting cal events from mailing lists (conference announcements etc).

16:25:48 <libby> maybne you end up doing a proper graph match tool in the end...?

16:25:50 <DanC> mortenf, the ntriple diff tools are available. What we need is somebody to integrate them into the workspace (http://www.w3.org/2002/12/cal/). i.e. somebody to go thru the trouble of getting cvs write access and do all the detail work.

16:25:51 <mortenf> that's what i though would be the showstopper.

16:26:10 <danbri> that was my plan to do a little cal stuff... maybe could squeeze testing into that (though I was hoping they'd do most the work)

16:26:33 <libby> I could do some of this stuff probbaly (on the website)

16:26:50 <libby> I'm always a little worried about doing the wrong thing!

16:26:51 <mortenf> ok, so that's a makefile doing testing - i could try it locally and hand it over?

16:26:52 <danbri> How about pick a realistic, if long schedule so folks no what to expect

16:27:06 <DanC> the wrong thing is *much* better than nothing, libby!

16:27:12 <mortenf> what else would be in "detail work"?

16:27:20 <DanC> I thought lack of progress on your actions was due to lack of time.

16:27:23 <danbri> libby, re website, the main thing is directories are hard to remove cos CVS is limited; anything else we can back out the changes.

16:27:42 <danbri> don't be put off from doing stuff! we're not that fierce and unforgiving ;)

16:27:52 <libby> some of it is lack of time, some lack of motivation...

16:28:01 <libby> - wasn;t sure where headed

16:28:44 <DanC> the detail work includes stuff like looking at all the test data that has accumulated and seeing whether we still believe it.

16:28:47 <libby> I think one task is to pick the testcases that are good, i.e. that we can generate the schema from, is that right

16:28:49 <libby> ?

16:29:02 <DanC> same thing, yes, libby.

16:30:01 <libby> mortenf, anything you would like to try is very welcome. I made some notes about generating the schema somewhere

16:30:10 * DanC would be more than happy to see mortenf and libby work toward (a) test automation and (b) testing the other half of the round-trip, i.e. toIcal.py

16:30:35 <DanC> libby's notes: http://esw.w3.org/topic/RdfCalendarSchema

16:30:46 * tim-cape_ more or less catches up

16:30:58 * mortenf was just reading that - unsure of the bottom-up approach.

16:31:11 <libby> the utilityy of it?

16:31:13 <tim-cape_> Theer are various bits to test generation.

16:31:22 <mortenf> yeah, i guess.

16:31:35 <libby> why?

16:31:49 <libby> I Like ity a lot, because it's based on real data

16:32:18 <DanC> we could/should switch to maintaining the schema by hand and checking it against the data, rather than assuming the data are good, at this point.

16:32:25 * danbri nods, stuff based on vocab people use less likely to bitrot

16:32:37 <mortenf> hmm, it just seems a waste of time trying to do it automagically.

16:32:46 <libby> ?

16:33:01 <libby> I like the automated approach very much

16:33:12 <libby> *much* better than writing by hand.

16:33:16 <tim-cape_> 1) Actually spottting the need fro aparticular test case. That often happens here.

16:33:18 <DanC> sniffSchema took about 15 minutes to develop. I don't consider it a waste at all.

16:33:18 <tim-cape_> 2) Coding it up in RDF, as part of a manifest

16:33:19 <tim-cape_> 3) actually writing code to use it.

16:33:21 <tim-cape_> I suggest we make sure we don't lose (1) and if anyone can help with (2) that would be great.

16:33:22 <tim-cape_> And put (3) off till we have more energy.

16:33:33 * libby puzzled why danc would suggets writing by hand

16:33:42 <mortenf> anyway, i wouldn't mind trying to do some tround-trip testing

16:33:50 <DanC> I didn't say "writing by hand". I said maitaining by hand.

16:34:06 <DanC> i.e. you add a property to the schema with emacs.

16:34:29 <DanC> the existing schema is kinda goofy; I think it declares log:forSome as a property.

16:34:30 <libby> mortenf, that would be cool

16:35:05 <DanC> any idea when you'd have something you'd like some folks to look at, mortenf? are you thinking in the 1 or 2 week timeframe?

16:35:09 <libby> ah, ok, it does look quite complex too, compared with the simplest sort of schema

16:35:13 <tim-cape_> mertenf, how familiar are you with

16:35:14 <tim-cape_> the RDF core test suite maifests and the cwm regression.n3?

16:35:17 <mortenf> couple of weeks.

16:35:24 <tim-cape_> tim-cape_ is now known as timbl

16:35:32 <mortenf> timbl: very close to not at all.

16:35:44 <mortenf> (have yet to run cwm here)

16:35:53 <DanC> are you interested in learning how to run the cwm regression tests?

16:36:00 <mortenf> sure.

16:36:04 <timbl> You could look at eth test test data format without running the tests.

16:36:09 * danbri wonders if libby and DanC have swapped views w.r.t. automation vs hand coding

16:36:18 <libby> heh

16:36:19 <mortenf> yep.

16:36:36 <DanC> (the cwm docs on how to run the tests are pretty bad; at least: when I last looked for them I couldn't find them.)

16:36:40 <danbri> maybe automation was a way of bootsraping us into having a usable and reasonably ical friendly vocab, but time to let it grow now...

16:36:44 <timbl> I made a python harness $SWAP/test/retest/py which could be hackable for a fifferent package easily.

16:36:52 <DanC> bootstrapping: quite, danbri.

16:37:00 <mortenf> heh, ok, collecting pointers...

16:37:03 <danbri> ok that makes sense

16:37:13 <DanC> mortenf, on installing and running cwm: http://esw.w3.org/topic/CwmTips

16:37:49 * DanC encourages danbri to move that problem report to www-archive+n3bugs@w3.org,

16:37:56 <timbl> Morton, http://www.w3.org/2000/10/swap/test/retest.py reads a cwm test mainfest and/or rdf test manifest.

16:38:16 <danbri> mortenf, If you're link collecting, see http://www.w3.org/2003/03/rdfqr-tests/ for similar efforts on query testing; and to further links on RDFCore and OWL test docs and repositories.

16:38:42 <DanC> timbl, there's more to it than that. you have to be in the right directory, with the right PYTHONPATH

16:39:08 * danbri always has trouble running cwm regression tests

16:39:26 * DanC has fallen out of the habit of doing it routinely :-{

16:39:58 <mortenf> ok, i'll take an action on thus, hoping the next meeting will be at least a couple of weeks from now, more links welcome anytime.

16:40:02 * timbl did it yesterday

16:40:03 <mortenf> s/thus/this/

16:40:30 <libby> thanks mortenf! care to action yourself under G?

16:40:35 * DanC would like to see the contract that cwm committers are party to documented better.

16:41:01 * timbl danbri, i think you found that the problem you mentioned on the wicky page was with an old version of expat, didn't you? maybe you should update the wiki to that effect.

16:41:11 <DanC> G:ACTION mortenf: take a stab at regression tests for .ics->.rdf and/or .rdf->.ics

16:41:11 <dc_rdfig> Added comment G1.

16:41:54 * libby proposes we move on

16:42:01 * DanC presumes a due date of ~2 weeks, assuming we'll meet around that time

16:42:05 <timbl> For me a round-trip test is one which starts with a given file, converts it to sometyhing else, converst it back and compares. It is also interesting o have some files which have a defined output.

16:42:51 <DanC> I know more about how to compare .rdf files than how to compare .ics files

16:42:59 <mortenf> yep, it seems there are both in ../test/ ?

16:43:05 <DanC> so we could do round-tripping .rdf->.ics->.rdf

16:43:06 <libby> yes

16:43:10 * danbri updates http://esw.w3.org/topic/CwmTips

16:43:25 <libby> yes re in /test/

16:43:37 <timbl> (How do you compare ICS files?)

16:43:51 <mortenf> ok, i'll talk with libby re test cases details, bother anyone around re cwm... :)

16:43:54 <libby> I've never found anything that did that

16:43:55 <DanC> me? I just said I don't know, tim

16:44:06 <timbl> Oh.

16:44:11 <libby> cool mortenf

16:44:13 * timbl misread

16:44:20 <danbri> mortenf++

16:44:27 <DanC> go mortenf!

16:44:29 <mortenf> DanC, you mentioned existing ntriples diff'ers, anything besides cwm?

16:44:39 <DanC> jang wrote one, mortenf

16:44:53 <DanC> bwm, do you know where jang's n-triples comparison tool is?

16:45:10 <DanC> I think it's in c/c++.

16:45:12 * timbl cheers Sir Mortnef as he rides out through the castle gates on his gallant quest

16:45:18 <DanC> but it only groks n-triples, not RDF/XML.

16:45:21 <mortenf> heh.

16:45:31 <timbl> Jena has one

16:45:33 <libby> jena does it, in java

16:46:10 <libby> - in ntriples, n3 or rdf/xml

16:46:31 <DanC> I think using cwm should work. { <f1.rdf>.log:semantics log:includes <f2.rdf>.log:semantics. <f2.rdf>.log:semantics log:implies <f1.rdf>.log:semantics } => { <f1.rdf> :sameGraphAs <f2.rdf> }

16:46:48 <dajobe> jang's one is linked from the rdf test cases WD

16:46:52 <DanC> s/implies/includes/

16:46:53 <dajobe> it's in the test cases area/CVS

16:47:03 <mortenf> thanks, dajobe.

16:47:07 <DanC> ok, so that's 3 choices, mortenf

16:47:16 <timbl> Or you can use cwm as a canonicalizer. Put both files throug cwm --quiet and teh resulting files should diff .

16:47:40 <DanC> no, timbl, we just agreed graph comparison was the taget.

16:47:40 <mortenf> will try that as well.

16:47:49 <DanC> cwm's canonicalization is *not* the same thing.

16:48:07 <timbl> Why not?

16:48:41 <timbl> The most practical way of comparing two graphs is to make a cannoncial serialization of each.

16:48:44 <DanC> for one thing, cuz jena and jang's tool don't grok it.

16:48:53 <DanC> for another, cuz graphs can't be canonicalized.

16:49:12 <timbl> In practical cases they can.

16:49:19 * bwm jena's proud father is telling everyone she heard yesterday she got a first in her first degree :>

16:49:26 * danbri wonders what's left on the agenda we might cover

16:49:31 <DanC> yes, it's practically very close. but you agree it's different, right?

16:49:35 <danbri> bwm: congrats! :)

16:49:44 <danbri> (hi Danny)

16:49:44 <bwm> I'll pass them on.

16:49:47 <libby> wow, cool

16:49:55 <DanC> what's a first?

16:50:04 * bwm realises he is intruding on a meeting

16:50:06 <libby> we have a few things to cover in the agenda: ok to move on?

16:50:12 * timbl congratulations to hp:Jena.dc:creator^fam:father

16:50:13 <mortenf> k.

16:50:13 <bwm> DanC: a v. good thing.

16:50:24 <libby> DanC: the best deagree you can get, very difficult to get

16:50:24 * danja apologises for lateness and dog ate homework

16:50:29 <DanC> DanC has changed the topic to: RDF Calendar meeting 1600Z Weds. http://rdfig.xmlhack.com/

16:50:45 <DanC> congrats to jena!

16:50:54 <libby> re agenda, I have a question about prodid

16:50:55 <bwm> DanC: top marks - hard but not necessarily top of the class - helps if you want to do phd

16:51:18 <timbl> ~~ suma cum laude

16:51:24 <danbri> wanna wrap up this item? are all the actions noted?

16:51:33 <libby> I think so

16:51:43 * danbri rsi-ing, takes a breaklet

16:51:51 <libby> re handling prodid (calendar agendum C)

16:52:02 <timbl> ______________________________

16:52:05 * DanC wonders if libby accepted chair nomination. if so, all she has to do is say "we're now on item C, handling prodid" et viola, we're there

16:52:09 <libby> I finally did some of an action item from 26th feb

16:52:27 <libby> I didn;t realize. must be firmer.

16:52:59 <libby> re prodid, I have some notes here: http://www.ilrt.bris.ac.uk/discovery/2003/02/cal/readme.txt

16:53:20 <libby> I _think_ you were after somethign like that danc?

16:53:22 <danbri> (offtopic: whee, http://news.com.com/2009-1032-1020641.html?part=dtx&tag=ntop resource discovery becomes competitive again...)

16:53:25 <DanC> C:some [http://www.ilrt.bris.ac.uk/discovery/2003/02/cal/readme.txt|notes]

16:53:26 <dc_rdfig> Added comment C2.

16:54:03 <libby> context: DanC and I had a hack for xproperties - nonpstandard icalendar properties, where you passed the relevant namespace to the ical2rdf code

16:54:18 <DanC> the "prodid. new 2003-06-24" stuff looks about right; I'd like to see an example of the output

16:54:31 <libby> we wanted to use prodid somehow to generate consistenly a ns for the right x

16:54:44 * DanC wonders why libby doesn't check this into the shared workspace

16:54:45 <libby> there is one somewhere

16:55:23 <libby> here's one: http://www.ilrt.bris.ac.uk/discovery/2003/02/cal/mimedir-parser/test/generatedrdf/20030122mtg.rdf

16:55:31 <libby> - http://www.w3.org/2002/12/cal/prod/Apple_Comp_32a91

16:55:50 <DanC> yes! xmlns:x='http://www.w3.org/2002/12/cal/prod/Apple_Comp_32a91'

16:56:07 <libby> I will check in revised testcaes,, I wanted to get confirmation that the thing was ok

16:56:15 <libby> before creating schemas etc

16:56:32 <DanC> C:[http://www.ilrt.bris.ac.uk/discovery/2003/02/cal/mimedir-parser/test/generatedrdf/20030122mtg.rdf|example output] shows "-//Apple Computer\, Inc//iCal 1.0//EN" becomes xmlns:x='http://www.w3.org/2002/12/cal/prod/Apple_Comp_32a91'

16:56:32 <dc_rdfig> Added comment C3.

16:57:00 <libby> here's the evolution one: http://www.w3.org/2002/12/cal/prod/Ximian_NON_ccc8d

16:57:09 * danja now caught up - congrats bwm!

16:57:10 <DanC> this looks like you hit the target spot on, libby. well played.

16:57:12 <mortenf> hmm, could we have a separator on the end?

16:57:14 <libby> looks alittle strange

16:57:24 <libby> a hash, yes?

16:57:29 <libby> thanks danc

16:57:34 <mortenf> yep, that'd do.

16:57:45 <libby> yes, there shoudl be one

16:57:59 <DanC> oh yes, Apple_Comp_32a91#

16:58:09 <DanC> good eye, mortenf

16:58:22 <libby> so there's a prod dir; I coudl add schemas in files Ximian_NON_ccc8d and Apple_Comp_32a91

16:58:29 <DanC> yes, please do

16:58:33 <libby> ok, cool

16:58:54 <libby> I could probably hack your perl code danc if you like. might make a mess of it though

16:59:26 <DanC> i feel obliged to hack the perl code. but if you're in the mood... shall we flip a coin?

16:59:35 <libby> heh

16:59:44 <libby> I'll have a go

16:59:52 <libby> you can revert if I make too much of a mess

17:00:04 <danbri> datum, flip a coin

17:00:06 <timbl> So will we end up with revere engineering the schema automaticaly by sending ics files to some agent soemwhere?

17:00:38 <timbl> www-archive+ics-example@w3.org maybe

17:00:51 <DanC> C:libby's action from 26Feb is done; she has accepted another one to add support to ical2rdf.pl and add a couple schemas under cal/prod

17:00:53 <dc_rdfig> Added comment C4.

17:01:06 <libby> yay, i did an action :)

17:01:15 <danbri> hurrah

17:01:17 <DanC> yay indeed!

17:01:23 <mortenf> go libby!

17:01:27 <libby> heh

17:01:32 <timbl> :)

17:01:34 * DanC noodles over timbl's idea

17:01:37 <libby> we're a supportive channel :)

17:01:41 <mortenf> the sha1 is of the entire prodid, right?

17:01:51 <DanC> I think so, mortenf

17:02:17 <mortenf> methinks there could be some collision if using the 10-char version

17:02:28 * libby was wondering about that

17:02:35 <DanC> homework assignment: compute the probabilty of a collision.

17:02:40 <libby> cough

17:02:50 <DanC> it's low enough for my needs.

17:02:53 * mortenf looks out the window

17:03:05 <mortenf> but why not use the entire prodid?

17:03:09 <timbl> What is the drawback of just encodingthe whoel string?

17:03:23 <timbl> It is a ns name, so it will nromamly only occur once in a doc

17:03:30 <DanC> er... only that this design is coded up and that one isn't.

17:03:43 <timbl> indeed.

17:03:52 <libby> yeah mortenf, makes sense

17:03:54 <DanC> and fitting the whole thing into the posix portable name conventions might be hairy

17:04:30 <timbl> (what are they? 8.3?)

17:04:51 <libby> there aren;t many different prodids anyway....

17:04:52 <DanC> no, not 8.3... they're names that tar is guaranteed to handle portably

17:05:48 <timbl> In fact, you could leave the / in (convert // to /) but just zap non-posix-portable chars

17:06:02 <mortenf> i just thought that if the first part is "always" company name, any company with a long name will have trouble upgrading...

17:06:06 <timbl> I'm wondering whether round-tripping these things is importnat.

17:06:34 <timbl> We did njust discuss round-trip tests after all.

17:06:47 <libby> ah, that's a good point

17:06:58 <mortenf> but the real prodid is still a property?

17:06:59 <libby> although the full prodid is also in the rdf docuemnt

17:07:02 <libby> yep

17:07:07 <DanC> the prodid is still there in the .rdf file

17:07:17 <timbl> As?

17:07:24 <timbl> As a property of the calendar?

17:07:25 <DanC> <prodid>...</prodid>

17:07:42 * DanC checks sample data linked from C:

17:07:44 <libby> as a prob of the vcalendar yes

17:07:53 <DanC> yes

17:08:06 <libby> <prodid>-//Apple Computer\, Inc//iCal 1.0//EN</prodid>

17:08:22 <timbl> As someone who likes to merge and filter calendars, I don't see a lot of use for calendar-wide properties at all actually, unless you can stick them on the events.

17:08:23 <DanC> this is just a convention for how we're going to manage the 2002/12/cal/prod/ directory

17:08:56 <timbl> But you are using it as a namespace for x- properties?

17:09:03 <DanC> I guess the schemas in 2002/12/cal/prod/* need to contain the full prodid

17:09:18 <DanC> yes, using as a namespace for x- properties

17:09:19 <timbl> That would complete the loop for those who read schemas.

17:09:23 <libby> as p[art of the namespace?

17:10:00 <libby> i.e. the whole -//Apple Computer\, Inc//iCal 1.0//EN (with //, spaces, comma removed?)

17:10:07 <libby> plus the hash bit?

17:10:10 <nym> hi

17:10:18 <libby> hi nym

17:10:23 <nym> cal meeting?

17:10:33 <libby> yep, for 20 minutes more

17:10:35 <nym> libby: are you a bot? i get so confused

17:10:44 * mortenf smiles

17:10:47 <libby> yes I'm a bot

17:10:48 <DanC> http://www.w3.org/2002/12/cal/prod/Apple_Comp_32a91 should contain something like { <> :sampleCalendar [ cal:prodid "-//Apple Computer\, Inc//iCal 1.0//EN"] }

17:10:48 <dc_rdfig> I: http://www.w3.org/2002/12/cal/prod/Apple_Comp_32a91 from DanC

17:10:50 <libby> no not really

17:10:57 <libby> I can;t spell well enough to be a bot

17:10:59 <DanC> phpht

17:11:06 <DanC> how do I delete I?

17:11:07 <nym> libby: have you ever held a flag with the words 'foaf' on it?

17:11:16 <libby> I'm afraid so

17:11:17 <DanC> I:|DeleteMe

17:11:17 <dc_rdfig> Titled item I.

17:11:26 <mortenf> put in another uri...

17:11:27 <nym> libby: i used your pic in me blog

17:11:29 <libby> although actually it didn;t ahve foaf on it

17:11:32 <libby> I saw :)

17:11:34 <nym> libby: it's keen

17:11:41 <nym> oh cool

17:11:48 <timbl> I=

17:11:52 <libby> so, danc, the calendar should have the full prdid, but not the namespace?

17:11:56 <nym> so i'm sorry to intupupt the flow, what's on the agenda?

17:12:05 <DanC> "not the namespace"??

17:12:06 <libby> see http://rdfig.xmlhack.com nym

17:12:07 <DanC> I don't understand

17:12:34 <mortenf> both.

17:12:37 <libby> I wasn;t sure if we had decided that the namespace for xproperties was ok with 1st 10 characters

17:12:40 <libby> or not

17:12:48 <libby> or whether it needed all

17:13:01 <libby> of teh characters from the prodid

17:13:03 <mortenf> 1st 10 + sha1 of all: ns

17:13:20 <libby> ok, full sha1, gotya

17:13:29 <nym> oh, cal as in calendar

17:13:31 <mortenf> the ns *document* contains a "reference" to the real prodid

17:13:32 <DanC> we decided 26Feb that the namespace name for x properties would be finite length with a hash. We could revisit that decision, but I'd rather not.

17:13:33 <nym> not as in socal

17:13:37 <timbl> To round trip an event which has got an apple x-whatever: line in it one has to preserve the whole prodid in the namespace name. Actually this breaks anyway if there are x- from >1 company in events in the same calendar. :(

17:13:40 <nym> you guys :)

17:14:00 <libby> socal?

17:14:06 <nym> sothern california

17:14:11 <libby> oh

17:14:14 <libby> no

17:14:15 <libby> :)

17:14:20 <nym> i have an event RDF module that might be of interest though

17:14:32 <libby> wanna chump it nym?

17:14:33 <nym> i'm using it at Burning Man to allow people to write events

17:14:36 <mortenf> why is that, timbl?

17:14:55 * DanC notes the toIcal.py tool doesn't grok x- properties

17:15:05 <nym>http://igargoyle.com/rss/1.0/modules/event/

17:15:06 <dc_rdfig> J: http://igargoyle.com/rss/1.0/modules/event/ from nym

17:15:21 <nym> J: RDF 1.0 Modules: Event

17:15:21 <dc_rdfig> Added comment J1.

17:15:27 <nym> J:| RDF 1.0 Modules: Event

17:15:28 <dc_rdfig> Titled item J.

17:15:40 <timbl> Well, the way to represent apple:foo seems to be that one put "Prodid: decode("apple")" at eth top of the calendar and then use x-foo in the event. This assume sthere is only one proidid fro all the data.

17:15:48 <nym> J1: My event module that I'm going to use for Burning Man. Comments appreciated

17:15:49 <dc_rdfig> Replaced comment J1.

17:15:55 <libby> so just to be clear, the hash should be the length that comes out of the algorithm (e.g. 30 chars) in teh namespace?

17:16:14 <libby> thanks nym

17:16:17 <timbl> Test case: Take one evolution 1-event calednar and one apple:ical 1-event character, merge them.

17:16:19 <nym> no problem

17:16:25 <libby> I'll have a look at that in a few monutes when done here

17:16:30 <nym> i figure it does work into the discussion somehow

17:16:39 <libby> definitely

17:16:44 <nym> i haven't looked at apple's schema

17:16:58 <DanC> if you take a calendar from apple ical and one from evolution, merge them together, it's not at all clear what prodid you should use for the result when you go back to .ics format. But the hashed namespace names will still work for the x-properties.

17:17:04 <mortenf> ah yeah, i see that, re multiple prodids, but not re "preserving in ns name"?

17:17:25 <danja> any background material on x- for us Macless rabble?

17:17:45 <mortenf> libby, methinks only first 5.

17:17:46 <nym> J: the event module is simple, it includes statedate, enddate, organizer, type (of event), and geo:point for locational purposes

17:17:48 <dc_rdfig> Added comment J2.

17:17:52 <mortenf> (of sha1)

17:18:01 <DanC> danja, can you see agendum C? i.e. http://rdfig.xmlhack.com/2003/06/25/2003-06-25.html#1056556072.203619 ? it has links to background

17:18:07 <nym> anyone seen danbri / anselm btw?

17:18:19 <libby> danja, icalendar allows you to create additional properties for use with icalendar. most sstems so this including icalendar and evolution

17:18:26 <mortenf> danbri is on a break...

17:18:30 <libby> danbri's about somewhere

17:18:34 <libby> thanks mortenf

17:18:43 <danbri> hi nym; we're in tail end of rdf-cal meeting, but i can multi-task...

17:18:49 <danbri> (in #foaf maybe?)

17:18:52 * libby notes taht we have 10 inutes or so less

17:18:57 <libby> s/less/left/

17:19:07 <libby> and a few more items

17:19:31 <libby> I'd be interested in talking about representing attendance (calendar agendum D) if danc is willing

17:19:37 <DanC> BLURB: next meeting (agendum K)

17:19:37 <dc_rdfig> K: next meeting (agendum K) from DanC

17:19:42 <mortenf> beet me to it, libby :)

17:19:48 <mortenf> s/beet/beat/ :)

17:19:56 <danja> DanC, thanks, but doesn't help much with origins of x-properties

17:20:09 <libby> we shoudl also decide on next meet (2 weeks time?)

17:20:28 <DanC> PROPOSED: to meet again in 2 weeks, 9 July. 1600Z for <=90min. http://www.timeanddate.com/worldclock/fixedtime.html?day=9&month=7&year=2003&hour=16&min=0&sec=0&p1=0

17:20:29 <dc_rdfig> Label PROPOSED not found.

17:20:30 <libby> and there's also E, and F

17:20:51 <libby> I thik that's ok for me

17:20:55 <mortenf> fine with me

17:21:10 <libby> seconded, DanC

17:21:19 <libby> any objections?

17:21:43 <timbl> J:"Use semantic augmentation if you desire to give i.e. a phone number to the organizer." semantic augmentation?

17:21:43 <dc_rdfig> Added comment J3.

17:21:53 * timbl regrests for 9 july

17:21:57 <DanC> L: PROPOSED: to meet again in 2 weeks, [http://www.timeanddate.com/worldclock/fixedtime.html?day=9&month=7&year=2003&hour=16&min=0&sec=0&p1=0|Wed 9 July at 1600Z] for <=90min.

17:21:57 <dc_rdfig> Label L not found.

17:22:06 <DanC> K: PROPOSED: to meet again in 2 weeks, [http://www.timeanddate.com/worldclock/fixedtime.html?day=9&month=7&year=2003&hour=16&min=0&sec=0&p1=0|Wed 9 July at 1600Z] for <=90min.

17:22:06 <dc_rdfig> Added comment K1.

17:22:36 <libby> J:a bit underspecified, but quite well used

17:22:36 <dc_rdfig> Added comment J4.

17:22:53 <DanC> as i recall "semantic augmentation" means confusing things vs. their names

17:22:56 <libby> ok, danc, care to talk us through D?

17:23:20 <libby> representing attendance (calendar agendum D)

17:23:21 * mortenf seconds libby's proposal

17:23:21 <DanC> J:cf [http://esw.w3.org/topic/ThingsVersusTheirNames|ThingsVersusTheirNames]

17:23:21 <dc_rdfig> Added comment J5.

17:23:52 * DanC tries to remember why he found attendance interesting

17:24:01 <libby> heh

17:24:07 <mortenf> zool's usecase?

17:24:11 <timbl> because you can't attribute an action to so. not there?

17:24:22 <timbl> Wg rules of good standing?

17:24:32 * libby also remembers adenda item H, wonders if DanC can talk about that too, maybe after meet?

17:24:32 <timbl> Because access to minutes should be allowed to all atrtendees?

17:24:52 <libby> also cool for doing pathcross maybe?

17:25:11 <DanC> ah now I remember: cuz I want to be able to run queries of the form "which of my colleagues plans to attend something that I plan to attend?" e.g. extreme markup in Augus

17:25:39 <timbl> Also, who can this photo be of, which I took during this event?

17:25:45 <danja> DanC - thanks, found what I was after in rfc2445

17:26:02 * DanC would rather do H

17:26:08 <danbri> Also "Who from company X attends meetings of type Y" could be useful (within X's intranet, or wider analysis)

17:26:13 <mortenf> danja, could you chump a seealso for that rfc?

17:26:36 <libby> perhaops we could folow up D using email, given that DanC has started the ball rolling

17:26:36 <danja> ok

17:26:58 <mortenf> that's fine.

17:27:02 <DanC> for D, I was thinking about { { ?WHO :location ?WHERE } :holdsDuring ?WHEN } vs. { ?WHO subAbstrac ?WHOWHEN. ?WHOWHEN :location ?WHERE }.

17:27:29 <DanC> and I was going to argue that :holdsDuring is probably less workable.

17:27:54 <libby> I think this is something that could profitably be moved to email

17:27:59 <libby> let's talk about H

17:27:59 <DanC> i.e. I don't want to rely on nested formulas for this.

17:28:06 <DanC> ok, on to H...

17:28:34 <DanC> H:cf [http://esw.w3.org/topic/EmailVocabulary|EmailVocabulary]

17:28:34 <dc_rdfig> Added comment H1.

17:28:56 <DanC> I finally published a schema http://www.w3.org/2000/10/swap/pim/email

17:29:03 <DanC> (using Dom's cool XSLT hack)

17:29:07 * timbl for D, yes, I agree .. don't use nested f's do [ start 10; end 11; at London; there me]

17:29:47 <DanC> the schema is supported by 2 tools:

17:30:01 <libby> ooh, like the xslt

17:30:15 <DanC> 1st http://www.w3.org/2000/04/maillog2rdf/aboutMsg.py which takes rfc822 formatted email and spits out some of the header info in RDF

17:30:43 <danja> morten - I can't remember how to do seealsos, chump item is C: iCalendar rfc at http://ietf.org/rfc/rfc2445.txt

17:31:10 <DanC> the other is http://www.w3.org/2000/04/maillog2rdf/mid_proxy.py . the name is kinda outdated. what it does is proxy imap queries and give you the results in RDF. mid_proxy.py still has @@ in the namespace name, though. hasn't been updated

17:31:48 <mortenf> C:seeAlso [http://ietf.org/rfc/rfc2445.txt|RFC 2445]

17:31:48 <dc_rdfig> Added comment C5.

17:31:55 <libby> that's neat

17:32:12 <libby> so you are using it for webont danc?

17:32:21 <DanC> i.e. you run mid_proxy.py and it logs into your imap server and starts servicing HTTP requests. it supports a form. you put imap queries in the form and the results come back in .html or .rdf

17:32:25 <mortenf> thanks danja

17:32:41 <timbl> Is there a running example server if the IMPA proxy onto eg www-archive or www-rdfig?

17:32:42 * libby notes that we're over time, proposes we carry on for 10-15 minutes

17:32:58 * mortenf is still here

17:33:07 <DanC> to build the email schema, I first ran aboutMsg.py on a bunch of data, used sniffSchema, and then manually cleaned up the results.

17:33:26 * danja can walk dog later

17:33:26 <timbl> It would be cool if the proxycraeted a dbview-like set of virtual RDF documents.

17:33:32 * timbl ok to continue

17:33:36 <DanC> er... the mid_prox.py gives access to all my mail when I run it.

17:34:05 <DanC> yes, mid_proxy.py should interoperate with dbview and joseki and the other remote query doodads.

17:34:15 <timbl> re err.... We would need an imap server running on the public archives.

17:34:21 <DanC> but so far it only groks IMAP query syntax

17:34:34 <timbl> ok

17:35:23 <DanC> I note we already have our public archives exported via HTTP, and I wished for RDF output, and EricP started tinkering.

17:36:10 <DanC> oh... one more point about the email schema and aboutMsg.py... there's a question of: how do I tell if a message was *not* sent to blort@example.org?

17:36:40 <DanC> so I added two properties, email:senders and email:recipients, which are strings that contain all the relevant email addresses, one per line

17:37:02 <DanC> so you can do { ?MSG email:recipients [ str:notContains "blort@example.org"] }

17:37:08 <danbri> (maybe i mised this)... do you consider To: to be an ordered list?

17:37:32 <DanC> no...

17:37:48 <danbri> cool.

17:38:05 <DanC> H:[http://www.w3.org/2001/sw/WebOnt/webont-lc-comments.rdf|webont-lcs-comments.rdf] is built using aboutMsg.py

17:38:05 <dc_rdfig> Added comment H2.

17:38:17 <danbri> I think some people consider the order to be significant, and it may carry subtle social hints but ok to abstract away from that detail

17:38:27 <DanC> [[[

17:38:28 <DanC> <email:to rdf:parseType="Resource">

17:38:28 <DanC> <email:phrase>Misha Wolf</email:phrase>

17:38:28 <DanC> <email:mbox rdf:resource="mailto:Misha.Wolf@reuters.com"/>

17:38:28 <DanC> </email:to>

17:38:28 <DanC> -

17:38:30 <DanC> <email:cc rdf:parseType="Resource">

17:38:32 <DanC> <email:phrase>Www International</email:phrase>

17:38:34 <DanC> <email:mbox rdf:resource="mailto:www-international@w3.org"/>

17:38:36 <DanC> </email:cc>

17:38:38 <DanC> -

17:38:40 <DanC> <email:cc rdf:parseType="Resource">

17:38:42 <DanC> <email:mbox rdf:resource="mailto:public-webont-comments@w3.org"/>

17:38:46 <DanC> </email:cc>

17:38:48 <DanC> ]]]

17:38:57 <DanC> RFC822 specifies order is insignificant. SMTP servers alphebetize, I believe.

17:39:21 <DanC> I made up the terms email:phrase and email:mbox, though the concepts are clearly in RFC822

17:39:42 <danbri> ok

17:39:58 <DanC> I was thinking about foaf:name and email:phrase, ala...

17:40:33 <DanC> { ?X a :PublicForum. ?MSG email:to ?X; email:from ?WHO. ?WHO email:phrase ?TXT } => { ?WHO foaf:name ?TXT }

17:40:57 <DanC> i.e. if you call yourself "Dan Brickley" in mail to a public forum, that's your foaf:name, or one of them, at least.

17:41:23 <danbri> seems a reasonable heuristic for many purposes, though I wouldn't pay much for datasets built that way

17:41:38 <DanC> you wouldn't accept that as an axiom?

17:42:14 <DanC> i see foaf as a community of internet citizens; the way internet citizens establish their names is on the From lines of mail to public fora

17:42:17 <danbri> I guess I would, but knowing that some people goof w/ their names there, so it might be more of a foaf:nick

17:42:20 <danbri> slidingscale...

17:42:30 * danbri nods, agrees

17:42:49 <DanC> if people goof, then the input data is false; it doesn't mean the axiom is wrong

17:43:07 <JibberJim> I think foaf:name == email from lines. seems entirely reasonable, there'll one day be more specific naming vocabs,

17:43:42 <DanC> also, email:mbox is pretty close to foaf:mbox, though I haven't figured out an formalization of the relationship

17:43:59 <danbri> in http://rdfweb.org/pipermail/rdfweb-dev/2003-June/011290.html Masahide Kanzaki writes, [[

17:44:00 <danbri> I would suggest that 'foaf:name or foaf:nick' be the minimal along with one

17:44:00 <danbri> inverseFunctionalProperty, because some people may not want to expose their

17:44:00 <danbri> real name on the web. Actually, the majority of the emerging japanese foaf

17:44:00 <danbri> use 'foaf:nick' but not necessarily 'foaf:name'.

17:44:02 <danbri> ]]

17:44:09 <danbri> that was interesting to learn!

17:44:10 * DanC is sorta rambling; pauses to see where others want to take the meeting

17:44:16 <timbl> goofing is part of the protol would mean the axiom was wrong.

17:44:20 <timbl> maybe foaf:aka

17:44:43 <DanC> did you mean "goof" as in "make mistake" or "goof" as in "goof around, play with"?

17:45:05 * danja interrupted by dogling - back soon...

17:45:12 * danbri has to head off...

17:45:29 <danbri> "play around with", mostly. Wear hats and roles, maybe. But also bad data.

17:45:39 <timbl> If goofing aorund by making up a silly nick is part of the accepted protcol, then you can't call ethh axiom correct and them wrojng.

17:46:00 <DanC> ah... yes, timbl's right.

17:46:23 <DanC> I thought you meant "goof" as in "make mistakes"

17:46:47 * DanC proposes to postpone the remaining items and adjourn

17:47:20 * timbl was referring to: danbri: "I guess I would, but knowing that some people goof w/ their names there, so it might be more of a foaf:nick"

17:47:39 <DanC> K:postponed: repeating events(E)

17:47:39 <dc_rdfig> Added comment K2.

17:47:55 * libby seconds danc's suggestion

17:48:02 <DanC> what else didn't we cover/

17:48:04 <mortenf> sounds great, needs food...

17:48:15 <DanC> K:postponed: InterpretationProperties, esp as applied to timezones (agenda item F)

17:48:15 <dc_rdfig> Added comment K3.

17:48:39 <DanC> D:pls continue in email

17:48:39 <dc_rdfig> Added comment D4.

17:48:39 <mortenf> H could be considered closed or so?

17:49:03 <DanC> yes, we're done with H to my satisfaction, thanks.

17:49:12 <libby> I would be interested din chatting more about H, when moree time

17:49:25 <libby> interested in things being used for

17:49:27 <DanC> E:not discussed.

17:49:27 <dc_rdfig> Added comment E2.

17:49:42 <libby> F?

17:49:49 <mortenf> pp'ed

17:49:57 <DanC> crap; I'm late for something. ciao, all!

17:50:03 <libby> thanks DanC, bye

17:50:04 <mortenf> thanks!

17:50:21 <libby> F:continued

17:50:22 <dc_rdfig> Added comment F2.

17:50:38 <mortenf> how can someone be late from a calendar meeting? :)

17:50:55 <libby> heh

17:51:06 <mortenf> ok, so that's it?

17:51:12 <libby> yep, we're adjourned

17:51:19 <mortenf> perhaps your action should be in capital letters, libby? ;)

17:51:20 <libby> thanks mortenf

17:51:25 <libby> action?

17:51:30 <mortenf> sure, and thank you

17:51:31 <timbl> bye everyone

17:51:44 <libby> bye

17:51:45 <mortenf> C4

17:51:56 <mortenf> list C

17:52:01 <mortenf> dc_rdfig, list C

17:52:01 <dc_rdfig> Not understood: list C

17:52:05 <dajobe> just do dc_rdfig:C

17:52:08 <mortenf> dc_rdfig:C

17:52:09 <dc_rdfig> Not understood: C

17:52:14 <dajobe> oops, or C:

17:52:15 <mortenf> dc_rdfig, C

17:52:15 <dc_rdfig> Not understood: C

17:52:21 <mortenf> dc_rdfig: C:

17:52:21 <dc_rdfig> blurb

17:52:22 <dc_rdfig> handling prodid (calendar agendum C)

17:52:23 <dc_rdfig> (1:libby) see [initial weblong action item|http://rdfig.xmlhack.com/2003/02/26/2003-02-26.html#1046279854.884486]

17:52:24 <dc_rdfig> (2:DanC) some [http://www.ilrt.bris.ac.uk/discovery/2003/02/cal/readme.txt|notes]

17:52:26 <dc_rdfig> (3:DanC) [http://www.ilrt.bris.ac.uk/discovery/2003/02/cal/mimedir-parser/test/generatedrdf/20030122mtg.rdf|example output] shows "-//Apple Computer\, Inc//iCal 1.0//EN" becomes xmlns:x='http://www.w3.org/2002/12/cal/prod/Apple_Comp_32a91'

17:52:26 <dc_rdfig> (4:DanC) libby's action from 26Feb is done; she has accepted another one to add support to ical2rdf.pl and add a couple schemas under cal/prod

17:52:28 <dc_rdfig> (5:mortenf) seeAlso [http://ietf.org/rfc/rfc2445.txt|RFC 2445]

17:52:31 <mortenf> thanks, dajobe.

17:53:17 <libby> hm, swapped an old action for a new one, but never mind :)

17:53:23 <mortenf> :)

17:53:26 <mortenf> biab...

17:53:32 <libby> bye

18:17:33 <sethl> timbl: if the outer most formula in cwm is open (which I'm assuming because I'm able to add triples to it), how can a inner formula, such as {:seth :has ?X} => {:seth :likes :?x}, ever return? How can it, theoretically, ever know it's gotten all the results for the query, and therefore, ever do the log:implies?

18:18:03 <sethl> in an open world, the query :seth :has ?x can't be proven to return all the matches, can it?

18:18:15 <sethl> if so, does that mean the world is closed for just a moment?

18:22:35 <sethl> I'm having a hard time writing a solution to this use case: I want to buy a Jimmy Buffet CD for Bob, one he doesn't have, and it's under $10 dollars. Steve and Justin know what Bob has, so ask them. Ask Amazon then for prices.

18:23:41 <sethl> it's the "CDs that Bob /doesn't/ own" that I'm having difficulty with. I'm worked out a owl:complementOf rules, but it's the act of computing the rdf:List of CDs Bob owns that gets me. I'm not sure if I can compute that List (or merge two Lists).

18:52:24 * bwm suggests someone change the topic back

18:52:27 <libby> bwm, that's really great about jena

18:52:39 <libby> what's she going to do next?

18:52:43 <bwm> thanks libby - I thought so too

18:52:50 <bwm> phd, maybe after a year out

18:53:01 <libby> what topic?

18:53:32 <DanC> J:the geo:Point example doesn't parse

18:53:32 <dc_rdfig> Added comment J6.

18:54:05 <DanC> DanC has changed the topic to: Semantic Web hack-n-chat. http://rdfig.xmlhack.com/ http://esw.w3.org/topic/FrontPage

18:54:11 <libby> I mean, what topic would jena's phd be on?

18:54:21 <bwm> she has some ideas she wants to try out - seems to know what area she wants to work in, but doesn't have specific topic

18:54:40 <bwm> her degree is in psychology

18:55:04 <libby> ->academia eventually?

18:55:37 * DanC wonders if there are tests for if vs. iff in rdfcore tests

19:02:42 <sethl> danc: how do you express iff in rdf?

19:03:06 <DanC> in english in the spec

19:03:23 <sethl> ah, right

19:03:23 * DanC tries to trace decision record re iff...

19:04:51 <DanC> BLURB:on RDFS semantics of subClassOf, range, etc.

19:04:52 <dc_rdfig> L: on RDFS semantics of subClassOf, range, etc. from DanC

19:05:14 <DanC> L:cf [http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Oct/0381.html|request for feedback on domain and range semantics] 25 Oct from Hayes

19:05:14 <dc_rdfig> Added comment L1.

19:16:04 <DanC> L:ugh; DanC was party to keeping this off the issues list [http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Oct/0386.html|25Oct]

19:16:05 <dc_rdfig> Added comment L2.

19:22:14 <DanC> L:jang notes if vs. iff in [http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jan/0003.html|his review of 6Jan]

19:22:14 <dc_rdfig> Added comment L3.

20:58:27 <mdupont> hello all you great mind in rdfig

20:58:39 <mdupont> please awake for a nice and oderly dicussion

20:58:53 <mdupont> of what it means to have an identifer in a progra

20:58:54 <mdupont> m

20:59:28 <mdupont> and how you can take reference in rdf to a idea of an identifer that can be resolved into a value in a binding in a file.

20:59:59 <mdupont> i mean, lets take the sourcecode of a program

21:00:09 <mdupont> what If i extract all the identifiers out of it

21:00:13 <mdupont> into an rdf file

21:00:18 <mdupont> and all the things that use it

21:00:25 <mdupont> and all the things that are referenced by it

21:00:40 <mdupont> does that not give you a map of the relationships in the program?

21:00:59 <mdupont> can rdf be used to describe such systems as statements,

21:01:45 <danbri>http://rss-jp.net/

21:01:45 <dc_rdfig> M: http://rss-jp.net/ from danbri

21:01:52 <danbri> M:|rss-jp.net

21:01:52 <dc_rdfig> Titled item M.

21:02:27 <danbri> M:RSS-related links in Japan.

21:02:27 <dc_rdfig> Added comment M1.

21:02:41 <danbri> mdupont, hi, sorry no time for tech chat now :(

21:05:50 <mdupont> hey danbri see you another time :)

21:20:37 <deluxe_> mdupont: gruezi ;-)

21:27:05 <mdupont> deluxe_: und sch?nen Tag noch!

21:28:47 <deluxe_> mdupont: haha 23:28

21:28:58 <mdupont> oho

21:32:53 <mdupont>http://wiki.dotgnu.org/DGEEChatReport?action=show

21:32:53 <dc_rdfig> N: http://wiki.dotgnu.org/DGEEChatReport?action=show from mdupont

21:33:14 <mdupont> N:this is a report over all the statements about dgee in dotgnu

21:33:14 <dc_rdfig> Added comment N1.

21:33:47 <mdupont>http://gnufans.net/intrspctr.pl?ScanEmails.Pl

21:33:47 <dc_rdfig> O: http://gnufans.net/intrspctr.pl?ScanEmails.Pl from mdupont

21:33:56 <mdupont> O:this scan.pl program groups statements from a subject or about a subject into hash

21:33:56 <dc_rdfig> Added comment O1.

21:34:34 <mdupont> this hash is used to determine what the speaker is, statements are grouped by speaker, sorted alphabetically

21:34:57 <mdupont> it would be good to have the time and the ability to determine the version of sofware that they are talking about

21:35:05 <mdupont> most of the bugs should have been solved by then

21:35:41 <mdupont> O|Simple Chatlog scannerl perscript

21:35:46 <mdupont> O:|Simple Chatlog scannerl perscript

21:35:46 <dc_rdfig> Titled item O.

21:35:58 <mdupont> O:|Simple Chatlog scanner perscript

21:35:58 <dc_rdfig> Titled item O.

21:51:43 <DanC> hey! somebody actually copied www-archive@w3.org when asking me an HTML question, the way I requested on my homepage. wonders never cease!

21:52:06 <DanC> ew; it's about frames

21:52:19 <mortenf> just say no.

21:52:25 <edd> heh, can't win em all

21:53:12 <DanC> . http://lists.w3.org/Archives/Public/www-archive/2003Jun/0055.html

21:53:40 <DanC> attached files (yeah!). zipped. (boo!)

21:54:09 <mortenf> do you get a lot of those?

21:54:13 <mortenf> q's?

21:54:28 <mortenf> (there was two more, anyway...)

21:54:57 <DanC> I get enough such questions that I bothered to write, on my homepage... "I cannot justify the time that it takes to answer questions privately; if you must send to me personally, copy www-archive@w3.org as a means of giving me license to share your question and the answer with the rest of the world vi the archive."

21:55:36 <mortenf> nice :/

21:57:25 <DanC> -- http://www.w3.org/People/Connolly/#contact

22:05:15 <danbri> and they went thru archive-approval too? (or i guess were already a www-* poster...)

22:05:58 <DanC> they did whatever was required.

22:06:29 * DanC is really depressed that all mail has to be assumed spam until proven otherwise

22:07:19 * danbri nods, grovels around in unknown-senders every few hours to keep on top of the spam/new mix...

22:10:03 <DanC> I'd really like to use evolution's filters to color my mail, but I'm not willing to record incredibly valuable knowledge about my world in some .evolution.rc file that (a) isn't usable by other email clients, and (b) might get blown away next time I upgrade if I'm not careful

22:10:18 <danbri> generate the .rc from rdf?

22:10:22 <DanC> I really want to write down my knowledge about my world in RDF and convert it to .procmailrc and such

22:10:27 <DanC> yes, exactly.

22:11:27 <DanC> I was really pleased when i was able to use aboutMsg.py and cwm rules and such to populate my evolution contacts list in Feb 2002

22:11:37 <danbri> I sortof do that with whitelists; i have a big dbm file derrived from the lists.w3.org accept list that I hashed (not publically available tho) plus my own whitelist; .procmail just consults that, alongside spamassassin

22:12:16 <DanC> yeah, I put "export sha1 whitelist" on my PDA todo list a LOONG time ago

22:12:17 <mortenf> hmm, could that accept list make it out as mbox_sha1sums?

22:12:39 <DanC> but now that so much email is forged, whitelists aren't all that reliable

22:12:49 <danbri> there was an objection that a list of common firstnames and family names, plus a list of internet domains, plus a bunch of cpu time, would allow the list to be reverse engineered...

22:13:02 <mortenf> bah.

22:13:21 * danbri thinks we need to ramp up pgp deployment, whitelists will buy us a year or two to get that done

22:13:46 <DanC> I put some energy into pgp in Budapest, but I'm still not very hopeful about it.

22:14:02 <danbri> I was looking at what it'd take to do client-side pgp checking in perl (eg. inside AmphetaDesk RSS client); but the perl mods have huge dependencies. GPG might be easier.

22:14:03 <DanC> I'm pretty much of a rocket-scientist, and I find it challenging to manage a pgp key.

22:14:15 <danbri> Yeah, I get confused easily...

22:14:35 <danbri> Maybe if we were taught it at school as kids?

22:14:41 <DanC> I like the idea of putting mail sending servers in DNS to fight forgery.

22:14:50 <edd> hmm, did i sign your key DanC? pretty sure i did it there and then.

22:15:01 <DanC> yeah, edd, I think you did.

22:15:23 <DanC> I haven't gone to the trouble of teaching evolution (both installations that I use) about my key.

22:15:33 <DanC> (not to mention the other email client I use for personal use...)

22:16:00 <DanC> can't I just tell my email client about my homepage and have it figure the rest out? sigh.

22:16:13 <DanC> (yes, I'd have to vouch for the key id)

22:16:47 <DanC> I guess with mozilla, I could use one of those roving profiles or whatever.

22:16:54 <danbri> that was bitsko was writing about recently, http://rdfweb.org/topic/FoafIdentityAssurance and prompted my investigation of PGP-on-desktop etc

22:17:07 <timbl> I like the idea of putting mail forgers in jail.

22:17:35 <edd> some item on radio 4 this evening about a guy MS are suing in error.

22:17:50 <edd> seems he bought a domain previously used for spamming.

22:18:01 <edd> got home to wife and kids to find a writ from BigMonsterCo.

22:18:14 <edd> i guess going to the media about the only way for joe public to fight such things

22:18:19 <DanC> putting folks in jail is pretty extreme... do we put check forgers in jail? not very often, I don't think. It's hardly fair to leave a pile of money out there in the open and then put folks in jail for taking it.

22:18:25 <dajobe> story: http://news.bbc.co.uk/1/hi/business/3020566.stm

22:18:34 <DanC> we should at least put some chicken-wire around the money

22:19:49 * DanC imagines title searches on domains...

22:20:05 <edd> DanC: i *believe* that if you simply turn PGP signing on, Evo will figure which key to use from your outgoing email address. keyid is only needed if you wish to disambiguate.

22:20:07 <timbl> I think the worst thing is the bouce messages from other systems saying that "my" mail had a virus or couldn't be delivered as npjne of johna, johna1, johnabrec, johnacru, johnbb@aol.com were valid emails. They are real messages by a machinbe trying to do right, erroneously sent to me.

22:20:24 <timbl> s/npjne/none

22:20:58 <DanC> email forgery is evil, no doubt. so let's put a little bit of technical difficulty in the way.

22:20:59 <timbl> chickenwire: yes.

22:22:03 <DanC> chickenwire proposal: RMX. http://www.ietf.org/internet-drafts/draft-danisch-dns-rr-smtp-02.txt

22:22:23 <DanC> from http://www.irtf.org/asrg/ Anti-Spam Research Group (ASRG)

22:22:31 <deltab> also SPF

22:22:36 <DanC> SPF?

22:23:23 <deltab> spf.pobox.com

22:24:26 <DanC> edd, the gpg preference dialog asks for a key id

22:25:14 <edd> DanC: i know, but i'm saying that you can leave it empty and it will still work

22:25:29 <edd> this is not documented or obvious, i'll grant.

22:25:45 * DanC wonders how to tell if it's working short of sending mail

22:26:19 <DanC> I don't even know where my key is

22:26:34 <edd> well, it should be in your keyring...

22:26:37 <DanC> i.e. it's not on this machine, I don't think

22:26:41 <DanC> where's my keyring?

22:26:41 <edd> ah, i see.

22:26:45 <edd> ~/.gpg

22:26:54 <edd> .gnupg

22:26:57 <DanC> ls: /home/connolly/.gpg: No such file or directory

22:27:10 <DanC> aha... I have .gnupg

22:27:14 <timbl> ls -lrta ~/.*

22:27:24 <DanC> what's the "wtf is my key?" command?

22:27:36 <edd> gpg --list-key connolly

22:27:38 <edd> does it for me

22:27:52 <DanC> gpg: error reading key: public key not found

22:28:09 * DanC thinks it's on my flash disk...

22:29:33 * DanC wishes linux would get with the '80s and automount the friggin disk

22:30:17 <DanC> ew... evil...

22:30:19 <DanC> Jun 25 17:29:53 dirk kernel: unable to read partition table

22:30:19 <DanC> Jun 25 17:29:53 dirk kernel: Device not ready. Make sure there is a disc in the drive.

22:30:19 <DanC> Jun 25 17:29:53 dirk kernel: sdb : READ CAPACITY failed.

22:30:22 <edd> there's some program called magicdev which does such things

22:32:08 <DanC> Jun 25 17:31:33 dirk /etc/hotplug/usb.agent: ... no modules for USB product ea0/6803/100

22:32:11 <DanC> this worked before.

22:32:38 * danbri gets that linux feeling

22:33:08 * DanC is just glad he's not waiting on hold for somebody to say "have you tried turning the machine off and on?"

22:34:09 <mortenf> nah, you're not on hold here...

22:37:02 * timbl wonders about http://www.sente.ch/software/GPGMail/English.lproj/GPGMail.html#Features MacMail GPG plugin

22:37:15 <DanC> there's a pretty reliable linux support algorithm: look at /var/log/messages, stick the most relevant thing into google, follow your nose...

22:38:42 <DanC> # modprobe usb-storage

22:38:42 <DanC> modprobe: Can't locate module usb-storage

22:38:51 <DanC> hmm... my usb-storage module seems to be in the wrong place.

22:39:48 <deltab> try adding .o

22:39:51 <mdupont> Q:This is interesting for doing documentation of the terms of the computing community

22:39:51 <dc_rdfig> Added comment Q1.

22:39:59 <DanC> hmm.. the usb-storage module is loaded...

22:41:36 <edd> norm++

22:41:47 <edd> i'm amazed, we might get some real rdf people to an xml conference!

22:41:54 <edd> could this be the start of a dialog :)

22:41:55 * DanC finds "Using a USB Compact Flash Reader in Linux" http://www.cs.sfu.ca/~ggbaker/personal/cf-linux

22:42:03 <DanC> yeah, norm++

22:43:43 * DanC discovers sg3-utils package of SCSI tools...

22:45:39 <DanC> sigh... it got assigned to /dev/sdc this time, while previously it was /dev/sdb

22:46:16 <edd> DanC: yeah, i get a different scsi device from my reader dependent on type of card, sda for memory stick, sdb for CF, etc. totally sucky

22:47:24 <DanC> does magicdevice handle that?

22:47:37 <edd> dunno, i don't use it :)

22:47:52 * DanC wonders if this problem is soluble

22:48:13 <DanC> the mac had a good idea: the disks themselves have names. when you put it in, the name of the disk shows up on the desktop

22:48:22 <edd> it's not clear to me that it is. a while back i wanted to write a script to read the CF serial number and then always mount it at a certain mountpoint

22:48:43 <edd> under linux hotplug currently, it didn't seem that that was trivial with some shell script, maybe not even possible without kernel mods.

22:49:05 <edd> like you, i was playing with keeping my GPG key on a USB memory device

22:49:54 * DanC doesn't see magicdevice among debian packages

22:50:37 <DanC> ah... http://www.ccs.neu.edu/home/bchafy/magicdev.html

22:51:13 <AaronSw> A:No, you're supposed to complain we're reinventing RSS.

22:51:13 <dc_rdfig> Added comment A6.

22:51:23 <edd> DanC: see http://www.apt-get.org/search.php?query=magicdev&submit=&arch%5B%5D=i386&arch%5B%5D=all

22:53:30 * DanC finds magicdev kinda scary

22:53:56 <DanC> autofs - A kernel-based automounter for Linux

22:54:03 * DanC wonders if that's any good

22:54:23 <edd> I use that for NFS mounts.

22:55:20 <DanC> holy cow... google is indexing https stuff

22:55:48 <DanC> "HOWTO Configure Linux to automount your USB digital camera or card reader" http://www.buberel.org/linux/usb-automounter.php

22:56:10 <JibberJim> such as what DanC? (https)

22:56:37 <DanC> such as https://listman.redhat.com/pipermail/valhalla-list/2002-July/008706.html <- http://www.google.com/search?q=automount%20usb

22:57:27 <DanC> chuckle... "Blinky lights should blink for a few seconds."

22:57:32 <edd> DanC: handy, but doesn't do what i wanted, which was matching CF serial numbers. e.g. i want ~/.gnupg linked into the device only when a certain card is input

22:58:16 * edd apologises for chat irrelevant to this channel but offers the trust -- gpg route as potential justification

22:58:19 <JibberJim> oh yeah, looks like it's got around 12 million of them (inurl:https anyway)

22:59:18 <DanC> hmm... what would happen if you put the wrong flash device in?

22:59:42 <DanC> I guess somebody could get you to believe a bunch of wrong keys

22:59:53 <edd> well, my script would dispatch on card id. so if the card ID didn't match...

23:00:19 <edd> actually on the card in question, my .gnupg directory is on a crypted filesystem, so i still need to type a passphrase in order to get at the .gnupg dir

23:00:59 <DanC> crypted filesystem makes sense... does it defeat the use of the cf disk for sneakernet to windows boxes?

23:01:10 <DanC> i.e. can you put more than one fs on the thing?

23:01:55 <edd> they only really support fat16, i think

23:02:16 <edd> i tried formatting one as ext2 once, worked, but very very very very slow

23:02:28 <DanC> how did you put a crypted fs in there? loopback?

23:02:29 <edd> the cryped fs would defeat use on windows, yes

23:02:38 <edd> yeah, loopback with cfs.

23:03:06 * DanC has been curious about this for a while...

23:03:23 <DanC> # apt-cache search cfs

23:03:23 <DanC> cfs - Cryptographic Filesystem

23:03:23 <JibberJim> can you not partition it edd? my USB can be partitioned in 2, so I have a clear, and a crypted one.

23:03:43 <edd> JibberJim: ooh, maybe i can. i can't say i've tried.

23:03:53 <DanC> hmm... cfs seems to rely on NFS. icky

23:03:53 <edd> two fat16 partitions eh

23:04:09 <edd> DanC: i tried several crypt solutions, they're all icky

23:05:03 * DanC wishes apt-get install would finish with a howto clue, ala "now try man cfs" or "look in /usr/doc/cfs/README for your next clue"

23:05:17 <aml> danc: yes. me too.

23:05:25 <aml> i usually have to dpkg -L

23:05:38 <edd> DanC: well, the standard is go read /usr/share/doc/<package>/README.Debian

23:06:06 * DanC wonders if that standard is documented

23:06:24 <dajobe> in debian, certainly

23:06:35 <DanC> when I first discovered apt, I found it *really* hard to find a path from the debian homepage to "here's what you should put in /etc/apt/sources" docs

23:07:09 <DanC> dajobe, where

23:07:10 <DanC> ?

23:07:26 * DanC wishes apt-get install would finish with a howto clue, ala "now try man cfs" or "look in /usr/doc/cfs/README for your next clue"

23:07:28 <danbri> +1

23:07:29 <edd> debian's website sucks for users. hard to see that changing in the near future

23:08:19 <DanC> debian's website rocks in lots of ways: weekly news, package search, security bulletins

23:09:33 <DanC> following my nose, I'm pleased to find "4.6 (How) Does Debian support Java?" in the debian FAQ: http://www.debian.org/doc/manuals/debian-faq/ch-software.en.html#s-java

23:10:47 <DanC> under "6 Basics of the Debian package management system" I find ready access to such arcana as "6.14 How do I build binary packages from a source package?" but no "so I installed a package. where do I find my next clue?"

23:12:22 <DanC> googling for "README.Debian site:debian.org" gets me a bunch of lintian reports

23:13:25 <DanC> aha... "Debian Policy Manual

23:13:25 <DanC> Chapter 13 - Documentation" http://www.debian.org/doc/debian-policy/ch-docs.html

23:13:44 <DanC> but that's for developers. I still don't see any user-facing documentation of the convention

23:17:32 <DanC> meanwhile, /usr/share/doc/cfs/README.Debian does pretty much have the clues i needed

23:20:54 <DanC> cmkdir: Invalid argument

23:22:02 <jllkifsh> jllkifsh is now known as jllykifsh

23:25:53 <DanC> hmm... I can cmkdir in my home directory, but not on the flash disk... I guess I need a loopback filesystem on the flash disk...

23:28:44 * edd runs out of time. up early to go and watch cricket tomorrow.

23:29:14 <DanC> "Using CryptoAPI" http://www.kerneli.org/howto/node3.php