Semantic Web Interest Group IRC Chat Logs for 2003-07-09

01:31:30 <MaryC> MaryC is now known as DanC_mac

01:43:06 * nb is away: afk

02:01:53 * DanC_jam resumes http://dm93.org/z2001/StudentOfMacOsX applescript/python stuff

02:07:05 <_joshua> AIEE

02:11:20 <em> em is now known as em-road

05:06:13 <mdupont> hi DanC_jam

08:09:44 * nb is back (gone 06:26:40)

10:03:18 <grault> grault is now known as earle

10:06:04 <earle> Hmm.

10:06:06 <earle> [[ ARTHUR C.CLARKE has another little brag: `I think I can claim to be the

10:06:09 <earle> godfather (with the good and bad implications that has) of the Web. Tim

10:06:12 <earle> Berners Lee acknowledges that my story "Dial F for Frankenstein"

10:06:15 <earle> (_Playboy_, January 1964) was one of the many inputs that started him

10:06:18 <earle> thinking on these lines.' (_Aerospace America_, May 2003) [CTLP] So, no

10:06:22 <earle> doubt, was Fredric Brown's slightly earlier `Answer' from 1954.

10:06:24 <earle> ]]

10:06:32 <earle> (from Ansible 192)

10:06:35 <mortenf> heh

10:17:28 <mortenf> ./ical2rdf.pl cal01.ics; no value type given, default unknown: tzid :/softwarestudio.org/Olson_20011030_5/America/New_York at ./ical2rdf.pl line 167, <> line 7.

10:41:00 * mortenf tries building python 2.2.3 from scratch...

10:41:15 <libby> this for cwm?

10:42:13 <mortenf> yep...

10:42:41 <libby> why build from scratch?

10:42:59 <mortenf> old system, new packages don't work

10:43:13 <libby> bummer

10:51:52 * mortenf wins, apparently

11:11:41 * mortenf wins, cwm seems to be running...

11:12:25 <libby> hurray!

11:28:18 * mortenf gets test.py running

11:28:28 <mortenf> http://www.w3.org/2000/10/swap/test/rdfcore-tests.n3 ; Sorry, Forbidden.

11:28:29 <dc_rdfig> A: http://www.w3.org/2000/10/swap/test/rdfcore-tests.n3 from mortenf

11:28:44 <mortenf> A:|RDF Core Tests

11:28:44 <dc_rdfig> Titled item A.

11:29:03 <mortenf> A:Sorry, Forbidden

11:29:03 <dc_rdfig> Added comment A1.

12:22:19 <sandro> mortenf, you can always get at it via dev.w3.org

12:24:16 <sandro> A:sorry, you can always use: http://dev.w3.org/cvsweb/~checkout~/2000/10/swap/test/rdfcore-tests.n3?rev=1.2&content-type=text/plain

12:24:16 <dc_rdfig> Added comment A2.

12:24:52 <sandro> A: or whatever, browsing from http://dev.w3.org/cvsweb/2000/10/swap/ or actually CVS checking out from there.

12:24:52 <dc_rdfig> Added comment A3.

12:29:24 <mortenf> ah, thanks.

14:58:04 <libby> BLURB: RDF calendar meeting today 2003-07-09, 1600 UTC

14:58:05 <dc_rdfig> B: RDF calendar meeting today 2003-07-09, 1600 UTC from libby

14:58:13 <DanC> .time utc

14:58:14 <datum> 2003-07-09T14:58:13

14:58:39 <libby> B:[2003-07-09, 1600 UTC|http://www.timeanddate.com/worldclock/fixedtime.html?day=9&month=7&year=2003&hour=16&min=0&sec=0&p1=0]

14:58:40 <dc_rdfig> Added comment B1.

14:59:24 <libby> BLURB: RDF calendar agenda item C: prodid support to ical2rdf.pl

14:59:24 <dc_rdfig> C: RDF calendar agenda item C: prodid support to ical2rdf.pl from libby

14:59:41 <libby> C:[see Libb's action|http://rdfig.xmlhack.com/2003/06/25/2003-06-25.html#1056556072.203619]

14:59:41 <dc_rdfig> Added comment C1.

14:59:54 <DanC> B:last time: [http://rdfig.xmlhack.com/2003/06/25/2003-06-25.html#1056550875.665269|25 June]

14:59:54 <dc_rdfig> Added comment B2.

15:00:29 <libby> BLURB:RDF calendar agenda item D: regression tests for ics->rdf and/or rdf->ics

15:00:29 <dc_rdfig> D: RDF calendar agenda item D: regression tests for ics->rdf and/or rdf->ics from libby

15:00:50 <libby> D:[see motenf's action item|http://rdfig.xmlhack.com/2003/06/25/2003-06-25.html#1056557026.871899]

15:00:50 <dc_rdfig> Added comment D1.

15:01:12 <DanC> B:see also rdf calendar task force [http://lists.w3.org/Archives/Public/www-rdf-calendar/|list archive], [http://esw.w3.org/topic/RdfCalendar|wiki topic], and [http://www.w3.org/2002/12/cal/|workspace]

15:01:13 <dc_rdfig> Added comment B3.

15:01:23 <libby> BLURB: RDF calendar agenda item E: InterpretationProperties esp as applied to timezones

15:01:23 <dc_rdfig> E: RDF calendar agenda item E: InterpretationProperties esp as applied to timezones from libby

15:01:50 <libby> BLURB:RDF calendar agenda item F: skical optimeset

15:01:51 <dc_rdfig> F: RDF calendar agenda item F: skical optimeset from libby

15:02:00 <libby> back soon....

15:14:53 <DanC> I'm experimenting with digitally signing my mail...

15:15:13 <DanC> I thought I knew pretty well how this works, but somebody suggested I hadn't signed my key.

15:15:15 <DanC> I thought I did...

15:15:25 <DanC> ... now... is there any reasonable documentation on how to use this stuff?

15:16:08 <DanC> sigh... debian docs start right in with minutiae... "Due to a bug in the way secret keys were encrypted ..."

15:16:24 * JibberJim wonders if there are any filters that can remove digital signatures at the mail-server...

15:16:57 <DanC> ok, /usr/share/doc/gnupg/README.gz seems like it starts from 'hello world'

15:17:27 <DanC> "The default algorithms are DSA and ElGamal"

15:17:54 <DanC> hmm... I was looking for info on whether DSA or RSA was preferred... oh... that was for ssh2

15:20:13 <DanC> README.gz refers to http://www.gnupg.org/docs.html

15:21:46 <DanC> sigh... nothing in the 'creating a key' section about signing your own key. http://webber.dewinter.com/gnupg_howto/english/GPGMiniHowto-3.html#ss3.1

15:22:07 <Karl-MTL> maybe http://www.skuz.net/pgp4dummies

15:22:33 <DanC> that was a mini-howto; I'm checking the more full manual now... http://www.gnupg.org/gph/en/manual.html#AEN26

15:22:56 <mattb> gpg --sign-key yourname

15:23:12 <mattb> looking at my keyring i see i've done it

15:23:24 <mattb> not sure based on what advice

15:24:07 * DanC dispairs that open source software will ever become usable

15:24:33 <mattb> crypto particularly

15:24:41 <DanC> take cwm, for example. I know the docs suck. do I fix them? no. I hack the next feature.

15:24:51 <mattb> cf http://usefulinc.com/edd/blog/2003/6/28#21:05

15:25:31 <nb> that's why we need more free software businesses

15:26:28 <nb> hackers write code. paying customers pay for usability and docs

15:27:16 <mattb> crypto is particularly tricky - commercialisation of pgp hasn't exactly taken the world by storm

15:27:18 <DanC> it doesn't have to be that way.

15:27:32 <DanC> (that way = no good docs until users pay)

15:28:19 <nb> pgp isn't fere software

15:28:44 <nb> with free software, everyone has the functionality, competition will be on usability and docs

15:28:56 <DanC> I hope for a future where unless your software has reasonably good 'getting started' docs, you just won't get noticed.... kinda like if you released code in today's community that doesn't build with ./configure;make;make installl

15:30:38 <DanC> packagin systems (e.g. debian) should incorporate smoke-testing and/or getting-started usage docs.

15:30:54 <DanC> i.e. after you install it, there should be some way to verify that it works, at least minimally

15:31:06 <DanC> (smoke-testing)

15:31:30 <DanC> and preferably, there should be good "how to X, Y, and Z" support for the top three features.

15:31:38 <Karl-MTL> hmmm :)

15:32:07 <Karl-MTL> documentation and good organisation of documentation are essential.

15:32:49 <Karl-MTL> But sometimes people who develop software, specs don't see the benefits of it.

15:33:23 <DanC> the norm seems to be "you can configure this package any way you want! it can do zillions of different things!" but they don't tell you *exactly* how to do the most popular/useful thing.

15:34:25 <DanC> google says the homepage of cwm is http://infomesh.net/2001/cwm/

15:35:35 <sandro> Too bad the title obeys the reverse naming principal. (Have I documented that anywhere? As soon as you name anything, even if the name was perfect, the name becomes wrong.)

15:35:53 <JibberJim> SBP's more google loved than w3?

15:36:04 <DanC> sigh... exhaustive enumeration of flags... http://www.w3.org/2000/10/swap/doc/cwm

15:36:32 <sandro> why sigh?

15:37:02 <danbri> so long as they're up to date, that seems a good thing

15:37:35 <DanC> exhaustive enumeration isn't useful unless you already have basic intuitions about how to compose them.

15:37:42 <sandro> This page or one like it was actually being generated automagically. Not sure about this one.

15:38:03 <DanC> exhaustive enumeration is why man pages suck

15:38:12 <sandro> huh.

15:38:43 <DanC> try man tar... only after a blizzard of details do you, maybe, see the basic idiom: tar zxvf foo.tgz

15:39:14 <DanC> ooh! I'm wrong! the tar man page starts with 3 examples!

15:39:18 <sandro> It's the old conflict between Tutorial material and Reference material, right?

15:39:42 <DanC> yes, it's an old conflict. but it should be clear by now who the winner is.

15:39:50 <DanC> or: which is essential for the first step

15:40:09 <DanC> somebody that wants to know all the cwm options can UTSL.

15:40:13 <sandro> I thought it was clear that both were important, and one shouldn't muddle them together.

15:40:34 <danbri> man zip, is worse

15:40:59 <danbri> wtf, Gnome keeps forgetting what alt-tab does... :/

15:41:08 <sandro> man zip, for me, is fine. It also has examples on the second page.

15:41:13 <DanC> nobody learns from exhaustive enumeration. people learn by successive elaboration. exhaustive enumeration is for reference, not for "how to use this thing"

15:41:36 <DanC> No manual entry for zip

15:41:37 <DanC> See 'man 7 undocumented' for help when manual pages are not available.

15:44:53 <DanC> hmm... does "forward chaining reasoner" belong in the hello-world description of cwm?

15:45:28 * danbri tries to think of an alternate gloss

15:45:36 * danbri fails

15:46:16 <DanC> currently, it says [[ Cwm is a general-purpose data processor for the semantic web. It is a

15:46:16 <DanC> forward chaining reasoner which can be used for querying, checking,

15:46:16 <DanC> transforming and filtering information. ]]

15:46:33 <DanC> the 2nd sentence could just start: It can be used for querying, ...

15:47:20 <dmwaters> {global notice} Good afternoon everyone, apparrently one of our main rotation servers has crashed, and the problem is being looked at. Any further messages will be given in wallops.

15:47:57 * bijan thinks so

15:49:54 <DanC> the "Cwm is written in python" sentence seems out of place

15:50:27 <danbri> yep, in that context it is ok, it introduces a bit of technical terminology (which will be familiar to many who are capable of making immediate use of it), but the intro gives some examples of usage

15:51:19 <DanC> which is ok, danbri? you prefer 'forward chaining reasoner' in or out?

15:51:32 * DanC gathered bijan prefers out, but isn't sure now

15:51:41 <bijan> Er..

15:51:42 <bijan> no

15:51:43 <danbri> in, given the context (but I don't care strongly)

15:51:44 <bijan> In

15:51:49 <danbri> I took bijan to mean 'in'

15:52:02 <bijan> Reasonably strongly prefer in, actually

15:52:06 <sandro> To the AI crowd, saying it's a "forward chaining reasoner" says a lot, I think. (not just about what it is, but about how we talk about it, etc.)

15:52:26 <danbri> yes, those three words carry a lot, for those who recognise them

15:52:30 <bijan> Well, also, it always me to look it up

15:52:39 * DanC puts it back in

15:52:51 <bijan> And there are plenty of good entry level descriptoins of what it is and strengths and weaknesses, etc.

15:54:10 <DanC> hmm... should I focus more on http://www.w3.org/2000/10/swap/ or http://www.w3.org/2000/10/swap/doc/cwm ?

15:54:11 <sandro> Meanwhile.... I'm reading the daml axioms (http://www.ai.sri.com/daml/axioms.kif) and thinking about how each rule has a name and documentation, and wondering how to do that in N3.

15:54:35 <sandro> { .... } a :Rule, rdfs:comment "...." ., ....

15:54:41 <DanC> it's kind of a pain in N3... yeah, like that...

15:55:04 <DanC> I don't think I'd bother

15:55:18 <DanC> or I'd just use #n3 comments

15:55:35 <DanC> I don't think it's all that interesting to institutionalize the rules.

15:55:39 <sandro> Depend on the value of collaborating with KSL and SRI, I think.

15:56:23 <DanC> "The python source files are available as a compressed tar file" <- do we really support that?

15:56:53 <sandro> Tim likes to. :-)

15:57:08 <DanC> tim likes a lot of stuff.

15:57:08 <Karl-MTL> Sandro, Danc, do you know if you had success modifying the templates for the wiki tools to make them valid?

15:57:39 <DanC> it's not as simple as modifying the templates, unfortunately, karl

15:58:08 <DanC> we asked the moinmoin folks about it, and they didn't even seem all that interested in accepting patches (or am I misremembering?)

15:58:25 <danbri> you're correct

15:58:34 <danbri> i've visited the code and #moin twice re that mission

15:58:40 <sandro> The MoinMoin folks want to do a major rewrite of some related section FIRST, and that's not done yet.

15:58:43 <danbri> i fixed some bugs, but more would have been a big effort

15:58:45 <libby> .time

15:58:51 <danbri> they said it is all in progress, nothing i could do to help

15:59:05 <Karl-MTL> hmm ok. I was about to help them to fix.

15:59:24 <danbri> you could always offer again, it might help encourage

15:59:24 <Karl-MTL> so it's better to wait they have fixed their engines.

15:59:30 <danbri> #moin

15:59:38 <Karl-MTL> ok. I'll do it :)

16:00:29 <libby> I think it's time for calendaring

16:00:34 * mortenf wonder if he's still connected...

16:00:37 <mortenf> ah, guess so.

16:00:38 <libby> -----rdf calendar meet ----

16:00:40 <mortenf> yep.

16:01:18 <DanC> is this the auto-generated page, sandro? http://www.w3.org/2000/10/swap/doc/CwmHelp

16:01:31 <libby> B:[libby miller attending|http://ilrt.org/people/libby/]

16:01:31 <dc_rdfig> Added comment B4.

16:01:44 <libby> is anyone else here for the calendar meeting?

16:02:00 <DanC> among other things, yes

16:02:06 <mortenf> B:[Morten Frederiksen|http://purl.org/net/morten/] attending

16:02:07 <dc_rdfig> Added comment B5.

16:02:18 <DanC> B:[http://www.w3.org/People/Connolly/|Dan Connolly] attending

16:02:19 <dc_rdfig> Added comment B6.

16:02:50 <danbri> B:[http://www.w3.org/People/DanBri/|Dan Brickley] attending

16:02:50 <dc_rdfig> Added comment B7.

16:03:04 <libby> would anyone like to add anything to the agenda?

16:03:27 <sandro> Pretty sure yes, DanC

16:03:42 <libby> i'll take that as a no.

16:03:54 <libby> if anyoe thinks of anything, addit to the chump

16:03:59 <sandro> see http://www.w3.org/2000/10/swap/doc/Makefile

16:04:20 <libby> cudl someone change the topic for me?

16:04:24 <danbri> libby chairing?

16:04:30 <mortenf> libby++

16:04:40 <libby> ok, will do, thanks

16:04:55 <DanC> DanC has changed the topic to: RdfCalendar meeting 1600Z Weds http://rdfig.xmlhack.com/ http://esw.w3.org/topic/RdfCalendar

16:04:57 <libby> so we have a rough agenda in http://rdfig.xmlhack.com

16:05:30 * DanC looks around for george...

16:05:33 <mortenf> looks fine

16:05:38 <libby> does anyone have any preferences about what to talk about first?

16:05:50 <mortenf> nah :)

16:05:58 <libby> I've done some on C, but not on E or F where I also have actions outstanding

16:06:06 <libby> ok, lets go with C

16:06:08 <mortenf> C sounds good

16:06:15 <libby> RDF calendar agenda item C: prodid support to ical2rdf.pl

16:06:55 <mortenf> i ran into a problem with that, i think

16:07:04 <libby> basically I've done some of this - I uploaded a reworked version of ical2rdf.pl earlier today, which seems to work (migh need a bit more tweaking)

16:07:10 <mortenf> nah, that wasn't it.

16:07:20 <libby> yeah, what was that mortenf?

16:07:57 <DanC> hmm... so ical2.rdf.pl isn't stand-alone any more. bummer

16:08:06 <libby> no

16:08:15 <libby> actually I thouht it wasn;t anyway

16:08:24 <mortenf> "no value type given, default unknown: tzid :/softwarestudio.org/Olson_20011030_5/America/New_York at ../ical2rdf.pl line 167, <> line 7."

16:08:31 <mortenf> for cal01.ics

16:08:35 <DanC> ah... Getopt::Long

16:09:01 <libby> right morten. that's one of thie things was thinking about.

16:09:23 <DanC> what input file are you using, mortenf?

16:09:34 <libby> re not standalone, I did wonder about that. I guess it would eb possible to put a sha1 functionin there - itwoudl just be a bit long

16:09:53 <libby> the problem motenf found is because in that file there's no prodid

16:09:55 <mortenf> cal01.ics

16:10:09 <mortenf> ah, works now!

16:10:12 * danbri waves to amy

16:10:17 <mortenf> i must have had an old version

16:10:18 <libby> I'm inclined to think that's a bug. I added the timezone stuff later to existing files

16:10:35 <libby> hey amy

16:10:42 <amy> hi libby!

16:11:06 <libby> anyway, I don;t know how you feel about me mangling your code DanC....my perl's not so good

16:11:30 <libby> ...but it does produce the same output as the java versin, so that's nice.

16:11:32 <DanC> you're not alone. most people's perl is ugly, to my eye. mine included

16:12:04 <DanC> ok, so I updated my copy of ical2rdf.pl ... do we have smoke tests? how can I tell if it's busted?

16:12:11 <libby> I still need toupdate the tests in the repository. also I started to use cwm to find all the xproperties so I could make soe schemas for them, but not finished that yet/.

16:12:34 <mortenf> i ran the new version on a couple of files, seems fine.

16:13:04 <libby> re smoke tests...I've only checked with one. file. I have the tools to check them all, I just havn;t had time yet.

16:13:38 <mortenf> seems like getting to my item on the agenda? :)

16:13:38 <DanC> when I do 'make testcase' it still uses the --xnames arg

16:13:50 <libby> but the checing would only be against the java stuff...which might also be busted :)

16:13:58 <DanC> sure, mortenf. no need to keep the agenda items totally separate

16:14:05 <libby> yep sorry, was a bit of a rush job

16:14:17 * DanC thinks you brits apologize too much

16:14:28 <mortenf> np, just a remark re the purpose of my item!

16:14:28 <libby> yeah, go for it motrten - that's all I have to report

16:14:42 <mortenf> not that have that much to report...

16:14:46 <DanC> er... let's not leave prodid altogether yet...

16:14:51 <mortenf> ok

16:14:54 <libby> ok

16:15:22 * danbri reminds DanC of the concept of a 'british apology' (cf The Empire; "sorry 'bout the mess, world..."...)

16:15:27 * danbri has to step out briefly; bbiab

16:15:47 <DanC> hmm... the old code for X- properties is still in there... die "X- property ($n) requires --xnames arg"

16:15:55 <libby> ok, well things to fix wrt prodid: update the tests in test/ ; fix makefile

16:16:17 <libby> yep and taht may still happen if there's an xprop but no prodid, I think

16:16:34 * DanC can't find the new code for x- stuff... hmm...

16:16:42 <libby> I was wondering if it made sense to say that that shoudl never happen tho - that it's an error

16:16:51 <DanC> oh... I see...

16:16:59 <libby> within the ica;l2rdf.pl file?

16:17:17 <DanC> $X_ns=createxns($field);

16:17:25 <libby> yep

16:18:10 * DanC wonders how you got the X_ns set up before the namespaces are written out... I thought that was going to be hard...

16:18:49 <DanC> please don't comment code out. just delete it. rely on CVS for history.

16:18:54 <libby> hm, I just made an enormous string of the rest of it and printed it out after first pass ver

16:18:58 <libby> ok

16:19:12 <DanC> enourmous string. ah.

16:19:55 <libby> which is much the same technique as I used in the java stuff

16:20:10 * DanC adds pointer to SHA1 module in cpan...

16:21:42 * DanC can't find the sha1 pointer now... are my eyes busted?

16:22:14 <libby> its in the cvs I think

16:22:30 <libby> http://search.cpan.org/author/GAAS/Digest-SHA1-2.02/

16:23:23 <libby> so, are we done with that topic?

16:23:38 <libby> I'll do a bit more tweaking

16:24:31 <DanC> hmm... should I leave the --xnames arg? issue a 'deprecated' warning?

16:24:53 <libby> do you think it's an error if no prodid and have an xprop?

16:25:00 <libby> I think I do

16:25:14 <DanC> let's say it's an error of there's no prodid *and* no --xnames arg

16:25:21 <libby> ok

16:26:05 <DanC> I added a warning if you supply both:

16:26:07 <DanC> perl ical2rdf.pl --xnames http://www.w3.org/2002/12/cal/prod_evo# test/cal01.ics >test/cal01.rdf

16:26:07 <DanC> clobbering --xnames http://www.w3.org/2002/12/cal/prod_evo# at ical2rdf.pl line 151, <> line 4.

16:26:41 <libby> cool I do the same in the java code too

16:26:57 <mortenf> does all the test cases have prodid?

16:27:18 <libby> I've not checked

16:27:41 <libby> because I copuied and pasted in tz info into some of them, they may be soe xprops snick in there by mistake

16:28:07 <libby> I think some of the files were handwritten

16:28:53 * DanC commits ical2rdf.pl,v 1.15

16:29:14 <libby> heh, I got one really werd prodid and it must have come from here: http://www.w3.org/2002/12/cal/test/gkexample.rdf

16:29:44 <libby> I need to check if you haveto ahve the -// thing in your prodid; I relied on that

16:30:08 <DanC> let's check that now...

16:30:28 <libby> this doesn;t ave prodid btw: http://www.w3.org/2002/12/cal/test/querymeetings.ics

16:31:19 <DanC> it's an error not to have a prodid: 'prodid' and 'version' are both REQUIRED -- 4.6 Calendar Components http://www.ietf.org/rfc/rfc2445.txt

16:31:22 <libby> hm, I didn;t know tnhat: 'prodid' and 'version' are both REQUIRED,

16:31:25 <libby> heh

16:32:01 <libby> 4.7.3 prodid: Description: The vendor of the implementation SHOULD assure that this

16:32:01 <libby> is a globally unique identifier; using some technique such as an FPI

16:32:01 <libby> value, as defined in [ISO 9070].

16:32:19 <libby> ok, so should not rely on -//

16:32:27 <DanC> quite.

16:33:21 <DanC> umm... perl is naturally ugly, but let's do indent the code, ok?

16:33:24 <DanC> do you use emacs?

16:33:28 <libby> no

16:33:46 <DanC> how do you indent your java code?

16:34:24 <DanC> hmm... did we really pick just the first 5 chars of the sha1? interesting.

16:34:44 <libby> hm, when I do, OI do it at the end using a pretttifier. I'm aware it's not ideal

16:34:54 <DanC> ah... you did include a pointer to http://search.cpan.org/author/GAAS/Digest-SHA1-2.02/ in the code; just not right next to the 'use' statement.

16:34:58 <libby> I throught that was what we initially decided.

16:35:15 <libby> though I didn;t check what we decided last meet

16:36:08 * DanC commits ical2rdf.pl,v 1.16 with fix for not relying on -//

16:36:58 <DanC> nothing about 1st 5 chars in http://rdfig.xmlhack.com/2003/02/26/2003-02-26.html#1046279854.884486 but maybe it's in the full log. no biggie.

16:37:10 <libby> 17:13:03 <mortenf> 1st 10 + sha1 of all: ns

16:37:10 <libby> 17:13:20 <libby> ok, full sha1, gotya

16:39:10 <libby> I came out of that meetign a buit unsure

16:39:14 <DanC> change to full sha1, or leave it until we find a collision?

16:39:38 <mortenf> shouldn't be that big a change?

16:39:45 <libby> so with a fell sha, is the filename going to be ximian_not_764237654723674628756

16:39:47 <libby> ?

16:39:51 <libby> i.e. a huge filename?

16:40:00 <DanC> something like that, yes. doesn't seem that huge to me.

16:40:07 <libby> it's not a big chenge until we 0put a schema there (I think)

16:40:21 <DanC> change 5 to 16 chars, maybe?

16:40:31 <libby> ok by me

16:40:56 <mortenf> what i thought we decided was 10 first chars + 5 chars of sha1 of full

16:42:23 <libby> i.e. something like /prod/ximian_not_76423

16:42:24 <mortenf> as opposed to + 5 chars of sha1 of first 5/10

16:42:27 <mortenf> yep.

16:42:49 * DanC commits Makefile with --xnames removed

16:43:24 <libby> I do;t really have an option. I guess the filename/ns will look uglier if longer, but probbaly doesn;t matter much. I don;t kow the probbalilities of collisions

16:43:25 * DanC commits ical2rdf.pl,v 1.17 with 16 chars of sha1

16:43:32 <libby> s/option/opinion/

16:43:41 <DanC> xmlns:x='http://www.w3.org/2002/12/cal/prod/Ximian_NON_f7c377930cb82492#'

16:44:02 <libby> ok

16:44:06 <mortenf> how long until they show up from cvs?

16:44:23 <DanC> umm... dev.w3.org lags by up to an hour

16:44:30 <mortenf> hm

16:44:53 <DanC> but http://www.w3.org/2002/12/cal/ical2rdf.pl lag is more like 15 seconds

16:44:55 <libby> coupla little things: I lowercased the prodid before sha1ing it; also replaces staces and // with _. make sense in tht order? I did not include the -// in sha1 because I figured it woudl appear in all of them

16:45:09 <DanC> little things: ok.

16:45:11 <libby> not sure why I lowercased it

16:45:13 <mortenf> hmm, strange, only just got libby's update from earlier

16:45:22 <libby> seemed lie a good idea at the time

16:45:47 <libby> s/staces/spaces/

16:45:51 <DanC> now I'm curious about round-trip testing

16:46:13 <libby> ah

16:46:25 <libby> hm, the prodid is still there in full in the rdf file

16:46:36 <libby> but you coudln;t roundtrip solley from the xns

16:46:44 <libby> sorry baout terrible spelling today

16:50:23 * libby wonders if people woudl like to move to D - regression tests for ics->rdf and/or rdf->ic

16:50:37 <mortenf> sure

16:51:06 <libby> cool

16:51:18 <libby> DanC? you done with C?

16:51:25 <libby> for now at any rate?

16:51:43 <DanC> for now, yes

16:51:49 <libby> ok, cool

16:51:51 <libby> RDF calendar agenda item D: regression tests for ics->rdf and/or rdf->ics

16:52:07 <libby> mortenf? ho0w are you doing with this?

16:52:14 <DanC> logger, pointer?

16:52:17 <DanC> logger_2, pointer?

16:52:17 <DanC> See http://ilrt.org/discovery/chatlogs/rdfig/2003-07-09#T16-52-17

16:52:20 <mortenf> not too much interesting to report (yet), see http://www.wasab.dk/morten/2003/06/cal/action.html

16:52:26 * nb is away: afk

16:52:39 <DanC> D:[http://ilrt.org/discovery/chatlogs/rdfig/2003-07-09#T16-52-17|irc log]

16:52:40 <dc_rdfig> Added comment D2.

16:52:40 <mortenf> i don't grok how regression.n3 works...

16:53:02 <mortenf> D:seeAlso http://www.wasab.dk/morten/2003/06/cal/action.html

16:53:03 <dc_rdfig> Added comment D3.

16:53:26 <mortenf> D3:seeAlso [http://www.wasab.dk/morten/2003/06/cal/action.html|Round-tripping experiences...]

16:53:26 <dc_rdfig> Replaced comment D3.

16:53:34 * DanC reads...

16:54:14 <DanC> D:hmm.. good fodder for [http://esw.w3.org/topic/CwmTips|CwmTips]

16:54:15 <dc_rdfig> Added comment D4.

16:54:21 <mortenf> yep :)

16:54:36 <mortenf> prolly rewritten...

16:55:04 <DanC> ok, so you've got cwm running, it looks like. the few, the proud, the cwm users...

16:55:20 <libby> heh

16:55:20 <mortenf> heh, and the two calendar converters

16:55:28 <libby> oh cool

16:55:30 <DanC> shall we work thru comparing a couple .rdf files now? or leave that for offline?

16:55:52 <mortenf> so, the question i have: what should the "Manifest" look like (i don't have access to the rdf-core Manifest)?

16:55:58 <libby> well this si really pretty much irt agenda-wise, so if you want to use the time

16:56:15 <libby> sorry, garbled. too hot to type here

16:56:40 <DanC> hmmm... manifest... just a bunch of pointers to test files, I'd guess

16:56:45 <libby> the last 2 agenda items are actions for me which I havn;t done. so we still have 30 mins plus we could use for this.

16:56:47 <mortenf> oh hang, i do...

16:57:01 <DanC> shall we start with cal01.rdf?

16:57:31 <mortenf> so how to define the test actions (re regression.n3, it seems to just accept cwm arguments)?

16:57:35 <DanC> libby, where are your testing notes again, please?

16:57:47 <libby> building the schema using cwm?

16:57:50 <mortenf> (pointers gladly accepted)

16:57:53 <DanC> I don't think regression.n3 is all that relevant to what we're up to

16:58:19 <libby> - http://esw.w3.org/topic/RdfCalendarSchema

16:58:19 <DanC> no, not building the schema... your notes on testing your java code... didn't you use a graph compare thingy to check its output against my perl code?

16:58:36 <mortenf> i can do something without cwm, but i thought it'd be nice for something that could prove the tests

16:58:49 <DanC> I propose that we try to work thru one round-trip test in this meeting

16:58:55 <libby> ok re java stuff

16:58:57 <mortenf> cool

16:59:05 <DanC> shall we start with cal01.rdf?

16:59:10 <mortenf> ysure

16:59:22 <DanC> the round trip goes .rdf->.ics->.rdf , not .ics->.rdf->.ics, right?

16:59:31 <libby> there's a bit there: http://www.ilrt.bris.ac.uk/discovery/2003/02/cal/readme.txt

16:59:40 <mortenf> hmm, isn't the originals the ics files?

16:59:58 <libby> yeah I thougt so

17:00:12 <DanC> D:[http://www.ilrt.bris.ac.uk/discovery/2003/02/cal/readme.txt|notes on java implementation] by libby includes testing notes

17:00:14 <dc_rdfig> Added comment D5.

17:00:42 * mortenf now see ical2rdf v 1.17

17:00:52 <DanC> we covered this last time, no? we don't have good tools for comparing .ics files, and we'd need them for doing .ics->.rdf->.ics, no?

17:01:19 <libby> D:there's a little more information about the java comparison [in the makefile|http://www.ilrt.bris.ac.uk/discovery/2003/02/cal/mimedir-parser/Makefile]

17:01:19 <dc_rdfig> Added comment D6.

17:01:33 <mortenf> ok, but then how'd we know where any errors occurred?

17:01:42 <libby> good point. but our original files are ics files

17:02:18 <DanC> we start by manuall inspecting the expected results. these are regression tests...

17:02:28 <DanC> ... they allow you to tell if a change you made in the code broke something.

17:02:43 <DanC> the machine can't tell whether you got it right in the first place. we have to do that manually.

17:02:47 <mortenf> that's ok for manual, but if the point was automated testing?

17:03:00 <DanC> it's automated *regression* testing

17:03:29 <DanC> i.e. once the correct answers are checked in, you can check them out and see if your installation is working

17:03:58 <DanC> let's work thru the 1st case; maybe it'll become clear

17:04:14 <mortenf> ok, i'll see if i can follow anyway; 1st case: go!

17:04:33 <DanC> ok, we start with 2002/12/cal/test/cal01.rdf ...

17:05:09 <DanC> we got it by 'cheating' , i.e. by taking a known-good .ics file and generating it with ical2rdf.pl .. but it doesn't matter how we got it.

17:05:32 <DanC> now we try the round-trip... let's convert it to .ics ...

17:06:48 <DanC> $ PYTHONPATH=../../../../2000/10/swap/ python ../../../../2000/10/swap/pim/toIcal.py cal01.rdf >,cal01.ics

17:06:58 <DanC> do you have swap/pim/toIcal.py ?

17:07:02 <mortenf> got it

17:07:49 <DanC> now we eyeball the ,cal01.ics file... and/or load it into evolution/apple-ical etc. maybe

17:08:50 <DanC> hmm... using diff to compare ,cal01.ics with cal01.ics (from CVS) produces lots of spurious output (as expected). for a Q&D hack, let's sort and compare...

17:09:11 * mortenf got line-ending issues, fixing...

17:09:56 <mortenf> hmm.

17:10:53 <DanC> yeah... line-end issues... this is no fun... let's assume ,cal01.ics is OK, for now...

17:11:15 <mortenf> hmm, does python really do xDxA ?!?

17:11:27 <DanC> let's convert it back to .rdf...

17:12:07 <DanC> $ mkdir ,temp

17:12:11 <DanC> $ perl ../ical2rdf.pl ,cal01.ics >,temp/cal01.rdf

17:12:57 <DanC> now we've got the input/expected-results in cal/test/cal01.rdf and the experimental results in cal/test/,temp/cal01.rdf

17:13:01 <DanC> let's compare them...

17:13:04 <mortenf> heh, enough round-trips and i think we reach <rdf:RDF/>

17:13:53 <mortenf> hm, toIcal doesn't handle the new prodids?

17:14:49 <DanC> doesn't handle... in what way?

17:15:14 <DanC> the new prodid design shouldn't matter in that direction, should it?

17:15:25 <mortenf> it makes "PRODID:-//w3.org/2002/01dc-nj/toICal.py//NONSGML v1.0/EN"

17:15:53 <mortenf> in ./,cal01.ics

17:16:05 <DanC> yes. what did you expect?

17:16:14 <DanC> it's only the namespace names that are mangled

17:16:32 <mortenf> well, the input rdf has a real Ximian prodid

17:16:44 <DanC> yup.

17:16:47 <earle> earle is now known as grault

17:16:51 <mortenf> ok, you're right, it's not the new stuff that's the reason

17:23:52 * DanC having trouble getting cwm to compare .rdf files...

17:24:10 <mortenf> how are you trying?

17:24:32 <DanC> $ PYTHONPATH=../../../../2000/10/swap/ python ../../../../2000/10/swap/cwm.py graphCompare.n3 --think --with cal01.rdf ,temp/cal01.rdf >,xxx.n3

17:24:41 <DanC> where graphCompare.n3 is, essentially [[

17:24:47 <DanC> {

17:24:47 <DanC> "1".os:argv log:semantics ?G1.

17:24:47 <DanC> "2".os:argv log:semantics ?G2.

17:24:47 <DanC> } => { ?G1 :testingAgainst ?G2 }.

17:24:50 <DanC> {

17:24:52 <DanC> ?G1 :testingAgainst ?G2.

17:24:54 <DanC> ?G1 log:includes ?G2.

17:24:56 <DanC> ?G2 log:includes ?G1.

17:24:58 <DanC> } => { :thisTest a :Success }.

17:25:00 <DanC> ]]

17:25:05 <mortenf> ah

17:25:07 <DanC> aha! missing log:uri...

17:27:13 <sandro> hm - does that work?

17:27:44 <DanC> not yet

17:27:48 * libby has to go in 5 mins

17:27:57 <libby> is there any bview on a next meet?

17:28:23 <mortenf> i'd like one; also to continue my action...

17:28:36 <danbri> sorry i've just been lurking this meeting, very happy to see this sort of detailed work happen... hope to be more helpful next time

17:28:43 <libby> right. not many people here but interesting stuff..

17:30:03 * mortenf now thinks he graps the concept of regression tests completely; the key is the existance of expected output at all stages...

17:30:31 * libby proposes 2003-07-23, 1600 UTC for next meet: http://www.timeanddate.com/worldclock/fixedtime.html?day=23&month=7&year=2003&hour=16&min=0&sec=0&p1=0

17:30:45 <libby> yeah I'm clearer too re regression tests

17:30:53 <mortenf> ok with me

17:31:00 <libby> objections?

17:31:13 <mortenf> DanC seems to be stuck in a graph somewhere...

17:31:19 <libby> heh

17:31:25 <DanC> 23jul... i think I'm at a TAG ftf then

17:32:03 <libby> could do 30th jul, not sure about 6th aug

17:32:09 <mortenf> 6th no good

17:32:23 <mortenf> on my way to Skanderborg festival then...

17:32:29 <libby> what's that?

17:32:44 <mortenf> www.smukfest.dk

17:32:50 <libby> you going to CCC mortenf?

17:33:05 <mortenf> no, not enough vacation time, i think; what date?

17:33:19 <libby> how about 30th july then?

17:33:28 <mortenf> fine with me

17:33:37 <libby> 7-10 aug I think

17:33:50 <DanC> 23jul... can somebody check the tag ftf date for me?

17:33:50 <libby> any other takers for 30th Jul?

17:33:52 <mortenf> uuh, can't do that.

17:33:59 <mortenf> ccc, that is

17:34:08 <libby> I make it only 21/22 danc

17:34:11 <mortenf> overlap...

17:35:21 <DanC> I return on Thu... 24th. http://www.w3.org/2003/07dc-yvr/yvr.rdf

17:35:39 * libby really has to run - leigh dodds visiting. I can do 23 or 30th and so can morten, so if DanC can only make one of those, swe should go withh that one.

17:35:41 <libby> ah

17:35:47 <libby> so how's 30th for you DanC?

17:36:16 <DanC> 30july looks ok

17:36:21 <mortenf> tell leigh hi!

17:36:31 <libby> will do!

17:36:38 <mortenf> 30th it is (chat consensus)!

17:36:46 * mattb waves hi to leigh

17:36:47 <libby> excellent

17:36:59 <DanC> logger_2, pointer

17:36:59 <DanC> See http://ilrt.org/discovery/chatlogs/rdfig/2003-07-09#T17-36-59

17:37:26 * DanC links to that pointer from http://calendar.sidekick.dngr.com/event?id=827

17:37:32 <libby> cool :)

17:38:02 <libby> B:[next meet 2003-07-30, 1600 UTC|http://www.timeanddate.com/worldclock/fixedtime.html?day=30&month=7&year=2003&hour=16&min=0&sec=0&p1=0]

17:38:02 <dc_rdfig> Added comment B8.

17:38:03 <DanC> hmm... cwm seems to think the cal01.rdf output doesn't match the input. I'm not sure why, though.

17:38:24 <DanC> sandro, should this work? [[

17:38:27 <DanC> {

17:38:27 <DanC> ?F1 :testingAgainst ?F2.

17:38:27 <DanC> ?F1 log:semantics ?G1. ?F2 log:semantics ?G2.

17:38:27 <DanC> ?G1 log:includes { ?S ?P ?O }.

17:38:27 <DanC> ?G2 log:notIncludes { ?S ?P ?O }.

17:38:28 <DanC> } => { ?F2 :missing { ?S ?P ?O } }.

17:38:29 <libby> is theer any way you can get a clue why? jena one doesn;t, just says fails. makes debuggiung very hard

17:38:30 <DanC> ]]

17:38:39 <mortenf> back in 10, would like to continue the "testing"...

17:39:13 <DanC> ah! can't use ?S inside {}...

17:39:41 <libby> I think we're 'officially' adjourned.... - see you later guys. thanks mortenf, DanC, danbri...

17:39:54 <DanC> thanks for chairing again, libby.

17:39:56 <mortenf> see you, thanks libby

17:39:59 <libby> no prob

17:40:01 <libby> bye

17:40:13 <DanC> DanC has changed the topic to: Semantic Web hack-n-chat http://rdfig.xmlhack.com/ http://esw.w3.org/topic/

17:40:59 <DanC> <,temp/cal01.rdf> gcomp:missing {:_g0 i:action i:display .

17:40:59 <DanC> },

17:41:03 <DanC> etc.

17:43:23 <DanC> I think I'm gonna try a smaller test case.

17:43:29 <DanC> mtg.rdf

17:46:54 <mortenf> ok, so here's how i see it: mtg.ics is input, mtg.rdf is generated by ical2rdf and inspected, verified; then toIcal can be run on that, and the result of running ical2rdf on output from that should then match with the semi-original mtg.rdf

17:53:40 <DanC> right.

17:56:06 <DanC> gcomp:thisTest a gcomp:Success .

17:56:06 <DanC> !

17:56:12 <mortenf> nice!

17:56:20 <mortenf> share, share...

17:56:34 * DanC checks in...

17:58:50 <DanC> D:[http://www.w3.org/2002/12/cal/Makefile|Makefile] updated to do one round-trip test...

17:58:51 <dc_rdfig> Added comment D7.

17:59:24 * mortenf will wait for update to www...

17:59:28 <DanC> D:... requires new [http://www.w3.org/2002/12/cal/test/graphCompare.n3|graphCompare.n3]

17:59:28 <dc_rdfig> Added comment D8.

17:59:42 <DanC> it's available now

18:00:10 <mortenf> nice, thanks.

18:00:16 <DanC> $ make test

18:00:17 <DanC> perl ical2rdf.pl test/mtg.ics >test/mtg.rdf

18:00:17 <DanC> PYTHONPATH=../../../2000/10/swap python ../../../2000/10/swap/pim/toIcal.py test/mtg.rdf >test/mtg.ics

18:00:17 <DanC> loading...file:/home/connolly/w3ccvs/WWW/2002/12/cal/test/mtg.rdf

18:00:17 <DanC> exporting...

18:00:17 <DanC> perl ical2rdf.pl test/mtg.ics >test/mtg,.rdf

18:00:19 <DanC> python ../../../2000/10/swap/cwm.py test/graphCompare.n3 --think --purge-rules --with

18:00:21 <DanC> test/mtg.rdf test/mtg,.rdf >test/mtg-compare.n3

18:01:05 <DanC> D:still todo: make something bad happen if the test fails

18:01:06 <dc_rdfig> Added comment D9.

18:01:21 <mortenf> should the original files be overwritten?

18:01:33 <DanC> original?

18:01:45 <mortenf> like test/mtg.ics

18:01:49 <DanC> you mean test/mtg.ics? hmm... it's less "original" now

18:02:12 <DanC> it's not optimal to overwrite mtg.ics, no

18:03:44 <DanC> it's no wonder the round-trip failed for cal01... toIcal.py doesn't support half that stuff:

18:03:46 <DanC> loading...file:/home/connolly/w3ccvs/WWW/2002/12/cal/test/cal01.rdf

18:03:46 <DanC> exporting...

18:03:46 <DanC> @@skipping i:transp of [Church] = [i:opaque]

18:03:46 <DanC> @@skipping i:sequence of [Church] = [2]

18:03:46 <DanC> @@skipping i:class of [Church] = [i:private]

18:03:56 <mortenf> yeah...

18:04:14 <mortenf> well, that just means there's still something to do :)

18:04:48 <DanC> quite!

18:05:09 <DanC> hmm... graphCompare.n3 really belongs in swap/util ...

18:05:22 <DanC> and toIcal.py belongs as much in 2002/12/cal as in 2000/10/swap/pim

18:05:42 <mortenf> heh

18:12:00 <DanC> you can try this, btw: make TESTCASE=test/cal01 test

18:12:10 <DanC> did you get make test working on your end yet?

18:12:33 <mortenf> haven't gotten to that yet, was waiting for a non-desctructive version

18:13:17 <DanC> oh...

18:14:38 <DanC> heh...

18:14:39 <DanC> $ perl ical2rdf.pl test/20030423mtg.ics >test/20030423mtg.rdf

18:14:39 <DanC> X- property (WR-TIMEZONE) requires PRODID field at ical2rdf.pl line 210, <> line 4.

18:15:08 * mortenf smiles

18:15:42 * DanC commits Makefile,v 1.9 wich doesn't clobber original .ics files

18:17:15 <DanC> hmm... there is a prodid in that file... but it's after the first x- property

18:19:39 <mortenf> wee, make test seems to work, only real output is "@@skipping i:sequence of [icalendaring #rdfig meeting] = [3]"

18:20:12 <mortenf> btw, could you move the "PYTHONPATH=swap " up, instead of on the invokation line?

18:20:27 <mortenf> (or at least use $SWAP)

18:21:01 <DanC> it does use $(SWAP), no?

18:21:14 <mortenf> oh it does, my mistake

18:22:00 <mortenf> anyway, putting on the command line still exludes all other paths

18:22:36 <DanC> huh?

18:22:52 <DanC> oh, it replaces the PYTHONPATH envariable, yes

18:22:57 <mortenf> i guess it's not strictly important, but yeah

18:23:18 <DanC> hmm

18:23:35 <DanC> gnu make supports some sort of += thing for envariables, no?

18:23:45 <mortenf> hmm, don't know :(

18:26:08 <mortenf> how does one determine whether the compare failed?

18:27:44 <mortenf> ah, by the absence of gcomp:Success.

18:29:16 * mortenf tries to grok graphCompare.n3...

18:29:41 <DanC> only the 1st two rules really matter.

18:29:50 <DanC> the others are an attempt to provide more than one bit of output.

18:29:58 <DanC> i.e. to say what the difference is. but they're not working.

18:30:30 <DanC> well, the next 2 rules might be OK... but the last 2 are definitely hosed.

18:31:12 <DanC> "1".os:argv is the 1st arg... it's short for [ is os.argv of "1" ]

18:31:29 <DanC> os:baseAbsolute converts it to an absolute URI

18:31:56 <DanC> so ?F1 log:uri "1".os:argv.os:baseAbsolute binds ?F1 to the thing whose URI is the absolute form of the 1st command-line arg. capice?

18:32:11 <mortenf> yep, figured as much; i guess the understanding lies within log:...

18:32:57 <DanC> log:semantics relates something like <mtg.rdf> with the formula that is its contents.

18:33:00 <mortenf> but it looks darn smart!

18:33:16 <mortenf> mmm...

18:39:36 * sandro looks at http://www.w3.org/2002/12/cal/test/graphCompare.n3

18:40:37 * sandro trips on his stupid opera bug, where it renders { as [

18:40:51 <DanC> problem is, when :S matches a bnode on ?G1 log:includes { :S :P :O }.

18:41:22 <DanC> cwm forgets it's a bnode when doing ?G2 log:notIncludes { :S :P :O }.

18:42:05 <DanC> I'm not sure it's fixable.

18:42:20 <mortenf> sandro, what version, i don't seem to have that problem (unless it's your chat client, which i haven't tried)?

18:42:20 <DanC> I think it's a basic limitation of how quanitification works in N3... or in cwm, at least.

18:43:04 <mortenf> forgets?

18:43:08 <sandro> mortenf, 7.1.0 B1 linux RH 9.0. I assume it's a font interaction.

18:43:13 * DanC considers moving all the test targets into a test/Makefile

18:43:20 <mortenf> ah.

18:43:59 <mortenf> danc, that'd be cool, leaving me to perhaps comparing .ics files

18:44:39 <DanC> comparing .ics files? say again?

18:44:57 <mortenf> not necessary, but then again...

18:45:33 <sandro> So the bNodes in two different formulas don't get treated as the same bNodes. Yeah, sounds like a natural consequence of quantification scope.

18:45:45 <mortenf> nah, i guess you're right, but then my action is done :)

18:45:56 <sandro> Of course bNodes are always going to be a big problem in graph comparison results.

18:46:14 <DanC> I don't think it's a consequence of quantification scope.

18:46:57 <DanC> it's a consequence of quoting wierdness in log:includes

18:48:13 <DanC> I'm trying to find formulas { :S :P :O} that are rdf-simply-entailed from ?G1 but not from ?G2

18:48:51 <DanC> e.g. { _:something a :Vevent }

18:49:19 <mortenf> hmm, thought: couldn't bnodes be named (in part) by the property involved, that might reduce conflicts?

18:49:59 <DanC> "the property"? bnodes are, in general, connected by lots of properties.

18:50:33 <mortenf> yeah, will have to think more...

18:50:37 * DanC considers using log:rawType hacks...

18:52:52 <sandro> Yeah, sorry, my mental model of this level of cwm is too weak.

18:53:32 <DanC> is there an analog in prolog?

18:53:47 <DanC> I'd be surprised if anybody's mental model fo this level of cwm is strong.

18:53:59 <sandro> I think Tim's is....

18:54:04 <DanC> I don't.

18:55:33 <DanC> crud... there aren't distinct rawTypes for URI and bnode

18:55:58 * DanC stops chasing this rabbit...

18:56:30 <sandro> re prolog: I don't think there is an equivalent of log:includes. You can unify terms, and search in structures, and combine those... so you could write something like that. This would all be operating on structures which would not naturally be related to prolog facts/rules themselves, at all.

18:56:44 * DanC commits graphCompare.n3 v1.2 with rules for finding missing triples deleted.

18:57:27 <DanC> log:includes operates on structures in the same way. it's using quoting

18:58:13 <sandro> There's an excellent introduction the problem n3 {...} are trying to solve in Woolridge's Multi-Agent Systems book. He says there are two approaches to solving it: modal logic and metalogics. N3 uses metalogic, and happen to use the same syntax at the meta level.

18:59:16 <sandro> So the fact that we're not clear about what a bNode looks like when it's quoted... is the problem here?

18:59:42 <DanC> I think Felten and company use neither of those. hmm... well, perhaps what they use could be called a metalogic. but it's quite different from cwm's style of metalogic. cwm's looks like quoting, to me. Felten and company use typed lambda calculus. real HOL.

19:00:17 <DanC> here.. yes, something like that... quoted terms should retain their variable-ness, I think.

19:00:35 <DanC> this business of using log:forSome triples bites.

19:00:50 <sandro> cwm's logic (and meta-logic, etc) is roughly Horn.... that's probably the difference

19:01:06 <sandro> no kidding

19:01:23 * DanC returns to updating cwm docs...

19:53:14 <DanC> ok, cwm home page updated: http://www.w3.org/2000/10/swap/doc/cwm v 1.37 2003/07/09 19:51:58

19:56:34 * DanC noodles on relationship between features tested in http://www.w3.org/2000/10/swap/test/regression.n3 and those documented in http://www.w3.org/2000/10/swap/doc/cwm#Closed

20:27:32 <DanC> sigh... lots of redudnancy w.r.t. http://www.w3.org/2000/10/swap/

20:27:52 <DanC> and then there's the whole http://www.w3.org/2000/01/sw/ advanced development thing

20:31:54 <sandro> challenging stuff.

20:35:51 <DanC> hmm... the june tour http://www.w3.org/2003/03/semantic-tour.html doesn't appear on http://www.w3.org/2001/sw/

20:36:49 <DanC> and there's nothing about WWW2003 under 'news and events'

20:37:43 <DanC> drool... "ZOOMON and Opera deliver SVG on Symbian" -- http://www.w3.org/Graphics/SVG/Overview.htm8

21:39:51 <DanC> ok, updated http://www.w3.org/2000/01/sw/ v 1.105

21:40:08 <DanC> oh yeah! link to swad-eu has been pending for ages...

21:44:53 <DanC> v 1.105 has it now.

22:05:06 <libby> E:not discussed

22:05:07 <dc_rdfig> Added comment E1.

22:05:12 <libby> F:not discussed

22:05:12 <dc_rdfig> Added comment F1.

22:21:22 <mortenf> btw, DanC, thanks for earlier; i'll be looking more at the rdf-core tests timbl pointed to, trying to understand what's going on; would like to do a manifest style doc, and/or a Makefile for cal/test/, based on what you showed earlier.

22:21:52 <DanC> yeah, we'll eventually want a manifest thingy, most likely

22:27:05 <danbri> DanC, do you have an rdf vocab somewhere for talking about status of rdf terms, ie substructure of schemas?

22:27:14 * danbri just saved http://www.w3.org/2003/06/sw-vocab-status/ns

22:27:59 <danbri> (greatly cut down from my original ambitions, http://ilrt.org/discovery/2001/06/process/index.html)

22:28:12 <danbri> needed a way to say which bits of foaf were half-baked and which only quarter-baked...

22:29:19 <DanC> substructure of schemas? no, I don't think so

22:30:53 * danbri trying a thought experiment out for real, borrowing 'stable','testing','unstable' from debian; just need to come up with plausible account now of what it is for 'foaf:name' to be considered 'stable'

22:31:08 <danbri> I expect the answer to that will depend on who is making the claim...

22:31:13 * danbri gives up for the night...

22:36:15 <DanC> chuckle... "You're better off, statistically, just scaring the bejesus out of drivers so they creep along at 2 miles per hour around the hairpins." -- http://joelonsoftware.com/articles/BuildingCommunitieswithSo.html

22:46:40 <DanC> idea: googly-foaf: spider my foaf.rdf file, find my friends, find their weblogs; show me articles that more than one of them cite.

22:47:16 * DanC wanders out for the day...

22:47:52 <Cardinal> A twist on Technorati?

23:07:04 <Karl-MTL> Karl-MTL is now known as karl-out