Semantic Web Interest Group IRC Chat Logs for 2003-09-18

01:53:42 <tav> tav is now known as tav|offline

03:32:33 <jql-zzz> jql-zzz is now known as jql

03:37:32 <DanC> hmm... http://esw.w3.org/topic/ProofChecking isn't a very good name for the pattern I think we're really after there.

03:37:59 <DanC> the problem/tension that the pattern resolves is between expressivity and tractability

03:38:43 <DanC> the solution in the pattern is: give up on an effective decision procedure and just agree on a proof checking mechanism

03:39:20 <DanC> hm... the size of the trusted computing base is one motivation...

03:39:43 <DanC> AnarchicScalability a guiding principle

03:52:39 <DanC> SocraticCompleteness is related...

04:39:09 <_dreaminofjeanni> _dreaminofjeanni is now known as adamhill

04:51:57 <DanC> wow... I have almost used up the 2 hours I gave myself to work on this, and I'm not much further than when I started.

04:54:57 <DanC> hmm... SemanticWebBus

05:36:39 <DanC> ok, that'll have to do for tonight: http://esw.w3.org/topic/SemanticWebBus

06:32:39 <GNUPredator> GNUPredator is now known as mdupont

08:19:34 <[1]adamhill> [1]adamhill is now known as _adam_work

08:40:30 <arnarl> hi

08:51:41 <adamhill> adamhill is now known as _dreaminofjeanni

09:43:37 <dajobe> uh oh, udell

09:43:45 <danbri> url?

09:43:50 <dajobe> . http://www.xml.com/pub/a/2003/09/17/udell.html

09:44:14 <dajobe> Only just clicked it

09:45:10 <dajobe> btw, I feel the time is near for me to add html-scrapery (via xslt or something) to raptor

09:48:17 <dajobe> udell's piece seems to tiptoe towards what norm walsh has done

09:48:31 <dajobe> "The canonical version of my weblog is no longer the published one. Rather, it's an XML document-database the structure (but not content) of which is evolving "

09:52:17 <danbri> is he using xhtml as the main xml format?

09:52:21 * danbri skims

09:52:25 <dajobe> don't bother

09:52:27 <dajobe> it's puff

09:53:01 <danbri> why are you being so rude about it/him?

09:53:01 <dajobe> or skim to "I don't think the Semantic Web..."

09:53:06 <danbri> ah!

09:54:21 <dajobe> it wasn't being either much of a critique of SW or much except making a light point about recording things in a natural way that people understand. To which the answer is - "Doh!"

09:55:35 <danbri> anything that might nudge bloggers towards reusing XHTML (+class attribute conventions or extra markup) rather than reinvent their own document format (echopie etc) is imho good

09:56:08 * danbri spent a little while in rss circles trying to make point that, having XHTML and RDF as generic, well known XML formats, any new format needs to find a niche that neither addresses

09:56:41 <danbri> I don't think ther's much middleground between the two: for doc formats, use XHTML, for data formats, use RDF; for corner cases, embed one in the other and live a wellformed but invalid life...

09:57:25 <dajobe>http://www.xml.com/pub/a/2003/09/17/udell.html

09:57:26 <dc_rdfig> A: http://www.xml.com/pub/a/2003/09/17/udell.html from dajobe

09:57:37 <dajobe> A:|Language Instincts, Jon Udell (XML.com)

09:57:37 <dc_rdfig> Titled item A.

09:58:09 <dajobe> A:a little "semantic markup" (html & css) and thoughts on recording things for people

09:58:10 <dc_rdfig> Added comment A1.

09:58:34 <dajobe> only ""s since somebody will argue that XYZ is semantic markup

09:58:37 <dajobe> for all XYZ

09:59:18 * nmg grins

09:59:24 <nmg> save us from pedantry

09:59:29 <danbri> A:I'm defintitely a supporter of using XHTML + conventions (class= etc) rather than abandoning HTML for a brand new XML doc format...

09:59:29 <dc_rdfig> Added comment A2.

09:59:32 <dajobe> that'll never happen nmg

09:59:47 <dajobe> A:XHTML2?

09:59:47 <dc_rdfig> Added comment A3.

10:00:19 <danbri> A:This basic approach works quite well on the W3C homepage, which is XSLT-[http://www.w3.org/2000/08/w3c-synd/#|converted] into two flavours of RSS.

10:00:19 <dc_rdfig> Added comment A4.

10:00:32 * dajobe digs out a link

10:00:47 <dajobe> oh, I remember now, xml.org is crap, I can't

10:01:02 <dajobe> xml-dev archives died in July and they didn't notice/care

10:01:07 <danbri> A:imho, if you're promoting an XML format for doc/data syndication on the public Web (rss/echo etc) you need a really really good excuse for not using either of XHTML or RDF (or a mix of both).

10:01:07 <dc_rdfig> Added comment A5.

10:02:10 <dajobe> A:but take care. You need to consider [http://norman.walsh.name/2003/09/16/escmarkup|Escaped Markup: Still Harmful], Norm Walsh.

10:02:10 <dc_rdfig> Added comment A6.

10:02:41 <dajobe> I've not seen a nonameyet update recently. has it got a name?

10:03:29 <nmg> danbri: are there any examples of people using XHTML modules in their own markup languages?

10:04:32 <danbri> not seen much thats formally specified

10:04:44 <danbri> sometimes you see fragments of xhtml inside other xml formats

10:04:53 <danbri> not necc using the full Modularization framework tho

10:05:55 <nmg> shame...

10:06:07 <danbri> but i haven't really gone looking that hard, to be honest

10:20:35 <arnarl> arnarl is now known as arnarl|lunch

10:21:31 <Mutiny-> Mutiny- is now known as Mutiny

10:30:12 <sandro> Hrm. esw.w3.org (aka swada.w3.org) has an 5.00*n random delay in responding to HTTP GETs. WTF? (I see lots of n=1, and a few n=0 or n=2. No n>2 yet.) n=0 from localhost or from a very nearby machine.

10:30:54 <sandro> the delay occurs before the entry appears in the access log.

10:31:26 <dajobe> sandro: can I bug you a sec?

10:31:37 <sandro> Sure, what's up?

10:31:45 <dajobe> cvs server: failed to create lock directory for `/sources/public/2000/10/swap' (/sources/public/2000/10/swap/#cvs.lock): Permission denied

10:32:12 <dajobe> I was trying to get some/latest n3 files from the anon cvs area

10:32:31 <sandro> any idea when this problem might have started?

10:32:37 <dajobe> no

10:33:31 <sandro> lessee....

10:33:43 * dajobe tries mailing web-human@

10:57:26 <sandro> dajobe, has the problem gone away? can you give me a recipe to reproduce it?

10:58:10 <dajobe> (I did email it to web-human) hold on

10:58:12 <sandro> re: 5.0 * n second delay -- it seems to be on my local machine!

10:58:25 <dajobe> $ cat CVS/Root

10:58:25 <dajobe> :pserver:anonymous@dev.w3.org:/sources/public

10:58:29 <dajobe> $ cat CVS/Repository

10:58:29 <dajobe> 2000/10/swap

10:58:39 <dajobe> $ cvs update -APd

10:58:43 <dajobe> ... then it happens

10:58:59 <dajobe> you can probably make a fake CVS dir/space and try this

10:59:11 <dajobe> touch CVS/Entries probably also needed

11:01:37 <sandro> can you get that with a one-line command, cvs -d something checkout something ? (I get other error messages, but my cvs-fu is weak)

11:04:34 <dajobe> yes

11:04:37 <dajobe> cvs -d :pserver:anonymous@dev.w3.org:/sources/public checkout 2000/10/swap

11:05:50 <arnarl|lunch> arnarl|lunch is now known as arnarl

11:05:54 <sandro> Ah, I see now. Hrm. Thanks.

11:17:57 <sandro> re: 5.0 * n second delay -- solved. it's my local DNS resolver. arg. /etc/hosts is a wonderful thing. :-)

12:24:03 <bitsko|dinner> bitsko|dinner is now known as bitsko

12:29:54 <libby>http://news.bbc.co.uk/1/hi/technology/3111004.stm

12:29:54 <dc_rdfig> B: http://news.bbc.co.uk/1/hi/technology/3111004.stm from libby

12:30:13 <libby> B:|glasses/camera combo from HP labs

12:30:13 <dc_rdfig> Titled item B.

12:30:33 <libby> B:"To tackle image overload, the HP system captures information about images, called metadata, too."

12:30:33 <dc_rdfig> Added comment B1.

12:42:41 <ear1grey> B: some years back, at a large music festival in the westcountry, i was taken to wondering how many photographs I might have inadvertently wandered into, given that there were 100,000+ people at the event. At the time I thought how useful it would be to combing a GPS log of my locations with gps, direction and viewfinder information of each image - thus my holiday would not just be captured by my photos, but by the 100K other p

12:42:42 <dc_rdfig> Added comment B2.

12:44:46 * danbri nods definitely to ear1grey

14:44:02 * DanC wonders what happened to my xchat process

15:15:29 * DanC noodles on swap/cwm/n3 support "office hours" where we'd offer real-time support for cwm

15:16:13 <DanC> ... priority given to folks that had mailed problem reports in advance, but nothing too strict...

15:24:47 <sbp> hehe: "Thing Mode is a GNU Emacs (pseudo-)mode which is sort of vaguely like a poor man's hypertext system: it finds textual things in your buffers and does stuff to them." - http://www.lib.uchicago.edu/keith/software/thing/thing-mode.html

15:27:44 <DanC> keith who? not the usability guy...

15:28:13 <sbp> Keith Waclena. apprently not the usability guy, no

15:34:58 <DanC> heh! "Next update: Never. By Keith Instone." -- http://usableweb.com/

15:41:49 * DanC hears the cheers "norm" thing in his head again

15:42:03 <ndw> Hey, DanC, I have a cwm question and you're my victim of choice.

15:42:21 * DanC can't find a place to hide

15:42:26 <ndw> Remember how you told me about ( a b c d ) as a way of making a list

15:42:36 <ndw> This generates daml:first/daml:rest pairs, right?

15:42:46 * danbri grins re "* DanC noodles on swap/cwm/n3 support "office hours" where we'd offer real-time support for cwm" from 1/2 hour ago

15:42:46 <DanC> er... maybe rdf:first/rdf:rest now

15:42:50 <danbri> were you reading the logs, norm?

15:42:54 <ndw> Uhm, yeah, exactly.

15:42:56 <ndw> What's up with that?

15:43:03 <ndw> No, danbri, did it come up earlier today?

15:43:11 <danbri> just re cwm support

15:43:31 <ndw> Jena barfs on rdf:first/rdf:last, fwiw. Did the WG change something, or is cwm buggy, or is cwm anticipating a change, or what?

15:43:32 <danbri> daml lists were a requiest webont wg made of rdfcore wg

15:43:41 * DanC hunts for docs on recent big cwm/n3 changes...

15:43:43 <danbri> which version of jena? v2 should do it

15:43:58 <ndw> Oh, probably still v1. I have v2 but I haven't tried to run my big stuff through it yet.

15:44:12 <ndw> i figure six things are going to break and it'll take me the better part of a day to get it all working again.

15:44:47 * DanC opens http://esw.w3.org/topic/CwmTips window...

15:45:20 <DanC> ugh... two evolution alarm daemons running.

15:45:35 <DanC> sbp, do you know where timbl put that change file?

15:46:18 <DanC> "Re-engineered in places - See doc changes.html" -- revision 1.135 changelog...

15:46:56 <danbri> Norm, Leigh reports in http://www.ldodds.com/blog/archives/000071.html "Just a quick note to say that I upgraded the FOAF-a-Matic Mark 2 to use Jena 2.0.0-beta1 and found the migration pretty painless. I'd read Shelley's Jena week articles as preparation so had a good idea about what I needed to do."

15:47:12 * danbri ponders having a definitiveish list of RDF IG member's weblog URIs

15:47:20 <danbri> missing ingredient: what is an RDF IG member

15:47:38 <danbri> we don't really have a prose definition, let alone an RDF/OWL one...

15:47:54 <danbri> "member of at least one w3c mailing list under the rdfig umbrella"?

15:48:05 <DanC> "List processing improves - see doc/changes.html" -- http://dev.w3.org/cvsweb/2000/10/swap/

15:48:24 <dajobe> is there a list of the umbrella contents ?

15:48:28 <DanC> aha! http://www.w3.org/2000/10/swap/doc/changes

15:49:11 * DanC really wants to try advogato style trust metrics in the W3C community for stuff like "developer of WAI-happy web sites" and such

15:49:26 <ndw> Thanks for the pointer danbri, maybe I'll try jena2 tomorrow, but I'm just trying to run a publish now. sigh.

15:49:58 <DanC> ndw, do you see "Lists are handled differently internally. ..." in http://www.w3.org/2000/10/swap/doc/changes ?

15:50:00 <danbri> dave, its a bit fuzzy edged. www-rdf-interest, www-rdf-logic, www-rdf-rules, www-rdf-calendar I'd say; others nearby, public-rdf-ruby, also www-rdf-comments, www-annotation...

15:50:19 * danbri nods re advogato, not sure what a gentle way in would be

15:50:24 <dajobe> the rdf-thes

15:50:36 <dajobe> I'd forgotten that one

15:50:41 <danbri> public-esw-thes yeah, though we've hardly started using it

15:50:52 <danbri> is one of those swad/ig bridgy things, in theory anyway

15:50:59 <ndw> Yes, Danc, I see it.

15:51:04 <dajobe> it's a moving bridge :)

15:51:04 * DanC offers http://esw.w3.org/topic/MailingLists as the community's list of lists

15:51:19 <DanC> does it answer your question, norm?

15:51:43 <DanC> (I'm not sure I understand your question nor timbl's answer, but I'm hopeful I can just play matchmaker)

15:52:11 <ndw> Yes. It's not a bug. I'll work around it today with a quick XSLT hack and see if Jena2 parses it as soon as I can.

15:52:39 <ndw> I thought maybe rdf: was a typo for daml: but if it isn't, it's just life being hard again. Nothing new there :-)

15:54:40 * DanC beams with satisfaction at addressing norm's issue...

15:54:51 * DanC wonders about an affero account for such occasions... http://www.affero.com/

16:01:37 <tav|offline> tav|offline is now known as tav

16:01:37 <_joshua> people use affero?

16:03:47 <tav> tav is now known as tav|offline

16:03:54 <tav|offline> tav|offline is now known as tav

16:05:45 * DanC -> webont telcon

18:01:58 <mdupont> peace sbp

18:06:04 <sbp> hey mdupont

18:06:21 <sbp> mmm... random peace. likewise

18:08:32 <mdupont> what you been up to? published anything lately?

18:10:34 <sbp> I've been working on... too many thanks. publications: not all that many, oddly. miscoranda.com generally tracks what I do publish...

18:10:55 <sbp> s/thanks/things/

18:11:30 * sbp watches as Aaron corrects one of the worst typos in decades

18:12:41 <sandro> where, sbp, where?!

18:13:06 <mdupont> name suggestion : Python Resource Description Format : PRDF ?

18:13:55 <sbp> sandro: his blog, Gay Marraige thing, first "shouldn't" was "should" previously...

18:14:05 <sbp> mdupont: heh, I think someone actually already suggested that to me

18:14:08 * sbp checks

18:14:18 <sbp> ah, no, PRF. close

18:14:34 <sbp> hmm. from prdf import graph... looks quite good

18:14:39 <sbp> s/g/G/

18:15:33 * sandro reminisces over typoing "pat hayes" as "pay hates" back when he was scared of him. :-)

18:15:49 <sbp> hehheh

18:18:40 <sandro> Wow, Aaron sure is the optimist. Hardwick stood for, what, 17 years and just barely got overturned.

18:20:26 <sbp> yeah. though he doesn't give a timescale... does sound like he's expecting it to all happen tomorrow, though

18:21:03 * sandro watches the millenium clock tick (virtually speaking)

18:25:55 <sbp> import time, datetime

18:25:55 <sbp> while 1:

18:25:55 <sbp> time.sleep(1)

18:25:55 <sbp> print datetime.datetime(year=3000, month=1, day=1) - datetime.datetime.now()

18:26:08 <sbp> [[[

18:26:08 <sbp> 363886 days, 4:34:31.610000

18:26:09 <sbp> 363886 days, 4:34:30.570000

18:26:09 <sbp> 363886 days, 4:34:29.530000

18:26:10 <sbp> ...]

18:26:12 <sbp> ]]]

18:38:21 * sandro hasn't had an excuse to play with the new datetime yet. yay

18:38:52 <sandro> "Naive datetime objects are easy to understand and to work with, at the cost of ignoring some aspects of reality."

18:40:00 <darobin> ooh python has new datetime stuff as well? funny, the perl world started a new and nice one some months ago too

18:40:14 <sandro> yeah, new in 2.3

18:40:18 * darobin still thinks there's insufficient technology crossover between those two worlds

18:40:38 <sbp> yeah. it's come in handy for me already

18:40:56 <danbri> is Parrot supposed to heal that rift?

18:41:19 <darobin> help, but it won't be enough to heal

18:41:35 <_adam_work> _adam_work is now known as adamhill

18:42:18 <darobin> it'll help if Dan Sugalski manages to make python running on it at least as fast as python by oscon 2004, in which case guido will very likely consider it (after having received pies in the face -- that's the bet)

18:42:32 * sandro prefers not to think about Perl at all any more. It's not surprisingly like (I imagine it would be to be) thinking about an ex-, after leaving 'em.

18:42:51 <darobin> but still, a python guy implementing smart stuff on python/parrot that's usable from perl won't see it used unless the communities talk to each other

18:44:03 * darobin thinks of python as that-girl-he-almost-went-out-with-but-didn't-for-no-obvious-reason

18:57:35 * DanC wishes the python runtime would switch to being more COM-like in the interest of cross-language interop

19:02:52 <darobin> well feel free to contact your local python implementer to ask him about parrot :)

19:03:11 <dajobe> there was a session at the open source conference on the .NET CLR stuff with py, perl, ruby etc.

19:03:19 <dajobe> turns out it doesn't do dynamic languages well

19:03:31 <dajobe> so I guess since CLR has some COM background, it'd need more work

19:12:23 <ndw> Everytime I look at another language, Perl seduces me back.

19:15:08 <darobin> hehe

19:15:48 <darobin> dajobe: dan sugalski has a few good entries in his blog about why perl/py/ruby/etc don't work well on .NET CLR or JVM

19:16:20 <darobin> basically iirc it boils down to "if you do closures, a stack-based VM will be terribly terribly slow"

19:16:42 <dajobe> ah

19:16:55 <dajobe> you saw how java runs very well on a .NET runtime such as Mono

19:17:10 <dajobe> I saw Miguel & Nat demo Eclipse running on Mono

19:17:20 <dajobe> which is cute

19:17:40 <darobin>http://www.sidhe.org/~dan/blog/

19:17:41 <dc_rdfig> C: http://www.sidhe.org/~dan/blog/ from darobin

19:17:57 <darobin> damn, I always forget about dc_rdfig, sorry

19:18:07 <dajobe> well, give it another url C:=newurl

19:18:38 <darobin> yes, but I wasn't intending to log any url :)

19:18:47 <dajobe> ok

19:18:56 <dajobe>http://ilrt.org/discovery/chatlogs/rdfig/

19:18:56 <dc_rdfig> D: http://ilrt.org/discovery/chatlogs/rdfig/ from dajobe

19:19:11 <dajobe> D:|RDFIG chatlogs, new style, dated indexes

19:19:11 <dc_rdfig> Titled item D.

19:19:22 <darobin> there's no reason why Java and C# wouldn't work just as well on pretty much the same VM

19:19:24 <dajobe> D:ok, the style isn't that new

19:19:24 <dc_rdfig> Added comment D1.

19:19:41 <dajobe> C:=http://www.w3.org/2003/08/owl-systems/test-results-out

19:19:41 <dc_rdfig> Replaced URL of C.

19:19:48 <darobin> thanks :)

19:19:49 <dajobe> C:|OWL Test Results

19:19:49 <dc_rdfig> Titled item C.

19:20:00 * dajobe eats

20:06:17 <golbeck_> golbeck_ is now known as golbeck

20:14:30 <adamhill> adamhill is now known as _adam_espresso

20:33:42 <Arnia22> Arnia22 is now known as Arnia

20:34:55 * DanC feels obliged to attack http://esw.w3.org/topic/WantedPages

20:35:08 * DanC needs to finish telcon minutes, though...

20:42:49 * danbri fixes 3, loads more left...

20:42:59 <danbri> will do some more, a fair few are mine

20:47:46 <_adam_espresso> _adam_espresso is now known as adamhill