Semantic Web Interest Group IRC Chat Logs for 2003-04-02

00:10:36 <nym> hey AaronSw

00:10:49 <AaronSw> hey

00:11:26 * Morbus nods head in Aaron's direction.

00:12:39 <AaronSw> hi

00:13:40 <AaronSw>http://advogato.org/person/Bram/diary.html?start=66

00:13:41 <dc_rdfig> A: http://advogato.org/person/Bram/diary.html?start=66 from AaronSw

00:13:58 <AaronSw> A:|W3C, IETF Announce New Policies

00:13:59 <dc_rdfig> Titled item A.

00:14:08 <AaronSw> A:"Today, the W3C announced a new initiative to emphasize ease of implementation in specification design. 'Encouraging adoption is the most important part of making protocols useful for the future internet' said Tim Berners-Lee."

00:14:08 <dc_rdfig> Added comment A1.

00:22:27 <AaronSw>http://www.mozillazine.org/weblogs/hyatt/archives/2003_04.html#002969

00:22:27 <dc_rdfig> B: http://www.mozillazine.org/weblogs/hyatt/archives/2003_04.html#002969 from AaronSw

00:22:35 <AaronSw> B:|Safari to Drop Table Support

00:22:35 <dc_rdfig> Titled item B.

00:22:48 <AaronSw> B:'Unfortunately word also leaked to the TAG (Table Advocacy Group), and they were outraged. Jim Lerners-Lee had this to say. "I hate all this newfangled crap. XHTML2, XFORMS, CSS3. It's all so complicated. Me, I like tables. Good old-fashioned tables. Especially when used with the blink tag. Safari is obviously taking a dangerous stance by dropping support for tables, since everyone knows they are the backbone of the modern Web

00:22:48 <dc_rdfig> Added comment B1.

00:23:56 <DanC> chuckle

00:32:23 <nym> woo

00:32:31 <nym> danbri helped me a lot

00:32:35 <nym> yay!

01:35:25 <nym> i'm trying to polish off http://igargoyle.com/rss/1.0/modules/spatial/ http://igargoyle.com/rss/1.0/modules/img/ and http://igargoyle.com/rss/1.0/modules/event/

01:35:42 <nym> if anyone can peer review them for me, i'd appreciate it

01:35:56 <nym> i'm going to post to rss-dev soon you see

01:42:26 <jhendler_>http://www.w3.org/News/2003#item65

01:42:27 <dc_rdfig> C: http://www.w3.org/News/2003#item65 from jhendler_

01:42:50 <jhendler_> C:| Web Ont WG releases five Last Call Working Drafts

01:42:50 <dc_rdfig> Titled item C.

01:43:10 <jhendler_> C: this is the W3C announcement of the publications

01:43:10 <dc_rdfig> Added comment C1.

01:43:32 <jhendler_> C: includes pointers to the Last Call drafts

01:43:32 <dc_rdfig> Added comment C2.

01:43:33 <jordan> hi jhendler_

01:43:43 <golbeck> hi jim!

01:43:59 <jhendler_> C: We would greatly welcome comments from the RDF IG community

01:44:00 <dc_rdfig> Added comment C3.

01:44:06 <jhendler_> hi Jen, jordan

02:05:59 * jhendler_ waves at timbl

02:42:28 <golbeck> golbeck is now known as golbeck_onadate

04:10:33 <oc|gone> -nick oc

04:10:50 <oc|gone> oc|gone is now known as oc

04:15:20 * DanC wanders by...

04:17:06 * timb| waves vaguely

04:17:17 <dnm> timb|?

04:17:22 <dnm> As in, timboy?

04:17:33 <timb|> dnm?

04:18:04 <dnm> Hrm. Might be a case of mistaken identity.

04:18:19 <dnm> I also thought I was in a different channel than this one. Woops.

04:18:30 <timb|> :-)

04:18:48 <timb|> no sweat

04:18:59 * DanC wonders about iPhoto and RDF image descriptions

04:19:17 <DanC> I haven't found iphoto data in plist.xml format.

04:19:21 * timb| checked in version of cwm with unicode things seeming to be fixed.

04:19:26 <DanC> but I gather there's an applescript interface

04:19:53 * timb| tries to remember what he had in plist format

04:19:59 * DanC feels somewhat obliged to cvs update swap, but somewhat nervous to do so

04:20:04 * timb| ... music i think

04:20:08 <DanC> probably music, yes

04:20:28 <timb|> Nervousness is in order...

04:20:34 * DanC starts with cvs -n -q update

04:21:11 <DanC> wow... many screenfuls. I'm very out of date.

04:21:17 <timb|> ... I have as threatened made the time and math incompatible, now generating numbres where appropriate. Use maths and times for old functionality, lovingly restored by our craftsmen

04:21:59 <timb|> So just slip that "s" into your namespaces if you are concerned. But you may find it doesn't bother you.

04:22:00 <DanC> ew... so you broke all the rules files that used math and time. bummer.

04:22:24 <DanC> doesn't bother? so maybe they didn't break as badly as I fear?

04:22:49 <timb|> Well, if something is generated as an int, it will still function as a string.

04:22:53 <DanC> cvs server: conflicts found in dbork/SqlDB.py

04:23:17 <timb|> mea non cupla

04:23:29 <timb|> s/pl/lp

04:23:55 <DanC> btw... if I make tested, it's not gonna cvs commit anything, is it?

04:24:05 <DanC> $ make tested

04:24:05 <DanC> make: *** No rule to make target `maths.rdf', needed by `package'. Stop.

04:24:09 <timb|> don't *think* so!

04:24:48 <timb|> Ah.... must have not added maths.n3 times.n3

04:25:25 * timb| checks them im

04:25:57 <DanC> make: *** No rule to make target `cwm_maths.html', needed by `package'. Stop.

04:27:34 <timb|> Sorry, cwm update cwm_maths.py cwm_times.py now

04:27:49 <timb|> Sorry, cvs update cwm_maths.py cwm_times.py now

04:28:14 <DanC> running regression tests now...

04:28:29 <DanC> crapped out at 9) Parse RDF/XML nodeID

04:28:40 <DanC> IOError: [Errno 2] No such file or directory: '/home/connolly/w3ccvs/WWW/2000/10/swap/test/nodeID/ex1.rdf'

04:28:52 * DanC remembers -d

04:29:41 <DanC> 12) N3 string, qname and IRI with utf-8 non-ascii characters

04:29:42 <DanC> passed

04:29:43 <jordan> haha timmy from #lisp != tim berners-lee.. they're quite different

04:30:21 <timb|> indeed ... nobody calls timbl timmy .. twice ;-)

04:30:24 <DanC> 40 tests passed...

04:30:53 * DanC didn't know about the timmy button... tucks that nugget away... 1/2; -)

04:31:25 <DanC> diff: ref/timet1s.n3: No such file or directory

04:31:31 <DanC> re 62) basic ISo time handling functions - stringy

04:31:42 <DanC> did you see the fencepost bug I reported re dayOfWeek, timbl?

04:32:18 <timb|> oops - added that

04:32:43 <timb|> I saw that bug fly by into some useful bugbin.

04:33:07 <timb|> (added ref/timet1s.n3 r1.1)

04:33:16 <timb|> dayOfWeek.

04:33:39 <timb|> I seem to remember that you found it one day out?

04:34:06 * timb| considers taking achance and shifting it one way of the other by 1 ;-)

04:34:23 * DanC adds pointer to www-archive+n3bugs archive to http://esw.w3.org/topic/CwmTips

04:34:49 <timb|> It isn't a very good tracker.

04:34:56 <DanC> I think the community self-support (via wiki) is going to work better than us trying to keep docs up to date.

04:34:57 <timb|> I have a mailbox where I keep copies.

04:35:03 <DanC> let's do a little bit of docs and defer the rest to the wiki.

04:35:27 <timb|> What about - bug reports? - TODO?

04:35:31 <DanC> cwm time:dayOfWeek fencepost bug http://www.w3.org/2002/02/mid/1047678399.25643.271.camel@dirk.dm93.org;list=www-archive

04:36:07 * DanC isn't sure where to draw the line... hmm...

04:36:48 <DanC> btw... did you realize time:dayOfWeek only works back to 1970?

04:36:54 <timb|> Did youget theough the tests now?

04:36:55 <DanC> I tried to compute easer 1776 and lost.

04:37:18 <timb|> No, I didn't know that . It should give an error before 1970

04:37:21 <DanC> tests... yes, focus... lemme grab that ref/ thingy...

04:37:24 <timb|> ... or just not calculate

04:37:54 <DanC> it craps out with a python exception if you try before 1970. just not calculating would be "correct" but boy is debugging N3 hard.

04:37:55 <timb|> It was useful to get the feedback of missing files!

04:38:20 <timb|> So just now I just ahve to add 1

04:39:01 * DanC noodles on a good name for a wiki topic on why debugging N3 is naturally hard...

04:39:36 <DanC> minimally constraining isn't conducive to early detection of errors.

04:40:15 <DanC> crapped out at 62 again...

04:40:28 <DanC> IOError: [Errno 2] No such file or directory: '/home/connolly/w3ccvs/WWW/2000/10/swap/test/time/t1s.n3

04:40:37 <DanC> would you please do a cvs -n -q update

04:40:46 <DanC> and make sure the ? files are really not supposed to be checked in?

04:40:47 <timb|> -q?

04:40:55 <DanC> -q is a bit less verbose

04:41:28 <DanC> -n is (1) like make, i.e. don't actually do anything, and (2) doesn't grab any locks

04:41:55 <timb|> -q is why I had found it difficult to work thu the list - make sit much easier

04:41:56 <DanC> do you review your changes when you check in? add cvs -n -q update to the ritual, pls.

04:46:10 * timb| cleans up and checks in various things, includinhg time/t1s.n3

04:46:57 * DanC noodles on IrcBasedPairProgramming

04:47:56 <timb|> I am going to wander off when you get the regresssion test through.

04:48:07 <timb|> Oh, I should explain the rdfcore test too

04:48:36 <timb|> cd test; python retest.py -f rdfcore-tests.n3 -f testmeta.n3 -i

04:48:42 <timb|> owtte

04:49:23 <DanC> rdfcore-tests.n3 ... ew.. you copied the manifest? why not do it with log:semantics and such?

04:49:27 <timb|> That checks against 108 positive parser tests which do not involve reification or parseType=Literal

04:49:36 <DanC> I guess I should have checked in my work on RDF Core tests.

04:49:51 <timb|> I only copeid it so that folks could run it in the cvs distribution.

04:50:03 <timb|> It is exactly an n3 copy of the manifest.

04:50:03 <golbeck_onadate> golbeck_onadate is now known as golb

04:50:06 <golb> golb is now known as golbeck

04:50:31 <timb|> I guess one could put a make dependency out of the SWAP source tree.

04:50:32 <DanC> exact copy... until the original changes

04:51:03 * DanC still sees no test/ref/t1s.n3

04:51:04 <timb|> rdfcore-tests.n3: ../../rdf-tests/rdfcore/Manifest.rdf

04:51:13 <timb|> .... $P $C --rdf ../../rdf-tests/rdfcore/Manifest.rdf --n3 > $@

04:51:36 * DanC wonders why convert it to n3

04:51:44 <timb|> t1s is in test/time

04:52:07 <DanC> how do I start 'make tested' at 40 or so?

04:52:08 <timb|> n3: If I am going to make a copy, i might as well make a readable one

04:52:21 <timb|> cd test;

04:52:32 <DanC> the existing manifest is readable to lots of folks

04:52:39 <timb|> python retest.py -n -f regression.n3 -s40

04:53:06 <timb|> And they have the existing Manifest to read. Yes, I could have made it RDF/XML just as well.

04:55:06 <DanC> Test worked, now can release

04:55:42 <timb|> Good.

04:55:58 <timb|> I must clean up my long list of "?"

04:56:31 <DanC> does the list that goes "people, documents, ..." have a clear end? what are the other members?

04:56:56 <timb|> You mean "people, documents, times, places, dollars....?'

04:57:19 <DanC> I was hoping for something without an elipsis

04:58:23 <DanC> other candidates: meetings, events, photos

04:58:33 <DanC> but I suppose those are derivatives, in a way

04:59:10 <DanC> is that list ( "people, documents, times, places, dollars....?") explained in context anywhere that you can think of? some slides or an essay?

04:59:30 * DanC is gonna put it in PersonalInformationDisaster or SemanticWeb1.0 or something

04:59:36 <timb|> as in http://www.w3.org/DesignIssues/diagrams/SemWebAppMetro.png ?

05:00:23 <timb|> The point of the slide is that there are always elipses .. taht is why it is a web not a centralized mega-app.

05:00:32 <DanC> yes, that's pretty close to what I'm after.

05:01:22 <timb|> Things like "parts" can have a local definition, but "time" connects tono end of things.

05:01:44 <DanC> yes, there are always elipses... but I'm hoping for some list of target vocabularies/tools that will form a critical mass.

05:01:50 <DanC> I guess we'll know it when we get there.

05:02:26 <timb|> Each demo which crosses app boundaries in a major way will give somebody that ah-ha moment.

05:02:55 * DanC resists the urge to coin the NormWalshMoment topic...

05:03:05 <timb|> ;-)

05:03:15 <DanC> SemanticWebKillerAppNot

05:03:43 * timb| wanders off... till tomorrow

05:03:45 <DanC> BLURB: cwm is I18N happy (er.. happier, at least)

05:03:45 <dc_rdfig> D: cwm is I18N happy (er.. happier, at least) from DanC

05:04:09 <DanC> D:timbl fixed a bunch of stuff; DanC reviewed it (well, got the tests running)

05:04:10 <dc_rdfig> Added comment D1.

05:04:18 <DanC> logger, pointer?

05:04:30 <DanC> logger_1, pointer?

05:04:30 <DanC> See http://ilrt.org/discovery/chatlogs/rdfig/2003-04-02#T05-04-30

05:04:42 <DanC> D:see [discussion|http://ilrt.org/discovery/chatlogs/rdfig/2003-04-02#T05-04-30]

05:04:43 <dc_rdfig> Added comment D2.

06:08:10 * DanC suffers galeon2 crash.

06:08:15 <DanC> sigh... had an edit buffer open

06:08:25 <sandro> arg....

06:08:57 <sandro> I wonder if there's a browser trick which hands the edit-buffer over to emacs. That would be nice.

06:09:25 <DanC> I think I've seen it done... with nasty mime type kludgery, though

06:44:13 <danbri>http://esw.w3.org/topic/RestaurantRecommendation

06:44:13 <dc_rdfig> E: http://esw.w3.org/topic/RestaurantRecommendation from danbri

06:44:25 <danbri> E:|RestaurantRecommendation

06:44:25 <dc_rdfig> Titled item E.

06:44:34 <danbri> E:"Many people see Restaurant Recommendation as a classic SemanticWeb use case - building community around food!"

06:44:34 <dc_rdfig> Added comment E1.

06:45:25 <danbri> E:Contributors welcome (it's a Wiki page). Looking for relevant links, data sources, usage scenarios, design choices etc...

06:45:25 <dc_rdfig> Added comment E2.

06:46:26 * DanC noodles on http://www.w3.org/2000/01/sw/ , how to update it to give a better view of what's up

06:47:37 <danbri> morning DanC!

06:48:10 <DanC> hi; it's late, here

06:48:34 * danbri grins re 24x7 hack'n'chat (the extra hour due to summertime change this week, and me being up early, closes the 24x7 loop here)

06:48:47 * DanC would much rather "based on the RDF standards" were "based on URI, XML, and RDF specs" or some such, in http://www.w3.org/2001/sw/

06:49:28 <DanC> and Overview: should be Contents:

06:49:38 <DanC> and the proxy activity statement is wierd.

06:50:30 <danbri> "by the W3C, in collaboration with a large number of researchers and industrial partners." is a bit redundant

06:50:58 <danbri> does it mean non-W3C-member industrial partners, for eg? Who? W3C is a bunch of industrial / research partners

06:51:45 <DanC> yes... redundant.

06:52:04 <DanC> bugs me in much the same way that "the web and ftp" bugs me. the web includes ftp.

06:52:51 <danbri> yup

07:00:19 <DanC> interesting... the http://www.w3.org/2000/01/sw/#L3028 annotation/collaboration section grew a bunch of nifty text that I didn't notice.

07:02:47 <DanC> I think I'll try to revive my Amaya installation...

07:04:06 <DanC> ... reagle told me the gnome/debian issues are resolved now...

07:05:30 <DanC> phpht. I still get Gdk-CRITICAL **: file gdkfont.c: line 397 (gdk_text_width_wc): assertion `font != NULL' failed.

07:06:29 <DanC> gtkfontsel works fine

07:07:26 <DanC> I think I'll obsolete http://www.w3.org/2000/01/sw/#Presentations in favor of http://www.w3.org/2001/sw/#events

07:10:31 * danbri has hunch there's an rss feed from that, somewhere...

08:39:30 <grault> grault is now known as earle

08:40:24 <earle> morning

08:46:22 <earle> anybody about? I'd like to register myself on esw, but I'm not sure how.

08:56:00 <libby> register yourself earle? on the wiki?

08:58:33 <earle> hey libby. yeah, in order not to be AnonymousUser... I was going to add something to the mention of Grubstreet I just saw on RestaurantRecommendation.

09:02:27 <libby> actually, I dont know how....:)

09:04:02 <libby> good idea tho..

09:05:17 * earle nods, can't see an obvious edit preferences link

09:05:45 <earle> or is there more WikiHonour in ContributingAnonymouslyForTheGoodOfThePublic? ;)

09:05:59 <libby> ? http://esw.w3.org/topic/HelpOnUserPreferences? but it doesnt say much. page it points to " how to make yourself known to the wiki, and adapt default behaviour to your taste"

09:06:04 <libby> heh

09:06:17 <earle> Ah! UserPreferences

09:06:36 <libby> ah :)

09:08:34 <earle> I'm in. </cheesy-movie-about-hackers>

09:08:56 <libby> heheh

09:13:03 <earle> okay, done.

09:29:14 <earle> morning dajobe

09:31:06 <dajobe> hi

10:43:37 <arnarl> hi

14:31:34 <JibberJim> Did no-one mention http://www.w3.org/TR/photo-rdf/ to nym who seemed to be recreating a lot of the technical schema there? in his RSS photo module. Also MattB's photo RSS stuff is probably worth a read.

14:32:16 <libby> yeah, depicts or depiction at least

14:33:33 <JibberJim> also of course rdfPic and other image description generating stuff, although I'm not sure a lot of the description vocab other than foaf:depicts/depiction is any less experimental than his work.

15:04:05 <Wack> hmm, you could get a lot of technical metadata from jpeg exif headers (practically all digital cameras add them)

15:05:13 <libby> true. I think Gerald oskoboiny (sp) has an RDf schema for that

15:06:29 <libby> here's a link http://impressive.net/people/gerald/2000/09/photo.html

15:06:38 * libby adds it to http://esw.w3.org/topic/ImageDescription

15:08:32 <libby> among others, I'm sure

15:15:22 <libby>http://esw.w3.org/topic/EswWp3

15:15:24 <dc_rdfig> F: http://esw.w3.org/topic/EswWp3 from libby

15:15:49 <libby> F:|SWAD-Europe workpackage 3 - marketing ideas

15:15:49 <dc_rdfig> Titled item F.

15:16:50 <libby> F:specifically, I was wondering if people would mind looking at [examples of Geographical location RDF|http://www.w3.org/2001/sw/Europe/200303/geo/]

15:16:51 <dc_rdfig> Added comment F1.

15:17:23 <libby> F:if I've missed something out or some something dumb, I'd like to know :)

15:17:23 <dc_rdfig> Added comment F2.

15:21:27 <libby> F:not sure if negative longs make sense for example...

15:21:27 <dc_rdfig> Added comment F3.

15:29:50 * DanC sees rdf calendar on his schedule for 11am today, recalls something about 1500Z vs 1700Z...

15:30:16 <libby> danc, it's next week right?

15:31:27 <libby> - http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Mar/0043.html

15:31:42 <libby> every other week, starting last week...

15:31:45 * libby hopes...

15:32:36 <DanC> gee, do you think we could use RDF to avoid some of this fiddling with our calendars? somebody should really do something about that ;-)

15:33:32 <DanC> ok, my WearableGizmo thought it was every week.

15:33:50 <DanC> .time CST

15:33:50 <datum> Wed, 02 Apr 2003 09:33:50 CST

15:33:56 <DanC> .time UTC

15:33:56 <datum> 2003-04-02T15:33:56

15:34:01 <libby> ooh

15:34:47 <DanC> swBot, if we agreed to meet every other weds at 1700Z, when is the next meeting?

15:34:55 <libby> wow

15:34:58 * libby waits

15:35:03 * DanC is dreaming

15:35:21 <libby> I got one of these: http://www.sonyericsson.com/P800/ but I can;t send it icalendar at the moment

15:35:34 <libby> but you can program it (in mini java or something)

15:36:02 <DanC> so that's your WearableGizmo now?

15:36:08 <libby> yeah :)

15:36:37 <libby> I want to send pics direct to webpage not manged that either yet. but has potential...

15:36:53 <DanC> can you email pics from it?

15:37:12 <libby> I would be able to if I cold get email workking. it should do sooon....

15:37:36 <DanC> phpht. "this site requires macromedia malarky?

15:37:42 <DanC> s/?/"/

15:37:52 <libby> the p800 one?

15:37:57 <DanC> yes

15:38:03 <libby> ew, sorry

15:39:02 <libby> that's a bit better, though big pics: http://www.mobile-review.com/review/sonyericsson-p800-e.shtml

15:40:08 <libby> I can't work out why email not working. very frustrating

15:40:18 <DanC> symbian OS... derived from psion, which I loved (see PsionWearableGizmo)

15:40:38 <libby> yeah, symbian seem to be doing well for themselves

15:42:00 <DanC> integrated camrea, infty. no querty keyboard :-{

15:42:05 <DanC> nifty

15:42:23 <libby> yeah, camera is not bad, about as good as my old palm one

15:42:38 <libby> keyboard, yes, will make irc a bit slow...

15:42:42 <DanC> infra-red port... I liked that in my NokiaWearableGizmo

15:43:02 <libby> I've been using bluetooth a lot

15:43:59 <libby> one annoyign thing: can;t transfer more than one photo at a time (syncing only works with windows at the moment, although it supports syncml)

15:44:02 <JibberJim> Are wearablegizmo's only available on contract at sensible prices DanC?

15:45:02 <DanC> hmm... syncml... I wonder if there's open source support for that.

15:45:10 <DanC> (i.e. debian or at least linux)

15:45:28 * DanC doesn't get JibberJim's question; wonders if it's a joke

15:46:01 <libby> prepaid phone vs contract phone?

15:46:04 <aaronofmontreal> syncml + oss : http://sync4j.sourceforge.net/web/theproject.html

15:46:05 <JibberJim> no, the wearablegizmo thingies in the US, can you only buy them with a phone contract?

15:46:17 <DanC> phpht... avantgo and lotus mobile notes... is there a plain old web browser?

15:46:38 <libby> in p800? yep and you can get opera for it

15:46:38 <JibberJim> (which would presumably need residency etc. - I'm going to be in the US for a number of months and would quite like one is all.)

15:47:33 <DanC> umm... the sidekick might be available without a contract... but it's like $450 in that case. I got it for $100 (after rebates) with a one year contract. For a few days it was available free after rebates with a contract.

15:48:05 <DanC> libby, what's the pricing model for data for your p800?

15:49:02 <libby> mine was something like 250UKP (after 80 quid off for an upgrade) + my contract + 0.2 pence a K download.

15:49:26 <libby> (I was going to wait for 3G but I lost my phone in boston :(

15:49:47 <libby> oop, got to go. I'll email you a photo danc, when I work it out....

15:50:32 <DanC> the flat data rate is the killer feature for the t-mobile sidekick plan.

15:51:07 * JibberJim wonders if they'll be nice to me roaming in the US...

15:51:18 <zool> edd's doing some syncml stuff for debian

15:52:57 <Eros> Eros is now known as mhgrove

15:54:19 <jhendler_>http://owl.mindswap.org

15:54:20 <dc_rdfig> G: http://owl.mindswap.org from jhendler_

15:54:54 <jhendler_> G:| MindSwap project "Semantic Web" home pages

15:54:55 <dc_rdfig> Titled item G.

15:55:15 <jhendler_> G: web site entirely driven from ontologies and RDF data, using lots of RDF tools

15:55:15 <dc_rdfig> Added comment G1.

15:55:27 <jhendler_> G: authorable in N3 (sorry, password protected for now)

15:55:28 <dc_rdfig> Added comment G2.

15:55:55 <jhendler_> G: all hierarchical indentation and the like is generated from RDF query results and the accompanying OWL

15:55:55 <dc_rdfig> Added comment G3.

15:56:05 <jhendler_> G: ontologies

15:56:05 <dc_rdfig> Added comment G4.

15:56:42 <DanC> have you gotten past the masochism stage yet? i.e. are there any tangible benefits that outweigh the PITA factor?

15:57:26 <DanC> with my travel tools, it took me about 9 months to get to the point where it was really a net win to get the computer to understand what's going on.

15:58:09 * JibberJim is about to start modelling his travels with RDF so he can draw little SVG docs of them, and generate a travelogue...

15:58:11 <jhendler_> G: the [http://owl.mindswap.org/2003/arbitrary/ | RDF Map] pulls all the classes and shows instances

15:58:11 <dc_rdfig> Added comment G5.

15:58:41 <jhendler_> DanC - fast answer is "barely" -- we have some nice stuff for authorability that, when we get a better GUI on it, will make it easier to use for things like adding papers and news items

15:59:34 <jhendler_> also, the http://owl.mindswap.org/2003/creations/ page, which shows everything created by one of the folks in it

15:59:54 <jhendler_> is a nice addition, and done from the dc: metadata that gets attached to the files

16:00:10 <jhendler_> we have a long way to go, and need a lot of other pages to link this to before it will really be a win, but

16:00:39 <jhendler_> I think it is a start at showing that OWL and RDF can have some benefits on the "normal" Web

16:02:18 <jhendler_> G: Credit to the Mindswap group and particularly the Alfords (bayta and ronwalf) and Bijan

16:02:19 <dc_rdfig> Added comment G6.

16:02:42 <DanC> G:bonus points for "oh yeah?" buttons

16:02:42 <dc_rdfig> Added comment G7.

16:03:11 <DanC> G7:I'd like to see "oh yeah?" buttons. 1500 bonus points if you can implement them

16:03:11 <dc_rdfig> Replaced comment G7.

16:03:45 <DanC> G:meanwhile, the [http://owl.mindswap.org/people/|people page] is already pretty interesting. How decentralized is it?

16:03:45 <dc_rdfig> Added comment G8.

16:05:17 <jhendler_> G: relatively - if you add a new person in N3, define a type for them and make it an appropriate sublcass, it will show up

16:05:18 <dc_rdfig> Added comment G9.

16:05:21 <DanC> hmm... <rdf:Description rdf:about="http://www.mindswap.org/papers/SMORE.pdf">

16:05:21 <DanC> <ns0:creator rdf:resource="mailto:golbeck@mindswap.org"/>

16:05:30 <DanC> that paper was written by a mailbox?

16:06:05 <jhendler_> DanC - we use the mindswap email address as the unique identifier for people - we'd love to figure out a better way to go

16:06:39 <jhendler_> on the other hand, we stole the idea from Foaf...

16:06:58 <DanC> no, the foaf world is clear about the difference between people and their mailboxes.

16:07:00 <jhendler_> let's try an experiment, JimH will add DanC as a "friend of the family" and see how long it takes...

16:07:02 <JibberJim> No, it's creator was a foaf:Person with foaf:mbox="..." ...

16:07:09 <JibberJim> in foaf world.

16:07:27 * DanC feels the urge to document this AntiPattern...

16:07:43 <jhendler_> true, you are right - we have a "swapper" class with the mbox in the right place, suspect wheover entered this

16:07:48 <jhendler_> paper did it wrong

16:07:57 * jhendler_ goes off to add DanC

16:08:48 <DanC> do you use OWL to find bugs? ala "mailboxes and people are disjoint. papers are written by people. you said that paper was written by a mailbox. I don't think you meant that."

16:12:19 <jhendler_> DanC now added - check people page

16:13:15 <jhendler_> would have been faster, but I forgot to end a little piece of N3 with a "."

16:13:36 <DanC> hm... the connection to http://www.mindswap.org/~danC/ seems bogus

16:13:41 <jhendler_> DanC - not yet using OWL reasoning to do these checks, that is next, soon as Bijan's OWL Lite reasoner can be hooked in

16:14:18 <DanC> so you answered my question about how easy it is to add stuff...

16:14:33 <jhendler_> Didn't have you regular web page - plus thought you might like to see the 404 error (except I set it wrong - try http://owl.mindswap.org/~DanC

16:14:34 <DanC> but not how decentralized it is. Is there a big central database of facts?

16:14:53 * timb| notes Eric newmann of Beyond Genomics needs a transtive property aware RDF query/reasoner system. Any suggestions?

16:14:54 <jhendler_> but, we have a long way to go - we need to hook our markup and editing tools in, that's not done yet.

16:15:33 <jhendler_> Ahh, I misunderstood - yes, currently we are using central DB.

16:15:47 <DanC> er, you don't think cwm is appropriate, timbl?

16:15:58 <jhendler_> Haven't figured out how to distribute yet, in part because we're all on one server at the moment

16:15:59 <DanC> parka is transitive-property-aware, right JibberJim?

16:16:17 <DanC> s/JibberJim/jhendler_/

16:16:35 <jhendler_> parka has good transitive property handling, the RDF front end still a little weak

16:16:46 <DanC> so if there's a central DB, you're competing with php.

16:17:32 <DanC> though arbitrary RDF/N3 input could be seen as a feature w.r.t. custom forms for every sort of fact you might want to enter.

16:17:43 <jhendler_> timbl, network inference stuff handles transitivity, but probably would constrain him to DL stuff

16:18:09 <DanC> meanwhile, I suppose RDF/N3 input isn't that novel w.r.t. a "type your SQL here:" form.

16:18:31 <jhendler_> DanC - yes re: phP in the sense of centralization, but we can handle arbitrary rdf input -- example, give me a pointer to an N3 page (not too big for now, I'm still figuring out my interface) and I can add it to the

16:19:04 <DanC> if I give you a pointer to an N3 document, can you add it *and track changes*?

16:19:17 <jhendler_> DB in one submit -- any rdfs:class(es) it would then show up in the RDF map

16:19:37 <jhendler_> track changes - only in the sense that we can tell what file it came from, and can remove all that RDF in one shot

16:19:41 <DanC> I keep lots of RDF about me in http://www.w3.org/People/Connolly/home-smart.rdf (and linked from there using rdfs:seeAlso)

16:19:44 * timb| logger_1, pointer?

16:19:51 <timb|> logger_1, pointer?

16:19:51 <timb|> See http://ilrt.org/discovery/chatlogs/rdfig/2003-04-02#T16-19-51

16:20:20 <jhendler_> let me see what happens

16:20:50 <DanC> one of the richer seeAlso link targets is http://www.w3.org/People/Connolly/travel-sched

16:21:07 <jhendler_> I think it will snarf that, but don't know what will show up since our display is "class" based

16:21:51 <DanC> do you follow-your-nose to find schemas for properties and classes?

16:21:53 <JibberJim> DanC, any chance of the cityl namespace coming out into the realworld?

16:22:04 <DanC> cityl namespace?

16:22:18 <JibberJim> xmlns:cityl="file:/home/connolly/w3ccvs/WWW/2000/10/swap/pim/cityLookup#"

16:22:35 <JibberJim> in your travelsched page.

16:22:39 <DanC> oops! that's http://www.w3.org/2000/10/swap/pim/cityLookup

16:23:09 <JibberJim> guessed it might be.

16:23:48 <DanC> aha... here's the problem, or part of it: @prefix : <cityLookup#>.

16:23:49 <DanC> @prefix cityl: <cityLookup#>.

16:23:57 <DanC> I could absolutize those.

16:24:12 * DanC grubles... xml-uri wars...

16:24:14 <jhendler_> DanC - it took your http://www.w3.org/People/Connolly/home-smart.rdf RDF just fine, but nothing shows up anywhere because you don't use any classes and it

16:24:41 <jhendler_> therefore just creates a graph that doesn't link to anything -- we'll have to think about how to display such things

16:25:22 <DanC> I recommend (1) an option to follow-your-nose to schemas when slurping in a file, and (2) an option to compute the RDFS closure (i.e. exploit domain/range)

16:26:01 <jhendler_> it had problem w/your namespace defn's for some reason

16:26:07 * DanC started documenting the follow-your-nose idea in http://esw.w3.org/topic/SelfDescribingWeb

16:26:12 <timb|> (danc, re cwm appropriate, I can show him how to use cwm, but i also want him to be able to use a afst dl reasoner when he has lots of data)

16:26:40 <jhendler_> timbl - the NI and Fact stuff are probably best if he can live in DL world

16:26:56 <DanC> some of my namespace defns are buggy file: thingies, as JibberJim noted

16:27:04 <jhendler_> NI = Network Inference (a W3C member org)

16:27:21 <DanC> speaking of owl tools... jim, I'm thinking of using the wiki to index them

16:27:52 <DanC> i.e. obsoleting http://www.w3.org/2001/sw/WebOnt/#Tools in favor of something like OwlTools

16:28:21 <DanC> and/or OwlFaq

16:28:55 * DanC noodles... ListMode is traditionally an antipattern...

16:29:37 <jhendler_> woohoo - mistake was on my end -- I put your RDF in correctly and it linked in immediately

16:29:43 <jhendler_> DanC - can you see http://owl.mindswap.org/2003/arbitrary/class.shtml?class=http%3A%2F%2Fwww.w3.org%2F2000%2F10%2Fswap%2Fpim%2Fcontact%23Person

16:30:14 <timb|> What is the antipattern for keeping lots of partial lists not in sync with no pointers between them?

16:30:49 <jhendler_> I'm tempted to say "Semantic Web" but suspect Timbl would correctly call me a wise ass...

16:30:57 <DanC> good question, timbl. I started discussing it in ESW:ProxyTopic

16:31:33 <timb|> suggest i.e. obsoleting http://www.w3.org/2001/sw/WebOnt/#Tools in favor of daml.org.

16:31:35 <timb|>http://www.daml.org/tools/

16:31:36 <dc_rdfig> H: http://www.daml.org/tools/ from timb|

16:31:53 <timb|> H:|DAML Tools

16:31:53 <dc_rdfig> Titled item H.

16:32:00 <DanC> er... well, I can't write to H:, can I?

16:32:20 <timb|> tools@daml.org

16:32:58 <timb|> If someone has volunteered and is being paid out of your tax dollars to accept and catalog infor sent to tools@daml.org, that is an offer i would take up.

16:33:16 <jhendler_> Dan - there is a "submit" on H:, but it is not externally writable (nothing on the DAML site is because of govt rules)

16:33:36 <DanC> a submit thingy is pretty good...

16:33:43 <jhendler_> timbl, Mike Dean managed the tools page

16:34:00 <DanC> are they indexing OWL tools? and being clear about the distinction?

16:34:05 <timb|> Yes, and he is a friendly person.

16:34:18 <jhendler_> hmm, used to be a form to fill out, now it just says Send additions/updates/corrections to tools@daml.org.

16:34:50 <jhendler_> my suspicion is that he is doing it by hand because people generated bad DAML when they did it themselves...the page is maintained from DAML I believe

16:34:50 <timb|> H:Send additions/updates/corrections to tools at daml.org.

16:34:51 <dc_rdfig> Added comment H1.

16:35:03 <DanC> is there some official "we're moving to OWL" thing I could point to yet?

16:35:19 * timb| suggests that comemnts on H be prepended by H and a colon, jh

16:35:20 <jhendler_> H: See http://www.daml.org/tools/tools.daml for the rdf/xml version

16:35:21 <dc_rdfig> Added comment H2.

16:35:40 <timb|> I pointed to the daml.org tools page from the RDF FAQ.

16:36:26 <timb|> When you do taht you have to make an explanatio of the term "DAML".

16:36:50 <jhendler_> danC - I don't think there is anything there yet - Murray told people at the last meeting, but not sure. I know the ontology library will now take OWL submissions

16:37:04 <jhendler_> timbl - not sure what you mean "explanation of term DAML"

16:37:08 <DanC> tim, you're thinking that "DAML" will come to mean "OWL"?

16:37:36 <timb|> H: I would like to see a rough categorization betwen commercial and academic projects, as now people are looking for serious products.

16:37:37 <dc_rdfig> Added comment H3.

16:37:59 <jhendler_> H: (DAML -> DAML+OIL -> OWL ) for those not tracking ontologies on the web...

16:37:59 <dc_rdfig> Added comment H4.

16:38:07 <timb|> I'm thinking that people will start to say "OWL" where they used to say "daml+oil".

16:38:10 <DanC> H:cited from [tools section of WebOnt WG|http://www.w3.org/2001/sw/WebOnt/#Tools] now.

16:38:11 <dc_rdfig> Added comment H5.

16:38:38 * timb| JimHG, see how to give a link in a comment )

16:38:45 <jhendler_> timbl - I hope so too, suspect for a while we're going to have both around

16:39:09 <DanC> H2:an [rdf/xml version| http://www.daml.org/tools/tools.daml] is available toofor the rdf/xml version

16:39:09 <jhendler_> timbl - you mean like I did in G: ?

16:39:09 <dc_rdfig> Replaced comment H2.

16:39:14 <DanC> H2:an [rdf/xml version| http://www.daml.org/tools/tools.daml] is available too

16:39:14 <dc_rdfig> Replaced comment H2.

16:39:19 * DanC hates InYourFaceUrls

16:40:02 <timb|> H: When someone tells you about the next cool tool, make sure they have sent it to this list.

16:40:02 <dc_rdfig> Added comment H6.

16:40:21 <jhendler_> ok, I sometimes use them when I want to share the "semantic content" in the URI -- that is, in that case that it was a ".daml"

16:40:50 <DanC> H2:a [daml/rdf/xml version| http://www.daml.org/tools/tools.daml] is available too

16:40:50 <dc_rdfig> Replaced comment H2.

16:40:55 <jhendler_> your way is better for this case - I yield (I was also being lazy I confess - I hate remembering syntax and this chump has become and entire command langauge)

16:41:04 <jhendler_> s/and/an

16:41:13 * DanC hates PoorMansHypertext too

16:41:46 <DanC> cf http://esw.w3.org/topic/PoorMansHypertext and http://esw.w3.org/topic/ImmersiveHypertextEditing

16:42:12 * timb| notes DanC just used a form of it

16:42:23 * timb| tx dan for elaborating

16:43:07 * timb| wishes we could have a set of current prefixes. Must make bot some time, or hack someone else's

16:43:31 <jhendler_> danC - footnote in ESW cute - I hadn't seen that before

16:43:49 <DanC> too cute. I don't want to invest too much in MoinMoin

16:43:59 <DanC> (it was added by the MoinMoin developer)

16:45:48 <DanC> ok, nearly 11am and I haven't finished reviewing my todo list; I have few appointments today, so it would be a shame to let the day go by without getting into something non-trivial.

16:46:12 <DanC> tim, gottaminute to help debug my papertrail bank stuff?

16:46:38 <DanC> the list bugs are gone/better, but now I'm seeing other wierdness.

16:47:18 * jhendler_ will let you guys get real work done - back to WebOnt stuff.

16:48:03 <timb|> timb| is now known as tim-away

16:48:22 <DanC> swap/ppt-bank$ cwm checking.n3 bankSW.n3 --think >,xxx.n3

16:48:24 <tim-away> danc, later

16:48:36 <DanC> doesn't do what I expect.

16:48:40 <DanC> ok, later, tim-away

17:05:57 <jhendler_> DanC?

17:06:30 <jhendler_>http://www.omg.org/cgi-bin/doc?ad/2003-03-40

17:06:31 <dc_rdfig> I: http://www.omg.org/cgi-bin/doc?ad/2003-03-40 from jhendler_

17:06:50 <jhendler_> I:| OMG Ontology Definition Metamodel RFP

17:06:50 <dc_rdfig> Titled item I.

17:07:39 <jhendler_> I: This RFP seeks a specification of a MOF2 Metamodel, UML2 profile, and

17:07:39 <dc_rdfig> Added comment I1.

17:07:51 <jhendler_> I: any additional information needed to support:

17:07:52 <dc_rdfig> Added comment I2.

17:08:08 <jhendler_> I: 1 - Development of ontologies using UML modeling tools

17:08:09 <dc_rdfig> Added comment I3.

17:08:26 <jhendler_> I: 2 - Implementation of ontologies in the W3C Web Ontology Language OWL

17:08:26 <dc_rdfig> Added comment I4.

17:08:40 <jhendler_> I: 3 - Forward and reverse engineering for ontologies.

17:08:40 <dc_rdfig> Added comment I5.

17:08:55 <DanC> yes?

17:12:34 <dajobe> DanC: re open source, syncml and bluetooth for p800. Parts are already in the kernel, in debian and in gnome or p800 developers near here are working on them.

17:15:20 <DanC> BLURB: syncML and RDF

17:15:20 <dc_rdfig> J: syncML and RDF from DanC

17:15:45 <DanC> J:libby got a p800, which caused DanC to wonder about SyncML in the opensource/debian/linux world

17:15:45 <dc_rdfig> Added comment J1.

17:16:21 <DanC> J:dajobe reports: Parts are already in the kernel, in debian and in gnome or p800 developers near here are working on them.

17:16:21 <dc_rdfig> Added comment J2.

17:17:24 <DanC> J:e.g. [sync4j|http://sync4j.sourceforge.net/web/theproject.html]

17:17:24 <dc_rdfig> Added comment J3.

17:17:48 * DanC collects thoughts on UML...

17:20:57 <tim-away> tim-away is now known as timbl

17:27:19 <earle> earle is now known as grault

17:30:43 <DanC> I:brain dump of relevant stuff: [http://esw.w3.org/topic/UML_20and_20RDF|ESW:UML and OWL]

17:30:43 <dc_rdfig> Added comment I6.

17:40:25 <DanC> D:see [ESW:AdvancedDevelopment|http://esw.w3.org/topic/AdvancedDevelopment] for notes on development process.

17:40:26 <dc_rdfig> Added comment D3.

18:17:06 <DanC> jhendler_, consider sending the LC message you sent to www-webont-wg to public-webont-comments too, to mark the LC epoch in the archives

18:17:39 * dajobe wonders did he get a reply to his p-w-c mail

18:17:59 * danbri_lap wanders past, waves

18:18:05 * jhendler_ good idea, will do

18:19:10 <danbri_lap> jim, could you send the webont LC to www-rdf-interest and www-rdf-logic? (perhaps redirecting followup discussion to the latter...?)

18:19:18 <danbri_lap> Or I could do it, whichever you prefer

18:19:29 <danbri_lap> congratulations btw!

18:19:56 <jhendler_> danbri - those are on my to-do list, am crossing t's and dotting i's, and then will do those probably within next hour or so

18:20:07 <danbri_lap> okay, thanks!

18:20:20 <jhendler_> thanks.

18:20:47 <danbri_lap> What are your expectations re LC reviews from RDFIG? Just that individuals sent comments directly, or do you want me to chase/solicit specific input of any kind?

18:21:00 <jhendler_> I'm open to either

18:21:06 <jhendler_> which would you prefer?

18:21:15 <jhendler_> how did you do this for RDF Core?

18:21:49 * danbri_lap suggests danbri not being bottleneck is good, so direct comments, but I could certainly encourage reviews as a followup to your announcement

18:22:08 * jhendler_ thinks that works all ways round

18:22:27 * danbri_lap nods

18:24:45 <timbl> Folks, when i am testing cwm parsers, etc, what is the graph comparison tool, or graph canonicalization tool, I should use?

18:25:06 <timbl> Rumors are that there are several.

18:25:15 <timbl> I don't want to use cwm, as I want to use it to test cwm.

18:25:33 <dajobe> either use ntc (c++, simple, fast)

18:25:40 <libby> I've been using the jena one, that seemed to work nicely, although no hint as to where you went wrong when there's an error

18:25:52 <libby> (which is a tricky problem I guess)

18:26:06 <dajobe> or the graph match API in jena (java, heavily optimised for bnode-heavy graphs)

18:26:10 <timbl> one - canonicalizer or comparer?

18:26:30 <dajobe> there is no canonicalizer

18:26:37 <libby> comparer

18:27:28 <dajobe>http://nicolas.bougues.net/syncml/

18:27:30 <dc_rdfig> K: http://nicolas.bougues.net/syncml/ from dajobe

18:27:35 <dajobe> K:|SyncML tools

18:27:35 <dc_rdfig> Titled item K.

18:27:47 <dajobe> K:partial implementation in PHP/MySQL

18:27:48 <dc_rdfig> Added comment K1.

18:28:26 <timbl> Thanks - I'll look at ntc

18:28:37 <dajobe> it's in the w3.org rdf test cases area

18:29:00 <danbri_lap> ntc looked neat and small, would be a student-project sized effort to rewrite in Python/Ruby/Perl (though would maybe be awful slow with bnode-heavy graphs?)

18:29:07 <danbri_lap> under util/ i think

18:29:57 <dajobe> the reason the jena one runs ok is the incredible amount of optimisation to colour the graph use bitmasks so that the java runs fast enough

18:30:27 <timbl> NTC is http://www.w3.org/2000/10/rdf-tests/rdfcore/utils/ntc/compare.cc

18:30:28 <dajobe> the simple ntc algorithm (moderate colouring) is fine for test case sized graphs

18:30:33 * libby likes nice slow java code

18:30:33 <dajobe> yes, that dir

18:30:50 <dajobe> ntc dies on graphs with ~15-20 bnodes

18:31:34 <danbri_lap> dave, any prospect ntc could get contrib'd to / bundled with redland? would be a handy addition to the toolkit...

18:31:40 <dajobe> yes

18:31:43 <danbri_lap> ooh

18:31:50 <timbl> I made a simple one in python http://www.w3.org/2000/10/swap/cant.py which, it seems, uses bascially the same start as ntc, using signatures. But cant is a cannonicalizer.

18:31:51 <dajobe> maybe I should make a redland todo wiki page

18:32:07 <danbri_lap> yeah, make a world-writable todo list!

18:32:09 <danbri_lap> (sucker ;)

18:32:31 <dajobe> timbl: canonicalizer is one way to do it, but you'll end up with ambiguity and sitll have to work through the combinations. That's what the comparers do sort-of anyway

18:32:57 <timbl> Not is tyhe canonicalizer removes the ambiguity

18:33:05 <timbl> (ie really canonicalizes)

18:33:15 <dajobe> ah

18:33:38 <dajobe> I don't believe that's possible

18:34:22 <dajobe> anyway, the test cases tend to have few bnodes - we made care to do that - except where it was useful, required

18:34:32 <dajobe> so that you can *nearly* compare with sort & uniq

18:35:08 <dajobe> danbri_lap: no chance, but requests or proto-designs might make it to world-writable

18:36:20 <timbl> ntc's signatures seem just to look at how many times a bnode appears in each position. cant's signatures include the node labales (URIs) where they exist.

18:36:27 <timbl> That makes the signatures much more distinct for a typical RDF graph with lots of URIs anad literals around.

18:36:30 <dajobe> yes

18:36:31 <timbl> So one can then order the bnodes.

18:36:34 <dajobe> that's what jena does, and more

18:36:46 * danbri_lap thinks yeah, what dave said, jena fingerprints are more detailed

18:36:49 <dajobe> but ntc was good enough, the C++ fast enough for the rest

18:36:57 <dajobe> jeremy wrote a whole paper on the compare algorithm

18:37:12 <timbl> in Jena.

18:37:16 <dajobe> you could probably just write that in python

18:37:17 <dajobe> yes

18:37:31 <dajobe> ... since it has similar features available in the language - dicts, hashes, sets or whatever

18:38:02 <timbl> If I use Jena I should use 1,6.1 or 2.alpha?

18:38:18 <dajobe> I'd say 1.6.1 but I don't use it much

18:38:32 <timbl> I don't see any reason to rewrite it - in fact several good reasons not to.

18:43:17 <libby> I've been using 1.6.1; 2 might be better but only a partial test release

18:43:31 <libby> s/better/have more functionality, dunno/

18:44:50 <timbl> thanks

18:46:00 * danbri_lap mails www-rdf-dspace to mention the ESW wiki, as they're thinking about collaboration tools currently

18:46:28 <dajobe> was there any reason to stick with the w3 usernames on the wiki?

18:46:40 <danbri_lap> w3 usernames?

18:46:51 <dajobe> yes, the ones for acls etc.

18:47:12 <danbri_lap> the new wiki doesn't have acls, beyond simple cookie setting thing for setting your name in 'latest changes'

18:47:28 <dajobe> ok

18:48:36 <timbl> AntiPatterns += SearchPath

18:49:17 <timbl> esw:SearchPath dc:title "The evils of search paths, class paths, etc."

18:54:36 * sandro is busy implementing search paths. :-)

18:58:56 * danbri_lap gets to bottom of mailbox

18:59:08 <jhendler_>http://www.geocities.com/ijmorrow/soap12-evil-feature.html

18:59:08 <dc_rdfig> L: http://www.geocities.com/ijmorrow/soap12-evil-feature.html from jhendler_

18:59:22 <jhendler_> L:| Soap Version 1.2 Evil Feature

18:59:22 <dc_rdfig> Titled item L.

18:59:32 <jhendler_> L: Dated April 1, 2003

18:59:32 <dc_rdfig> Added comment L1.

19:01:06 * jhendler_ wonders if April 1 date for WOWG LC announcement is a bug or a feature...

19:03:37 * timbl assumes Jena has a command line command for things like comparing graphs but can't find command line doc

19:06:23 <dajobe> you probably need the src/jena/rdfcompare.java program

19:06:31 * DanC realizes the owl comment he's looking for is in his unknown-senders inbox; finds it full of zillions of email bounces, drek, spam, and a few real messages that he should have seen

19:06:48 <timbl> tx

19:07:15 * timbl thanks dajobe

19:07:27 <dajobe> np

19:07:57 * jhendler_ found it amusing recently to discover that mail from the University's Office of Research Administration got a 10+ from spamassassin and ended up in my spam mbox - found all my financial memos by accident...

19:13:51 <danbri_lap>http://www.artima.com/weblogs/viewpost.jsp?thread=4204

19:13:51 <dc_rdfig> M: http://www.artima.com/weblogs/viewpost.jsp?thread=4204 from danbri_lap

19:13:58 <danbri_lap> M:|Sloppy and Forgiving versus Strict Systems

19:13:58 <dc_rdfig> Titled item M.

19:14:06 <danbri_lap> M:by Bill Venners

19:14:06 <dc_rdfig> Added comment M1.

19:14:18 <danbri_lap> M:"For social programming to work, specifications must in theory be strictly adhered to by both providers and consumers. But some systems seem to work in practice with loose adherence to specifications, systems in which sloppy providers and forgiving consumers are the norm. How important is strict adherence to contracts?

19:14:19 <dc_rdfig> Added comment M2.

19:14:27 <danbri_lap> M2:"For social programming to work, specifications must in theory be strictly adhered to by both providers and consumers. But some systems seem to work in practice with loose adherence to specifications, systems in which sloppy providers and forgiving consumers are the norm. How important is strict adherence to contracts? "

19:14:29 <dc_rdfig> Replaced comment M2.

19:16:22 <danbri_lap> M:This, recast somewhat, could be a good argument for XML wellformedness + namespaces as a flexible baseline, with RDF providing some scruffy-friendly structure on top of that...

19:16:23 <dc_rdfig> Added comment M3.

19:17:01 <danbri_lap> M:But I don't like all these noises I'm hearing lately about illformed "XML" being an acceptable thing to publish, citing RSS as motivation...

19:17:01 <dc_rdfig> Added comment M4.

19:30:34 <Morbus> at M4

19:36:10 <Wack> hmm

19:36:41 <Wack> why is the parseType="Collection" introduced in the revised rdf syntax spec?

19:37:00 <Wack> that is, why not use a bag? (which results in less triples as well)

19:37:03 <dajobe> lol

19:37:16 * dajobe hides, points at jhendler_

19:38:24 <Wack> sure, a collection costs less to iterate (like linked lists), but that cannot be the only reason it was added :]

19:38:39 <jhendler_> Wack - the collection is there to have a mechanism for a "closed list" - we need it a lot in OWL

19:39:31 <jhendler_> for example, suppose you want to say that the only kinds of wine you can reason about are red or white

19:39:44 <jhendler_> You can use a owl:oneOf statement (with a collection) to handle that.

19:40:05 <jhendler_> You cannot let someone else come along and add another type to that list (as they could generally do with a bag)

19:40:21 <jhendler_> so your bag would have to have all that first, rest, nil stuff in it to make it a closed set

19:40:31 <jhendler_> parsetype=collection is a shorthand for that.

19:40:36 * timbl gets jena.rdfcompare working as comparison function in regression test butfind it doesn't like datatypes

19:40:39 <jhendler_> dajobe - is that a fair summary?

19:40:59 <dajobe> yes - closed collections

19:45:30 <Wack> hmm

19:47:29 <dajobe> timbl: ntc does datatypes, I added them

19:48:03 * dajobe notes the rdf validator is now way broke/out of date

19:48:13 <Wack> like, say it was a bag with wine types, you could make a (bag) =rdf:_x=> (foo) statement where x is the number of resources in the original bag + 1, whereas with a collection there is an effective 'stop' with the nil resource

19:48:36 <Wack> (?) :]

19:49:23 <timbl> dajobe - does ntc do everything you need for the rdfcore tests?

19:49:30 <dajobe> yes

19:49:55 <timbl> I should have asked that about Jena befroe I started. Still, its good to have it around.

19:49:59 <jhendler_> wack - not sure I understand - the "stop" is what we want - a way to say "and no others"

19:50:51 <Wack> i'm thinking too much in triples

19:51:37 <jhendler_> yeah, dajobe and I spent long hours discussing the differences between my thinking about it in terms of functionality and his in terms of triples

19:51:43 <Wack> but isn't this more a matter of authority rather than making it 'physically' impossible to add more?

19:52:20 <jhendler_> no, not really - we needed a way to say a particular class was DEFINEd by the closed list. So we either would need a content mechanism and a bunch of other stuff

19:52:37 <jhendler_> (so we could say Hendler defines the class wine to have only these two members)

19:53:06 <jhendler_> or else we need a way to say in triples that this is a closed list -- which is what we chose (together w/RDF Core)

20:01:47 <jhendler_> gotta go. back later.

20:01:54 <jhendler_> jhendler_ is now known as jimh-away

20:46:49 <timbl> _____________________

20:47:28 <timbl> Anyone know - Is an RDF literal with no type something with an unknown type or something with a default type?

20:47:43 <dajobe> it has no type

20:48:30 <timbl> I'm looking at how to reify it. How do I say that it has no type?

20:48:49 <dajobe> reify? you've lost me

20:49:13 <timbl> In a more practical, quoting, reification system - not the std one.

20:49:20 <timbl> I am experssing abinding in a proof

20:49:40 <timbl> "?x" bindsTo the literal "fred".

20:49:47 <dajobe> I had wondered if adding an rdf:PlainLiteral or something would be useful

20:49:57 <dajobe> cf rdf:XMLLiteral

20:51:09 <timbl> :binds [ a reify:Variable; reify:URI "http://...x"] ; :bindsTo [ a reify:PlainLiteral; reify:stringValue "fred"].

20:51:17 <timbl> that would work.

20:51:18 <dajobe> all you've got is the rdfs:Literal

20:51:22 <dajobe> which is the superclass

20:51:47 <timbl> :binds [ a reify:Variable; reify:URI "http://...x"] ; :bindsTo [ reify:plainLiteralStringValue "fred"].

20:52:09 <timbl> That makes it shorter, and we can say that the domain of plainLiteralStringValue is a PlainLiteral.

20:52:42 <timbl> Similarly, [] :binds [reify:variableURI "http://...x"] ; :bindsTo [ reify:plainLiteralStringValue "fred"].

20:53:11 <timbl> And with language, is it the same?

20:53:53 <timbl> I need some sort of "None".

20:54:10 <dajobe> no, with language makes a different literal

20:54:51 <timbl> Well, a datatype makes a different literal, no?

20:55:21 <dajobe> yes

20:56:43 <timbl> So there are treated about the same. Special-casing the plain literal sounds reasonable. Then for a nonPlainLiteral, one would look for a dt and a lang. The lang if not present could be ""

20:56:58 <dajobe> hmm

20:57:06 <timbl> the dt if not present could be some made up type which says nothing.

20:57:11 <dajobe> xml:lang="" means absence of language

20:57:20 <dajobe> and I think that's the same as not mentioning it

20:57:35 <timbl> Oh, i didn't know that xml:lang was defined.

20:57:42 <timbl> Oh, i didn't know that xml:lang="" was defined.

20:57:46 <dajobe> it is now

20:57:56 <dajobe> IIRC rdf core asked xml core

20:58:15 <dajobe> it was something related to c14n and xml fragments

20:58:20 <timbl> Ok, so if "" is the same as omission, then any plainLiteral has a lang of ""

20:59:06 <timbl> Is there anything wrong ith giving a URI to the datatype which you get when you don't mention a datatype?

20:59:46 <dajobe> that would be the imaginary rdf:PlainLiteral

20:59:53 <dajobe> I'm not sure

21:00:09 <dajobe> such a type would be special in the rdf semantics anyway, same as rdf:XMLLiteral

21:00:42 <timbl> I'm trying to think whether the class of the reifiedobject is the same as the datatype. I hadn't thought so.

21:01:07 <dajobe> empty xml lang errata - http://www.w3.org/XML/xml-V10-2e-errata#E41

21:01:29 <timbl> Similarly, [] :binds [reify:variableURI "http://...x"] ; :bindsTo [ reify:literalStringValue "fred"; rei:lang "en"; rei:dt xsi:Integer].

21:01:40 <timbl> s/Similarly//

21:02:06 <timbl> rei:dt here is the reltion between the literal (as a concept) and its datatype.

21:02:11 <dajobe> (not really an xsi:integer there ;)

21:02:25 <timbl> ?

21:02:33 <dajobe> "fred" an integer?

21:02:34 <timbl> oh, fred.

21:02:38 <timbl> scuse me!

21:02:44 <dajobe> I see what you are doing

21:02:51 <timbl> well, the examples are pretty weird with langauegs and datatypes

21:02:55 <dajobe> yes

21:03:20 <dajobe> I'd have to check if it is significant in the rdf dt entailment. I think it isn't

21:03:29 <timbl> [] :binds [reify:variableURI "http://...x"] ; :bindsTo [ reify:literalStringValue "123,000"; rei:lang "en"; rei:dt xsi:Integer].

21:06:06 <timbl> I wouldn't want to use the xsi:Integer class the class of RDF literals whose datatype is Integer. I would prefer to use it (as a class) as the class of Integers. I assume RDF has pronounced on that

21:06:44 <dajobe> I don't recall

21:16:27 <danbri_lap> oh, thanks hex/grault for improving the Grubstreet mention in http://esw.w3.org/topic/RestaurantRecommendation :)

21:18:03 <timbl> hasn't anyone modelled RDF in owl before?

21:23:52 <jimh-away> DanC

21:24:01 <jimh-away> jimh-away is now known as jhendler

21:26:12 <timbl> Duh... I odn't hink I have to model teh datatypes, just use a literal of the same type.

23:07:35 <danbri_lap>http://www.mozilla.org/roadmap.html

23:07:36 <dc_rdfig> N: http://www.mozilla.org/roadmap.html from danbri_lap

23:07:48 <danbri_lap> N:|New Mozilla Development Roadmap

23:07:48 <dc_rdfig> Titled item N.

23:08:28 <danbri_lap> N:Summary: moving to focus on Phoenix, slimdown, keeping XUL-based UI, tighter control over checkins etc.

23:08:28 <dc_rdfig> Added comment N1.