Semantic Web Interest Group IRC Chat Logs for 2004-03-26

00:11:28 * chaalsMEL waves to danbri_dna

00:16:21 <danbri_dna> ello

00:18:00 * chaalsMEL thinks SWAD-E will be having a workshop on image annotation in June, in Madrid. But still confirming availability of room.

00:26:03 <danbri_dna> w/shop, cool

00:59:37 <[Core]Chaos> hello..

00:59:52 <[Core]Chaos> is this for axis?

01:02:54 <hyper\> is this room soap related at all?

01:03:10 <sbp> barely, but there are probably some SOAPers about

01:03:26 <hyper\> do you perhaps know of a better soap channel?

01:03:57 <sbp> unfortunately not; this is a pretty good clearinhouse for such issues if you can't find anywhere else, though

01:04:28 <hyper\> im trying to find out what soap toolkits can handle persistent http1.1 connections for soap calls ;/

01:05:18 <sbp> apparently http://www.pocketsoap.com/pocketsoap/ can handle them

01:05:23 <hyper\> yeah

01:05:25 <hyper\> i saw they did

01:05:34 <hyper\> but they are tiny guys i need to see if some big toolkits handle it

01:06:10 <hyper\> you know PHB approval :( !

01:06:47 <sbp> I see. well, I'm not an expert; if you hang around, perhaps someone will see your question and chip in. otherwise, there are plenty of SOAP lists that you could ask on

01:06:56 <sbp> PHB approval: heh

01:07:01 <hyper\> ^_^

02:16:56 <zImage> I'm a newbie to RDF and have a few questions. Anyone have their ears on?

02:17:23 <IsoosI> maybe

02:18:25 <zImage> I'd like to make something similar to wordnet, but for biological taxonomy (Kingdom, Phylum, Class, ...).

02:19:03 <zImage> I have some code, but I'm hitting walls, since I don't know a whole lot about RDF.

02:19:22 <zImage> code = RDF vocabulary

02:19:29 <IsoosI> yep

02:19:39 <deltab> like http://tolweb.org/tree/phylogeny.html ?

02:20:12 <zImage> deltab- do they already have RDF?

02:20:21 <zImage> I was using itis as my data source.

02:21:33 <zImage> http://www.itis.usda.gov/ (You can download their database)

02:21:34 <dc_rdfig> A: http://www.itis.usda.gov/ from zImage

02:22:10 <MarkB> hyper, I can't think of a single toolkit that doesn't use persistent connections by default

02:22:38 <zImage> I have a Class called "TaxonomicTerm".

02:23:05 <zImage> the Classes "Kingdom", Phylum, etc are all subclasses of TaxonomicTerm.

02:25:02 <zImage> I've then got

02:25:21 <zImage> <Species rdf:about="H.Sapiens">

02:25:34 <zImage> <rdfs:subClassOf>

02:25:51 <zImage> <Genus rdf:about="Sapiens"/>

02:25:58 <zImage> </..?

02:26:01 <zImage> </..>

02:26:49 <zImage> This is wrong since I'm subclassing a Species instead of a Class. I don't know how to think about this problem, so I'm stuck...

02:27:23 <zImage> I thought I could do...

02:27:42 <zImage> <Class about="H.Sapiens">

02:28:16 <zImage> <tax:rank><tax:Species/></tax:rank>

02:28:30 <zImage> <subClassOf>

02:28:49 <zImage> <Class about="Sapiens">

02:28:55 <zImage> <rank ...

02:28:56 <zImage> ...

02:28:56 <zImage> ..

02:29:13 <zImage> but that seems crufty.

02:29:28 <zImage> any tips?

02:34:56 <hyper\> marb: still around ? when you say you cant think of any toolkits that dont use persistent connections, do you mean they stay open after the soap call is made?

02:35:13 <hyper\> markb: cause most toolkits close the socket after one soap call

02:36:12 <MarkB> hmm, Axis didn't IIRC

02:36:20 <zImage> Is there a better place I can go? I guess I just need to talk through my class model.

02:38:04 <hyper\>http://wiki.apache.org/old/AxisProjectPages/KeepAlive

02:38:04 <dc_rdfig> B: http://wiki.apache.org/old/AxisProjectPages/KeepAlive from hyper\

02:38:12 <hyper\> it doesnt look like it

02:38:22 <hyper\> it seems you can with non normal libraries ;(

02:52:31 <zImage> \quit

03:09:20 * DanC wonders if my debian linux box can do anything with this m-audio USB midi keyboard

03:09:48 <DanC> 1st bad sign: the light on the USB hub went on for just a split second and then off when I plugged it in.

03:11:46 <DanC> hmm... no usb-midi module in /lib/modules/2.4.16/

03:23:23 <DanC> more bad signs: not many clues in google

03:24:56 <IsoosI> hmm

03:25:04 <IsoosI> have you looked at ALSA?

03:26:02 <DanC> er... well, yes, I just surfed the ALSA web site.

03:26:12 <DanC> I think I've tried to get it running before and failed.

03:26:16 <DanC> I never tried really really hard.

03:27:40 <IsoosI> hmm

03:28:33 <DanC> # /etc/init.d/hotplug restart

03:28:34 <DanC> Restarting hotplug subsystem: input pci usb input[failed] pci** can't synthesize pci hotplug events

03:28:34 <DanC> usb.

03:28:35 <DanC> ew.

03:29:26 <DanC> ah... /etc/init.d/hotplug restart

03:29:28 <DanC> now I see

03:29:29 <DanC> Bus 002 Device 003: ID 0a4d:0090 Evolution Electronics, Ltd

03:31:00 <DanC> # usbmodules --device /proc/bus/usb/002/003

03:31:00 <DanC> audio

03:33:28 <DanC> hmm... "getting a midi keyboard working under linux" http://jeffcovey.net/music/midi/linux-midi-keyboard.html

03:35:14 <IsoosI> sweet

03:38:58 <DanC> that was just a pointer to somebody else's software.

03:39:12 <DanC> ... which software isn't packaged for debian.

03:39:20 <IsoosI> grr

03:39:28 <DanC> I just tried fluidsynth, based on its apt-cache description. no joy:

03:39:29 <IsoosI> thats irritating

03:39:38 <DanC> $ tail /var/log/messages

03:39:38 <DanC> tail: cannot open `/var/log/messages' for reading: Permission denied

03:39:38 <DanC> connolly@dirk:~/gfweb/docroot$ fluidsynth

03:39:38 <DanC> cca_open_socket: could not connect to host 'localhost', service '14541'

03:39:38 <DanC> cca_init: could not connect to server 'localhost' - disabling ladcca

03:39:38 <DanC> fluidsynth: pcm_params.c:2220: snd_pcm_hw_refine: Assertion `pcm && params' failed.

03:42:10 <DanC> [[ The easiest way to start the synthesizer is to give it a SoundFont on

03:42:10 <DanC> the command line: 'fluidsynth soundfont.sf2'. ]]

03:42:18 <DanC> ok, where the flip do I get a sound font?

03:43:05 <IsoosI> hmm

03:43:19 <IsoosI> they were very poopular back in the ol'd days of the Gravis Ultra sound

03:43:29 <DanC> "Soundfont is a standard for creating instruments (also called patches) for use in composing MIDI music." -- http://smurf.sourceforge.net/

03:44:59 <DanC> ya know what's really annoying? that hardware support docs don't use the customer-visible identifiers. they say stuff like "supports the philips WCA chip" or some such. how the heck do I know what chip is inside? tell me what USB id you support!

03:46:45 <DanC> aha... some soundfonts... http://hammersound.net/cgi-bin/soundlink.pl

03:49:46 <IsoosI> cool

03:50:02 <DanC> fluidsynth still dies as above

03:50:25 * DanC tries notedit , a kde thingy

03:50:38 <DanC> I haven't had much luck with arts, the KDE audio system

03:50:55 <IsoosI> indeed

03:52:26 <DanC> notedit spews several screenfuls of diagnostics and keels over

03:52:54 <IsoosI> oh dear

03:53:04 <IsoosI> anything particularly enlightening?

03:53:32 <DanC> well, this is somewhat telling:

03:53:34 <DanC> mv: cannot stat `/home/connolly/.kde/share/apps/kio_help/cache': No such file or directory

03:53:43 <DanC> and

03:53:45 <DanC> TSE3: Alsa scheduler error.

03:53:45 <DanC> Alsa is not running on this machine

03:53:45 <dc_rdfig> Label TSE not found.

03:53:58 <DanC> cannot create an ALSA MIDI Scheduler

03:54:56 * DanC plays another round of debian package roulette... muse ...

03:55:58 <DanC> muse emits .5 screenfuls of diagnostics, but does start up...

03:56:07 <DanC> ALSA lib seq_hw.c:446:(snd_seq_hw_open) open /dev/snd/seq failed: No such deviceCould not open ALSA sequencer: No such device

03:56:08 <DanC> ...

03:56:09 <IsoosI> well

03:56:11 <IsoosI> thats a good start

03:56:13 <DanC> cca_open_socket: could not connect to host 'localhost', service '14541'

04:00:09 <DanC> ok, going up the platform stack from debian to gnome, I find http://beast.gtk.org/ "BEAST - Your Solution for Music Composition and Synthesis"

04:01:18 <DanC> very nice web site... incl sample song http://beast.gtk.org/beast-ftp/bse-files/partymonster.ogg

04:03:03 <DanC> ooh... debian packages...

04:03:33 <DanC> Get:1 http://ftp.us.debian.org unstable/main beast 0.5.5-5 [2121kB]

04:04:03 <DanC> ding! starts with no diagnostic spew...

04:06:11 <DanC> help menu... with quickstart...

04:06:41 <DanC> demo right there in the menu... that seems to work, though it hungers for CPU

04:07:00 <DanC> very slick, this thing... now... how to get it to talk to my keyboard...

04:08:31 <DanC> " There's a #beast channel on the GimpNet IRC Network (Server: irc.gimp.org, Port: 6666)"

04:22:07 <IsoosI> wow, nice.

04:26:30 * DanC is getting real-time support from the developer now...

05:27:07 <DanC> BLURB: a search for open source usb-midi support

05:27:08 <dc_rdfig> C: a search for open source usb-midi support from DanC

05:27:41 <DanC> C:started [http://www.ilrt.bris.ac.uk/discovery/chatlogs/rdfig/2004-03-26.html#T03-09-20|a few hours ago].

05:27:41 <dc_rdfig> Added comment C1.

05:28:05 <DanC> C:I'm giving up, leaving a question in [http://wiki.debian.net/index.cgi?QandA|QandA] in the debian wiki

05:28:06 <dc_rdfig> Added comment C2.

09:41:07 <aharth> anybody figured out how to use firefox to display application/rdf+xml?

09:45:59 * aharth found mozilla bug id 61839 talking about application/rdf+xml

09:49:35 * aharth has already voted for mozilla bug 61839, encourages others to do the same

09:50:06 <dajobe> mime type?

09:50:18 <aharth> yep, application/rdf+xml in mozilla

09:50:25 <dajobe> I read it

09:50:26 <aharth> you are the other vote :)

09:51:06 <Remosi> has everyone voted for SVG? :)

09:51:34 <aharth> i hate that mozilla is not able to display RDF with the (maybe soon-to-be) correct content-type

09:51:42 <aharth> remosi: bug id?

09:52:08 <dajobe> there has been work in the last 2 weeks on registering rdf/xml as rfc

09:52:25 <dajobe> but the mysteries of the ietf process elude me

09:53:03 <aharth>http://www.ietf.org/internet-drafts/draft-swartz-rdfcore-rdfxml-mediatype-04.txt

09:53:04 <dc_rdfig> D: http://www.ietf.org/internet-drafts/draft-swartz-rdfcore-rdfxml-mediatype-04.txt from aharth

09:53:15 <Remosi>http://bugzilla.mozilla.org/show_bug.cgi?id=122092

09:53:15 <dc_rdfig> E: http://bugzilla.mozilla.org/show_bug.cgi?id=122092 from Remosi

09:53:21 <dajobe> yeah, we've asked for that to be advanced

09:53:41 <Remosi> you don't need a internet draft to get a mime type

09:53:51 <Remosi> you just fill in the form, and provide a URI to the standard

09:53:52 <dajobe> sigh, I don't want to discuss that again

09:54:00 <Remosi> to my understanding?

09:54:05 <Remosi> dajobe: uri?

09:54:08 <aharth> D:|RfC for mime-type application/rdf+xml

09:54:08 <dc_rdfig> Titled item D.

09:54:34 <aharth> D:|RfC draft for mime-type application/rdf+xml

09:54:34 <dc_rdfig> Titled item D.

09:56:22 <aharth> D: mozilla firefox support for application/rdf+xml would be nifty

09:56:22 <dc_rdfig> Added comment D1.

09:56:38 <dajobe> D:what would such support do?

09:56:39 <dc_rdfig> Added comment D2.

09:56:56 <aharth> D: bug id 61839 on bugzilla

09:56:56 <dc_rdfig> Added comment D3.

09:57:37 <aharth> dajobe: display RDF in mozilla browser window instead of saving to disk for starters

09:57:47 <aharth> maybe that's possible but i didn't figure out how to do that

09:58:02 <dajobe> it's unlikely to do that

09:58:13 <dajobe> like all other application/*, it tends to offer to save them

09:58:19 <dajobe> since it's not text/*

09:58:31 <dajobe> but that might be a choice of mozilla rather than mandated

09:59:08 <aharth> the current behavior makes it complicated to debug my services that return the correct content-type

09:59:21 <dajobe> I make mozilla run a helper app that displays the triples

09:59:29 <aharth> browse URI/save to disk/more the file

09:59:50 <aharth> what helper app do you have for that?

10:00:00 <dajobe> it's an example program that comes with raptor

10:00:08 <aharth> apt-get install raptor?

10:00:11 <dajobe> a Gtk viewer program called grapper

10:00:18 <dajobe> hmm, I don't think it's in the debs

10:00:23 <dajobe> since it's an example

10:00:35 <dajobe> maybe I should add it to raptor-utils

10:00:41 <aharth> would be cool

10:00:51 <dajobe> it's only 1 c file, you can surely compile that

10:01:10 <aharth> ok, i'll look into it

10:01:25 <dajobe> but I was asking if you thought mozilla should do something clever with the rdf

10:01:30 <dajobe> ... with the rdf/xml

10:01:33 <dajobe> since it does know rdf

10:01:55 <dajobe> currently they use a different unregistered mime type internally

10:01:55 <aharth> i don't think mozilla will be able to do something clever with RDF for quite a while

10:02:33 <aharth> hm... displaying the rdf in the calender if it's ical/rdf would be interesting

10:04:32 <aharth> what would be a "clever" use of RDF in mozilla?

10:04:46 <dajobe> that's what I asked you ! :)

10:05:15 <dajobe> for some application, mime type recognition could be inside the code of the app

10:05:20 <dajobe> and mozilla does know rdf

10:05:35 <dajobe> otherwise the mime type is unknown, so it can only offer other choices such as a user dialog

10:05:45 <dajobe> there might be a different bug for mozilla

10:05:52 <dajobe> which makes it recognise the '+xml' mime types internally

10:05:58 <dajobe> and view them using the XML view

10:06:14 <aharth> hm... having the XML view would be a start

10:06:21 <Remosi> er

10:06:21 <dajobe> but that's a different thing from recognising the application/rdf+xml correctly

10:06:30 <Remosi> doesn't mozilla use RDF internally for history/bookmarks/etc?

10:06:57 <dajobe> that's what I was saying

10:07:10 <dajobe> it knows, uses rdf, but communicates it externally via an unregistered mime type

10:07:14 <aharth> but i think to make "clever" use of rdf you have to look to more than just one file (combining information from files/repositories is the interesting thing imho)

10:07:53 <dajobe> files would be a start

10:10:06 <aharth> using flat files doesn't scale very well, and you get into problems with updates and staying current

10:10:28 <dajobe> now you are changing the subject ;)

10:10:28 <aharth> but it's a start

10:10:58 <aharth> i try to talk about _my_ work all the time *eg*

10:11:44 <aharth> hmm raptor's configure says no xml parser available but i install both libxml-devl and expat

10:12:05 <dajobe> libxml2?

10:12:55 <aharth> yes, that works

10:12:57 <aharth> thanks

10:17:16 * aharth installs libgtk1.2-dev to compile grapper

10:17:45 <dajobe> gtk2

10:19:17 <fidothe> morning all

10:19:20 <dajobe> hi

10:19:51 <fidothe> dajobe: btw where's the right place to make Redland python interface RFEs?

10:20:21 <dajobe> to redland-dev, here or email, whatever!

10:20:27 <fidothe> okay

10:20:45 <dajobe> patches attached to the email help make it go faster :)

10:24:19 <fidothe> I've run into a couple of problems with redland classes emulating some but not all the bits of python sequence types.

10:24:52 <fidothe> I don't know how hard some of that extra sugar would be to implement - I know nothing about C or the python/C bridge

10:24:57 <fidothe> unfortunately

10:25:33 <fidothe> it's not massively problematic: you can work around all the problems I've run into so far pretty easily

10:26:24 <dajobe> it probably can't do all of a python sequence api

10:26:34 <dajobe> since they are move-forward-only type things

10:26:47 <dajobe> you can't change them or do random access, etc.

10:27:03 <sh1m> moin

10:27:06 <fidothe> it's things like length that are the big bug bears: you're supporting iterators already

10:27:15 <dajobe> length's not possible

10:27:18 <dajobe> it's a streaming api

10:27:25 <fidothe> ahhhhh

10:27:33 <fidothe> that makes more sense now...

10:27:35 <dajobe> like efficiency :)

10:27:39 <aharth> dajobe: thanks, grapper works perfectly!

10:27:45 <dajobe> oh good

10:28:24 <fidothe> in that case implementing __nonzero__ or something would be cool: you can't do things like

10:28:36 <fidothe> if redland_stream: do things...

10:29:08 <dajobe> interesting

10:29:15 <dajobe> I'd have to look that up; I'll note it down

10:29:48 <fidothe> it's a shame because there's no way of knowing if a stream has any contents unless you iterate through it, so you replace that if statement with things like:

10:30:03 <dajobe> I see what you mean

10:30:31 <fidothe> startNode = None

10:30:31 <fidothe> for statement in siteModel.find_statements(RDF.Statement(None, parentPredicate, None)):

10:30:31 <fidothe> print statement

10:30:31 <fidothe> startNode = statement.object

10:30:31 <fidothe> break

10:30:31 <fidothe> startNode = None; for node in stream; startNode = node; break

10:30:39 <fidothe> and then if testing startNode

10:32:45 <fidothe> anyway, I'm really liking Redland

10:32:45 <fidothe> :)

10:32:46 <dajobe> wouldn't __nonzero__ be: def __nonzero__(self): return !self.end()

10:32:46 <dajobe> I'm guessing, not having looked at the python docs or tried it

10:32:59 <dajobe> thanks

10:35:24 <fidothe> hummm, yeah that looks like it would do the trick nicely

10:35:32 <fidothe> morning lib

10:37:15 <libby> mornin'

10:38:35 <fidothe> this morning i am mostly writing an RDF vocab. I imagine there'll be panicked cries for help every so often...

10:38:42 <fidothe> never done one before

10:39:04 <libby> heh. what's it for?

10:39:09 <fidothe> CLR

10:39:17 <fidothe> sorry, central lettering record

10:39:22 <libby> ah

10:40:00 <fidothe> at http://www.centralletteringrecord.org/

10:40:27 <fidothe> built most of an XML form system, now trying to tie RDF vocab-based stuff in to the framework

10:40:37 <fidothe> which means I need an RDF vocab...

10:40:59 <dajobe> nice

10:41:15 <fidothe> hmmm. is the shelley powers RDF book right on its terminology

10:41:21 <fidothe> ?

10:42:01 <fidothe> cos if it is i think when i say vocab i mean ontology.

10:42:01 <libby> which bit?

10:42:07 <fidothe> bollocks

10:42:09 <libby> sounds cool fidothe

10:42:26 <fidothe> slightly terrifying, actually

10:44:02 <fidothe> heh, is CLR's very own founder Nicholas Biddulph any relationship to #rdfig's very own mattb, i wonder

10:44:08 <sh1m> hi libby

10:44:31 <dajobe> I've asked him

10:44:34 <libby> hi sh1m

11:07:43 <danbri>http://web.sfc.keio.ac.jp/~kaz/www2004/

11:07:43 <dc_rdfig> F: http://web.sfc.keio.ac.jp/~kaz/www2004/ from danbri

11:07:52 <danbri> F:|WWW2004 Workshop on Content Labeling -Technical and Socio-Cultural Challenges and Solutions

11:07:52 <dc_rdfig> Titled item F.

11:08:14 <danbri> F:Deadline is today! I'd be interested to see some papers on PICS/RDF migration...

11:08:15 <dc_rdfig> Added comment F1.

11:36:47 <eaon> leo sauermann has joined irc

11:36:51 <eaon> i love crikey! ;)

11:37:03 <LeoBard> hoy eaon

11:37:27 <eaon> #franse?

11:37:28 <LeoBard> want to check out latest about iCal-RDF, going to give a talk about it...

11:38:48 <eaon> "Changes to iCalender ontology to make calendar files diffable"

11:39:03 <eaon> isn't that exactly what we needed? :)

11:45:00 <sh1m> the calander stuff is funky

11:45:09 <sh1m> there was a calander meeting the other day

11:45:12 <sh1m> wed i think

11:45:13 <libby> maybe but it's not strightforward in terms of what we've been trying to do, i.e. consistently and solely to translate icalendar to rdf

11:45:18 <sh1m> or you culd just pester libby

11:45:29 <libby> yeah wed at 1530 utc

11:46:13 <sh1m> libby: whats utc? is that like gmt|bst /

11:46:43 <libby> the main thing is that DanC has updated the schema to use OWL/RDFS. It also no lonjger uses the testcases to generate it but it generated with GRDDL from an annotetd XHTML version of the RFC 2445

11:46:47 <libby> yeah sh1m

11:47:30 <libby> DanC has got graph checking working - i.e. generating the rdf from icalendar files, cheking the generated graphs with repect to teh originals

11:47:37 <libby> but not roundrtripping though yet

11:47:38 <LeoBard> hm, I give a talk to fellow RDF/SemWeb people in Vienna, so I start with basics like http://esw.w3.org/topic/RdfCalendarPresentation

11:47:54 <libby> yeah good place to start

11:48:09 <libby> we were chatting about this yetserday too, getting the tests working

11:48:32 <LeoBard> @libby: btw: anything new about the swad-e t-shirts ? (i liked the brains..)

11:48:57 <libby> we have new ones; now we have to decide who to give them to :)

11:49:15 <sh1m> libby: do you know if Norm Walsh's page is automated? Like does it sycronise from his calander and do the geo stuff itself?

11:49:40 <sh1m> libby: some people will swap beer for shirts :P

11:49:42 <libby> new tshirts: http://liz.xtdnet.nl/swad/isofinal3.gif

11:49:44 <libby> :)

11:50:08 <libby> hm I think norms stuff is autiomated yep. but not sure presicely how

11:50:23 <libby> the brains are no more... :/

11:50:24 <LeoBard> I like the style - it "reminds me" of a röyksopp video :-))))

11:50:41 <libby> :)

11:50:53 <LeoBard> I used to show the röyksopp video "remind me" to other RDF fewllows at HP/Galway

11:51:30 <dajobe> that's a nice video, infographics style

11:53:44 <eaon> i want such a tshirt!

11:54:42 <fidothe> they're nice - but do they validate?

11:54:48 <LeoBard> @libby: can eaon and /me print us the shirts ourselves, are there copyright/left issues ?

11:56:17 <libby> I think there will be copyright issues

11:56:27 <libby> the designer, liz turner owns the copyright

11:58:24 <eaon> so are they available anywhere?

11:59:00 <libby> they are project tshirts, we can;t sell them, only give them away

11:59:24 <eaon> oh i see :)

11:59:47 * sh1m grins

11:59:53 <sh1m> be really nice to people michael then

12:00:00 <libby> we need to come up with some criteria for giving away.... :)

12:00:12 <fidothe> batting eyelashes?

12:00:18 <sh1m> lol right

12:00:31 <danbri> yeah, I'm wondering about some kind of fun competition that developved the decision to the community a bit...

12:00:44 <danbri> like people submit cool SemWeb-related hacks, and others vote

12:00:51 <libby> ouch

12:00:57 <danbri> we got a bunch anyway but they'll soon go

12:00:59 <libby> sounds a bit competitive

12:01:01 <sh1m> danbri: you see that why we all like you, you didn't even consider bribery

12:01:23 <danbri> heh, what do you think foaf:tipjar is for? ;)

12:01:27 <danbri> .google foaf:tipjar

12:01:27 <datum> foaf:tipjar: http://rdfweb.org/mt/foaflog/archives/2004/02/12/20.07.32/

12:01:42 <danbri> competitive, yeah, i wouldn't want it to be...

12:02:12 <danbri> tricky, in a competition! its just to set some limit and to use it to get people to announce stuff they've done that is cool...

12:02:18 * libby was wondering about getting people to write about what they've done, more for collecting information about what's out there than as a competition/CV

12:02:33 <danbri> yep, bestwriteup not besthack

12:02:45 <libby> hm, we've thinking along the same lines anyway

12:02:48 <danbri> s/best/first100/

12:02:51 <dajobe> "Here's my 200M triple store in 20 lines of python"

12:02:52 * sh1m came, thought, wrote, deligated.

12:02:54 <sh1m> fin.

12:03:37 <LeoBard> @dajobe: does it do owl-dl, too ? kewl!

12:04:08 <sh1m> libby: oh yeah you know how I was raving about rdf author yesterday?

12:04:12 <danbri> yes, and if you've got CSS support it blue-screens if you make instance-level assertions about classes and properties

12:04:22 <libby> oh aye

12:04:23 <danbri> hi kota

12:04:36 <kota> hi danbri

12:04:42 <sh1m> i got home, downloaded the jar file on my linux machine typed java -jar RDf...jar and it worked.

12:04:54 <sh1m> you would have paid to have seen my face.

12:04:54 <sh1m> :)

12:05:05 <libby> yay! it's not fully functional though

12:05:16 <sh1m> do you know what doesnt work?

12:05:28 <libby> hmmm I'll ask shellac

12:05:29 <dajobe> if it was OSX or a set-up desktop, you can just click on it

12:05:51 <danbri> saving?

12:05:57 <sh1m> dajobe: the OS X one did indeed work as it said on the tin.

12:06:32 <dajobe> saving thought. it's an app, click on it -> works

12:07:22 * sh1m waves

12:07:30 * shellac waves back

12:07:59 <dajobe>http://www.semanticwave.com/blog/

12:08:00 <dc_rdfig> G: http://www.semanticwave.com/blog/ from dajobe

12:08:04 <shellac> you wanted to ask about something, libby tells me

12:08:06 <dajobe> G:|Semantic Wave

12:08:06 <dc_rdfig> Titled item G.

12:08:08 <sh1m> someone remind me to buy Damian a beer

12:08:10 <sh1m> ;)

12:08:39 <sh1m> yeah a few things about RDf Author. Libby reckons the java version isn't fully functional and I was wondering what.

12:08:58 <shellac> well a) I hate swing

12:09:08 <sh1m> I also had some UI things in mind but I haven't got around to tinkering with the source yet. It's been a while since I did any _proper_ coding.

12:09:11 <shellac> but b) it needs export added

12:09:20 <shellac> i.e. save as rdf

12:09:36 <danbri> you can get stuff out via copy/paste from text-area, though?

12:09:46 <shellac> probably

12:09:53 <sh1m> RDF/XML and N-Triples?

12:09:56 <sh1m> oh right I see

12:09:58 <shellac> yep

12:09:59 <sh1m> yeah

12:10:13 <Arnia> Hmm... is there any work being done on devising some sort of RDF-using language ala FABL (http://www.fabl.net)

12:10:28 <shellac> to explain - I wrote an application class, similar to NSApplication in OS X

12:10:34 <shellac> to ease porting

12:10:57 <shellac> but I didn't add the export/save as bit

12:11:17 <shellac> so it's that framework that needs fixing, really

12:12:23 <danbri> FABL does look interesting. Chris Goad talked me through it when some of us visited Tijuana ( http://danbri.org/archives/2004/02/13/17.44.54/) at end of ETCON last month...

12:12:49 <shellac> see the world with the semantic web :-)

12:13:08 <danbri> yeah...

12:13:52 <dajobe> and deliver t-shirts?

12:13:56 * danbri has a week or so to finish learning Katakana before -> japan again

12:14:29 <sh1m> shellac: the one thing I would really like is to be able to search for properties or maybe press a key and jump to letters.

12:14:45 <libby> I thought it did that...

12:14:46 <sh1m> the problem is the schemas a huge and scrolling through a massive list is a PITA

12:14:55 <sh1m> hmm not when I last tried

12:15:15 <shellac> oh - in the schema window?

12:15:27 <shellac> good idea

12:15:49 * sh1m grins

12:15:56 <shellac> how big are your schemas?

12:16:01 <sh1m> I work in a usability lab, some of it rubs off

12:16:10 <sh1m> just thinking of my inept foaf knowledge actually ;)

12:16:51 <jbreslin> jbreslin is now known as jb_away

12:16:56 <shellac> afraid I'm just about to go out

12:17:03 <sh1m> dont worry about it

12:17:10 <sh1m> I am around

12:17:11 <shellac> but that should be easy

12:17:16 <sh1m> thanks for comming to chat though :)

12:19:11 <sh1m> a-z sorting too perhaps

12:19:51 <shellac> sorting in RDF is always an issue :-)

12:20:24 <sh1m> not of the schemas props

12:20:29 <sh1m> hehe

12:20:42 <shellac> ok, danbri getting pushy - must go...

12:20:50 <shellac> talk later...

12:20:52 <sh1m> tell him not to whip you too hard

12:20:53 <sh1m> ciao

12:28:08 <eaon> eaon is now known as eaon|out

12:44:59 <fidothe> fidothe is now known as fidothe_atlunch

14:01:33 <ndw> ndw is now known as docbook

14:02:12 <docbook> docbook is now known as ndw

14:04:11 <dajobe>http://jodi.ecs.soton.ac.uk/Articles/v04/i04/Veltman/

14:04:11 <dc_rdfig> H: http://jodi.ecs.soton.ac.uk/Articles/v04/i04/Veltman/ from dajobe

14:04:21 <dajobe> H:|Towards a Semantic Web for Culture, Kim H. Veltman

14:04:21 <dc_rdfig> Titled item H.

14:04:27 <dajobe> H:*87* page PDF

14:04:27 <dc_rdfig> Added comment H1.

14:04:35 <dajobe> H:any volunteers to summarise?

14:04:35 <dc_rdfig> Added comment H2.

14:04:54 <dajobe> H:it's got it all, semweb wedding cake, layers, Sowa ...

14:04:54 <dc_rdfig> Added comment H3.

14:05:07 <JibberJim> Er, lots of pages of PDF about the sem web?

14:05:15 <JibberJim> or do you want a slightly longer summary?

14:05:22 <dajobe> what?

14:05:26 <Morbus> i'll get to reading it in about 4 months.

14:05:31 <Morbus> i'll get back to ya then ;)

14:05:49 <Morbus> it's .. .. #37 in my reading list.

14:09:51 <dajobe> H:the [http://jodi.ecs.soton.ac.uk/Articles/v04/i04/editorial/|issue editorial] has a paragraph summary *He argues that it is based on a restricted and static sense of meaning.*

14:09:52 <dc_rdfig> Added comment H4.

14:10:11 <dajobe> H4:the [http://jodi.ecs.soton.ac.uk/Articles/v04/i04/editorial/|issue editorial] has a paragraph summary *He argues that {the semantic web} is based on a restricted and static sense of meaning.*

14:10:12 <dc_rdfig> Replaced comment H4.

14:17:16 <fidothe_atlunch> fidothe_atlunch is now known as fidothe

14:35:41 <sh1m> any maths buffs around?

14:43:39 <sh1m> ok I wanted to know the name of the backwards E used to denote predicate of classes in peano notation

14:43:44 <sh1m> any ideas?

14:44:09 <DanC> hmm

14:44:18 <Morbus> Existence.

14:44:21 <Morbus>http://members.aol.com/jeff570/set.html

14:44:21 <dc_rdfig> I: http://members.aol.com/jeff570/set.html from Morbus

14:44:54 <sh1m> yeah I got that, but is 'existance' that actual name then?

14:44:58 <Morbus> I:|Earliest Uses of Symbols of Set Theory and Logic

14:44:58 <dc_rdfig> Titled item I.

14:45:23 * sh1m has been chewing on a book on category theory this morning

14:45:27 <DanC> I can't recall a name for the backwards E, other than backwards E. It's read as "there exists..."

14:45:43 <sh1m> ok :)

14:45:45 <DanC> category theory... breakfast of champions. :)

14:46:07 * DanC really likes suber's class notes on math/logic/set theory

14:46:09 <sh1m> i haven't touched anything this interesting since I did my futher maths a-level

14:46:15 <DanC> .google peter suber set theory

14:46:16 <datum> peter suber set theory: http://www.earlham.edu/~peters/writing/infapp.htm

14:46:24 <libby> "existance operator"?

14:46:39 <libby> !l2k yo

14:46:40 <wh4> ヨ

14:46:44 <DanC>http://www.earlham.edu/~peters/writing/infapp.htm

14:46:44 <dc_rdfig> J: http://www.earlham.edu/~peters/writing/infapp.htm from DanC

14:46:52 <tav> tav is now known as tav|offline

14:46:58 <DanC> J:|A Crash Course in the Mathematics Of Infinite Sets

14:46:59 <dc_rdfig> Titled item J.

14:47:20 <DanC> J:a favorite of mine.

14:47:20 <dc_rdfig> Added comment J1.

14:48:43 <DanC> J:Suber writes at exactly the speed I like to read.

14:48:43 <dc_rdfig> Added comment J2.

14:49:18 <Morbus> i love suber.

14:49:23 <Morbus> he was an ampheta user years ago.

14:49:31 <DanC> J:hmm... seems to be an appendix to [http://www.earlham.edu/~peters/writing/infinity.htm|an essay]. hope to read that sometime.

14:49:31 <dc_rdfig> Added comment J3.

14:49:33 <Morbus> and he did the great Nomic and Logical Rudeness stuff.

14:50:39 <sh1m> Logical rudeness? that sounds interesting

14:50:57 <Morbus> http://www.earlham.edu/~peters/writing/rudeness.htm

14:55:39 <sh1m> curious.

14:56:09 <sh1m> Reminds me of a disucssion I was having with a budist and a hippie (both of whom I work with) a few days ago.

14:57:46 <DanC> .time jst

14:57:46 <datum> Fri, 26 Mar 2004 23:57:47 JST

15:19:21 <eaon|out> eaon|out is now known as eaon

15:34:51 <bengee> sh1m, still looking for the foaf:name of the backwards E?

15:35:11 <sh1m> bengee: the foaf:name ?

15:35:20 <bengee> ;-)

15:35:35 <sh1m> As oposed to the maths name?

15:35:36 <bengee> the maths:name then?

15:35:38 <bengee> ah

15:36:52 <bengee> the german term is "existenzquantor" which translates to "existential quantifier"

15:37:07 <danbri> yep

15:38:49 <bengee> now we know that danbri knows german. so the ml spam from a couple of days ago _was_ yours! ;-)

15:39:19 <danbri> i wish

15:39:59 <shellac> hey, so 'Existenz' wasn't just a wacky name, but German?

15:41:43 <bengee> you mean the movie?

15:42:04 <shellac> yep

15:42:35 <shellac> codepoint 2203 would have been geekier...

15:43:08 <fidothe> but what if they didn't have unicode in the future?

15:43:29 <fidothe> 2203 might have been something lame like lowercase A or something

15:49:34 <sh1m> fidothe: yet another question I have :) any mac people know how to make the character map appear. I heard it was in the 'input' menu after some googling

15:50:08 <fidothe> erm, am i obviously a mac user - i thought i'd been washing my hands enough...

15:50:12 <shellac> I do - do you have a flag on your menu bar?

15:50:23 <sh1m> fidothe: nah but you used the unicode issue

15:50:35 <fidothe> shim: it's in the country flag menu bar item

15:50:41 <sh1m> shellac: nope, i type in british and bad american ;)

15:50:41 <sh1m> oh right ok

15:50:54 <fidothe> shellac: oh, cool - you beat me to it

15:51:48 <sh1m> oooh 'Japanese Kana Pallette' sounds much more interesting that what I was doing.

15:52:14 <fidothe> thinking about (hello, fidothe's in serious work avoidance mode) there's a character set version of the Office's fanny-means-butt keith/dawn joke involving MacRoman and iso-8859-1 out there somehwere

15:52:29 <shellac> i recommend the character palette - browse unicode for hours :)

15:52:41 <danbri> its via one of the preference panels (he says, helpfully)

15:53:00 <sh1m> is ? in unicode?

15:53:25 <sh1m> thats the apple apple is irc broke it

15:53:30 <sh1m> s/is/if/

15:54:19 <fidothe> i'm not sure if the apple-apple is in unicode - there's probably a reserved slot for platform-specific oddities like it though

16:01:06 <fidothe> looking at the charts, apple-apple is probably in the range UF0000 ? UFFFFD

16:01:29 <fidothe> Supplementary Private Use Area-A

16:05:36 <sh1m> I dunno, its alt+shift+k

16:05:39 <sh1m> ;)

16:06:13 <fidothe> you know they're holding the 25th unicode conference this year

16:09:13 <sh1m> well that was excessively rubbish

16:09:21 <fidothe> ?

16:09:34 <sh1m> never use apple+q to quit the character pallette

16:11:14 * fidothe falls off his chair

16:11:20 <fidothe> lol - nice

17:01:08 <fidothe> dajobe: can i ask you a using-redland question?

17:13:37 <fidothe> dajobe: is there any easy way to follow a chain of arcs from one node to its end?

17:13:53 <dajobe> not really

17:14:00 <fidothe> *grins*

17:14:24 <fidothe> just an iterative query then

17:14:39 <fidothe> cool - that's fairly easy

17:14:46 <dajobe> yeah, a walk along the path

17:14:54 <dajobe> there are rdf query/path languages around

17:16:12 <fidothe> like squishQL and stuff

17:16:48 <dajobe> that's not a path ql as such

17:16:55 <dajobe> you have to specify all the triples you expect

17:18:10 <fidothe> okay - where do I look for path query languages?

17:18:36 <dajobe> it'll probably divert you from getting the job done

17:18:51 <dajobe> there are a few - versa, rxpath, rdfpath (maybe)

17:19:09 <fidothe> probably, but i've got to run to catch a train in 5, so i'm already distracted ;)

17:20:26 <fidothe> hmmm, interesting: thanks very much

17:20:32 <fidothe> right, got to run

17:20:41 <dajobe> ciao

17:22:01 <sh1mmer> bye

18:08:43 <Bader> hello

18:09:13 <Bader> i'm sarching for an OWL API, but I doen't know which one to use ? any ideas ?

18:11:28 <alk> what choices did you find?

18:12:16 <Bader> alk: OWL API and jena2, but there are in java.

18:13:14 <alk> but? that makes them interoperable at least with C/C++.

18:13:27 <alk> via gcj

18:13:31 <Bader> in fact i want to port their api in python :)

18:13:51 <alk> pretend they are C++ then, use gcj, and commence to port

18:14:51 <alk> i know of no python owl api, but there is pyrple

18:15:43 <alk>http://infomesh.net/pyrple/

18:15:44 <dc_rdfig> K: http://infomesh.net/pyrple/ from alk

18:15:55 <alk> in case it is of use 2 you

18:16:04 <Bader> thanks

19:18:42 <ndw_> ndw_ is now known as ndw

19:55:38 * ndw suffers from a thinko

19:56:01 <ndw> n3 has a comment character, yes?

19:56:46 <ndw> #, i guess, but only at the beginning of a line.

19:56:46 * ndw blinks

19:58:15 * bryceb likes that term "thinko"

20:01:14 <ndw> Can I have an RDF property that has as its value a list of property/value pairs

20:01:14 <ndw> my:property (my:first "Norman", my:last "Walsh)...yeah, I know that specific syntax doesn't work :-)

20:06:30 <Arnia22> Arnia22 is now known as Arnia

20:06:57 <Reto> you could have a property with a collection of reified stmts as value

20:08:28 <Reto> onmitting the rdf:subjets properties of the statements

20:08:59 <ndw> Hmmm.

20:08:59 <ndw> A colleague is trying to bend RDF to support DTD-like rules and I don't think it's going to work

20:09:43 <Reto> me neiter

20:09:44 <ndw> soccos, for example, you can have a Name class that has First and Last properties, that's easy

20:09:52 <ndw> (Sorry soccos, I meant "so, for...")

20:10:11 <ndw> But there's no way to express that First and Last are ordered...because they just aren't in RDF. Right?

20:11:55 <Reto> right. but you could add a sequence of properties to the class

20:12:56 <Reto> so serializers would know how to order the properties of instances

20:23:37 <ndw> Sorry, whatever explanation there might have been was lost as VPN fell over

20:24:13 * ndw goes off to find the logs

20:24:48 <ndw> What does it mean to "add a sequence of properties to a class"?

20:26:59 <alk> which part is unclear?

20:27:04 <alk> what is the context?

20:28:10 <ndw> I was trying to help a colleague struggling to do this illegal thing:

20:28:26 <ndw> my:name = (my:First "Norman", my:Last "Walsh")

20:30:14 <dajobe> you want a collection [set] of property/value pairs

20:30:17 <ndw> It's easy to make a class with two properties, the question then was, is there any way to constrain the "order" of the properties

20:30:17 <ndw> The idea is to provide some schema-like validation, which I think is basically not possible, but I thought I'd mention it here before giving up

20:30:33 <sbp> could do: ?person my:name [ rdf:first [ my:First "Norman" ]; rdf:rest [ rdf:first [ my:Last "Walsh ]; rdf:last rdf:nil ] ] .

20:30:44 <dajobe> ETOOLONG

20:30:49 <sbp> yeah

20:31:11 <sbp> s/rdf:last/rdf:rest/

20:31:22 <dajobe> ( [ my:First "Norman"] [ my:Last "Walsh"] ) a my:name .

20:31:37 * dajobe making things up

20:31:48 <sbp> same thing, more or less

20:32:03 <ndw> Ah

20:32:33 <dajobe> that's an ordered collection, if that matters

20:33:32 <sbp> but really, if you're using my:name and you can declare its range yourself, you can do: my:name ("Norman" "Walsh") . and then do some trickery such that { ?person my:name (?first ?last) } => { ?person my:First ?first; my:Last ?last } .

20:34:32 <ndw> Interesting.

20:35:38 <sbp> and vice versa, actually, constructing the list from the properties (if they've got a cardinality restriction of 1 on them)

20:36:57 <DanC> constraining what you say, rather than what you mean, is hard in RDF. it's much easier in XML.

20:37:47 <ndw> Yes. I don't think this approach is going to work, but we'll see.

20:38:02 <DanC> hendler just gave an example earlier today. in owl, you can say "every contract has a purchase order number." Then fred says "here's a contract" and bob says "what's the purchase order number?" and fred says "I dunno. it's got one, but I dunno what it is."

20:38:29 * ndw chuckles

20:39:50 <DanC> meanwhile, in N3, you can say "every contract *document* says what its purchase order number is."

20:40:21 <ndw> How?

20:40:28 <DanC> { ?D a :Contract } => { ?D log:semantics [ log:includes { ?D :poNum _:N } ] }.

20:40:58 * ndw 's parser falls off the rails part way through

20:41:39 * DanC hunts for relevant part of cwm/n3 tutorial...

20:41:43 <ndw> What do log:semantics and log:includes mean?

20:41:53 <ndw> And what does _:N map to/from?

20:42:05 <DanC> "Reaching out onto the Web" http://www.w3.org/2000/10/swap/doc/Reach

20:42:20 <DanC> _:N is an existentially quantified variable. read: "for some N"

20:42:38 * ndw goes to read

20:42:56 <DanC> log:semantics relates a document to the formula expressed by that document. (i.e. it's HTTP GET + rdf parse)

20:43:43 <DanC> log:includes is rdf simple entailment. e.g. { :book :author :dan; :date 1990 } log:includes { :book :date 1990 } .

20:44:21 <DanC> strictly speaking, the syntax I used for ?D won't work above, since I use ?D at two different levels of {} nesting. I'd have to say this log:forAll :D.

20:44:50 <DanC> and in stead of _:N I could have just written []. i.e. "something". but I thought _:N was more mnemonic.

20:45:11 <DanC> s/the formula/a formula/

20:45:50 * ndw mutters something about turtles all the way down, finds his perplexed face, and bolts it on

20:48:51 <ndw> Hmm. There's more N3 syntax than I know about.

20:48:57 <ndw> I find {"MA"^state:code.state:borderstate^city:state city:name ?n} totally opaque

20:49:12 <DanC> yeah, that's a punctuation nightmare.

20:49:52 <ndw> There's a bug in http://www.w3.org/2000/10/swap/doc/Reach

20:49:59 <DanC> but did you get the InvalidRequest example?

20:50:01 <ndw> After the "Albany", "Amherst", etc. examples,

20:50:12 <ndw> and so on.

20:50:12 <ndw> To do more fancy things, you can run cwm in a mode which loads the

20:50:12 <ndw> By now yo

20:50:27 <ndw> "loads the" ... there's text missing there

20:50:49 <DanC> ah... yes, you're supposed to read timbl's mind at that point in the text. ;-)

20:51:00 <ndw> Ah

20:51:16 <ndw> I often find that my brain doesn't interface very directly with Tim's

20:51:39 <ndw> Perhaps replace the white space with "log:mindReading" :-)

20:52:17 <DanC> ok, enough poking fun at the typos... did the explanation of log:semantics and log:includes in the top half make sense?

20:52:17 <ndw> No, I don't really get the InvalidRequest example. I find that syntax a bit opaque too

20:52:51 <DanC> oh... he used path syntax there. hmm... something of a distraction

20:52:52 <ndw> I'm not sure how to interpret "<input.data>.log:semantics"

20:53:10 <DanC> for "<input.data>.log:semantics" read [ is log:semantics of <input.data> ]

20:53:30 <DanC> you can think: log:semantics(<input.data>)

20:53:55 <sbp> open('input.data').logSemantics()

20:54:31 <ndw> Ok, so like it says, take the input data, the axioms, and the rules and think about them, producing conclusion :f

20:54:39 <DanC> yes...

20:54:42 <ndw> Then what does :f log:conclusion :g . do?

20:54:53 <DanC> er... :f is before thinking. :g is after.

20:55:10 <DanC> :f is just the conjuction.

20:55:15 <ndw> Uhm.

20:55:32 <ndw> (input axioms rules) log:conclusion :f .

20:55:40 <ndw> How is :f a conjunction?

20:56:04 <ndw> Sorry. read the word wrong. It says "log:conJUNCTion :f."

20:56:15 <DanC> ( { :book :author :dan } { :book :price 20 } ) log:conjunction { :book :author :dan; :price 20}.

20:56:29 <DanC> yes... conJUNCTion

20:56:36 <ndw> Ok.

20:56:44 <ndw> So take stuff and form the conjunction :f.

20:56:47 <DanC> yes

20:56:54 <ndw> From :f conclude :g

20:56:58 <DanC> yes

20:57:16 <ndw> If :g doesn't include :request a :ValidRequest, that implies that :request a InvalidRequest

20:57:22 <DanC> yes

20:57:30 <ndw> Spiffy

20:57:35 <DanC> whew.

20:58:19 * DanC had the same conversation with timbl probably 2 years ago... no longer remembers what it's like to not understand it.

20:58:44 * ndw is happy to have finally had his very first RDF question answered.

20:58:52 <ndw> That question was "how do I express 'not'" :-)

20:59:10 <ndw> Well, maybe that wasn't my *first* question... :-)

20:59:19 <DanC> ah... I think I remember that. I think I gave you a pointer to this same document at the time. takes a while to absorb, I suppose.

20:59:35 <DanC> or maybe I gave you a pointer to a test case back then, with no prose around it.

20:59:36 <ndw> You might very well have done so. And I probably didn't grok a word of it

20:59:58 <ndw> I don't remember.

21:00:18 <ndw> I have a couple of "nots" that I want to express in norman.walsh.name, so this is handy timing

21:00:54 <ndw> (If an essay doesn't have :talkback "no", assume :talkback "yes" and something similar for CSS formatting)

21:00:56 * DanC looks forward to the relevant blarticle

21:01:38 <ndw> I've apparently crossed some magic threshold in cwm, btw. Grinding over my .palm.rdf went from 5 minutes to 25 minutes at some point when I wasn't (alas) looking

21:01:52 <DanC> ew. ouch.

21:01:59 <sbp> .googlecount blarticle

21:02:03 * ndw shrugs

21:02:03 <datum> blarticle: 37

21:02:37 <ndw> "Did you mean: partical"?

21:02:43 <ndw> Uh, particle even

21:02:50 <DanC> no, I meant weblog article

21:03:12 <ndw> Yes, I deduced that. I just went and did the google that sbp googlecounted

21:03:17 <DanC> oh. duh.

21:03:49 <ndw> I decided to call them "essays". I didn't think "article" really worked.

21:04:36 <DanC> hmm... I think of essays as persuasive, exclusive of expository writing.

21:04:52 <ndw> hmm

21:05:09 <sbp> the original sense of essay means to try out or explore an idea, from French assayer or somesuch

21:05:23 <ndw> Oh, that sounds like a win

21:06:47 * DanC tries out "blessay"... nope.

21:07:35 <DanC> sbp, where did "weblog" originate?

21:08:43 <sbp> Jorn. "In 1998 there were just a handful of sites of the type that are now identified as weblogs (so named by Jorn Barger in December 1997)." - http://www.rebeccablood.net/essays/weblog_history.html

21:08:55 <Morbus> i started mine in 98.

21:09:21 <sbp> when did you first start calling it a "weblog", though, Morbus?

21:09:36 <Morbus> probably around the time i blogger-enabled it.

21:09:39 <Morbus> don't know when that was though.

21:09:47 <Morbus> i know i wasn't calling it that in 98 ;)

21:10:28 * Morbus searches archives for first meniton.

21:10:39 <verbosus> I know the etymology of "blog", but not "weblog"

21:10:45 <DanC> hmm... extreme markup deadline is today.

21:11:01 <ndw> not for papers, for tutorials and peer review apps

21:11:20 <Morbus> sbp: i mentioned weblog 2005/05.

21:11:38 <ndw> You will, Morbus? :-)

21:11:39 <Morbus> that was the earliest, but it was in reference to another site.

21:11:42 <Morbus> er. heh, heh.

21:11:44 * DanC wonders if exml is on ndw's radar... indeed... http://norman.walsh.name/2004/03/26/extreme

21:11:47 <Morbus> 2000.

21:12:45 <ndw> Yes, Danc, I plan to write about DocBook in RELAX NG and blah blah blah. By 16 Apr. Gulp.

21:13:29 * DanC finds it confusing to think of "paper" as a class dijoint from "peer reviewed submission"

21:13:50 * DanC swats away a cookie from the extreme web site. sigh.

21:14:06 <sbp> I can just see hack philologists in 100 years' time explaining that "weblog" is formed from "we" and "blog", and that the origin of the particle "blog" is obscure

21:14:13 <DanC> chuckle... "Do you FO?"

21:14:40 * ndw chuckles at sbp

21:15:40 <DanC> oh! it's not _submissions_ that are due today, but "can I be on the program committee?" requests.

21:21:44 <ndw> I thought I said that, I said "apps" where it probably would have been clearer to say "applications". But maybe not

21:22:46 <DanC> I got confused long before you wrote "apps"

21:22:56 <DanC> my office is hot and stuffy. hmm.

21:23:36 * ndw had the windows *open* today! For the first time since...way too long

21:24:38 <JibberJim> It's been very nice here in the UK for the first time in ages too!

21:25:30 <ndw> What does "this log:forAll :car ." mean?

21:25:30 <Arnia> The light burns my eyes

21:25:34 <ndw> lol

21:27:14 <sbp> this log:forAll :car . is a CWM hack to make :car a (universally quantified) variable

21:27:33 <sbp> so for: { this log:forAll :car . :car ... } read { ?car ... }

21:27:57 <ndw> alk, ok. but the this form is needed for certain nesting cases, I gather

21:28:28 <sbp> yes, but it's to be replaced with "@forAll car."

21:30:06 * sbp gets his big figurative "[DEPRECATED]" rubber stamp out

21:31:07 <ndw> And what's the dot notation?

21:31:12 <ndw> :car.auto:specification mean?

21:31:14 <sbp> insane

21:31:23 <sbp> [ is auto:specification of :car ]

21:32:13 * dajobe has a different rubber stamp for that one

21:32:17 <yonderboy> B:|Apache, Axis and the HTTP 1.1 Keep-Alive header

21:32:20 <dc_rdfig> Titled item B.

21:32:28 <ndw> And, as long as I'm admiting the depth of my ignorance...

21:32:39 <sbp> it's a traversal. so just as in OOP usually instance.property evaluates to the property of the instance, :car.auto:specification gets the auto:specification of that :car

21:33:02 <ndw> [ is auto:specification of :car ] is, um, a bnode that is the value of the auto:specification property of :car?

21:33:18 <sbp> right

21:35:54 <ndw> So...I'm still confused. In this example, it looks like auto:color is a property of the auto:specification property. Is that really what's intended?

21:35:56 <ndw> { :car.auto:specification log:notIncludes {:car auto:color []}}

21:36:24 <ndw> s/is a property of/is a property of the value of/ I guess

21:37:21 <sbp> apparently, :car.auto:specification returns a formula

21:37:23 <sbp> a formula being a set of triples in N3

21:37:29 <sbp> so roughly analogous to a document: it's the information in that car's specification as a set of RDF triples

21:37:58 <sbp> and what that whole this is saying is that the car's specification does not include any information about the car's colo[u]r

21:38:00 <ndw> Hmmm. So can that be rewritten for the simple case where auto:color is a direct property of :car?

21:38:22 <sbp> I'm not sure what you mean

21:39:18 <ndw> Rather than having this indirection through the formula returned by auto:specification, I simply want to deal with properties directly on :car

21:39:54 <ndw> I want a rule that says if :car does not have a auto:color property, then imply something

21:39:59 <sbp> ah. you want to make a rule checking that the current document does not have a triple stating the colo[u]r of the car?

21:40:02 <sbp> okay

21:40:10 <ndw> right

21:40:11 <sbp> then use "this" instead of :car.auto:specification

21:40:33 <sbp> hmm. no

21:40:45 <sbp> that's going to match the formula inside that rule... that's irritating

21:41:08 * ndw is out of his depth

21:41:23 <sbp> I guess <>.log:semantics would be what you're after, though DanC or timbl would know for sure

21:41:53 <sbp> i.e. { <>.log:semantics log:notIncludes {:car auto:color []}}

21:42:34 <DanC> you can't make a rule that "if a car does not have a color property...". you can only make one that says "if formula X doesn't give a color for a car, then..."

21:42:45 <ndw> I can't see how that "iterates over" all the log:forAll :cars...but I'm not saying it doesn't :-)

21:44:32 <ndw> Ah

21:44:51 <DanC> (actually, you could make a rule "if a car does not have a color property..." but you'd be hard pressed to convince cwm that it didn't.)

21:45:21 <ndw> Ok.

21:45:44 <ndw> So I have a bunch of n3 statements in a file, does that file constitute a formula?

21:45:54 <ndw> I.e. can I ask if a color is not defined in that file?

21:45:55 <DanC> the log:semantics of that file is a formula, yes.

21:46:04 <DanC> in that file, yes.

21:46:48 <sbp> that's what I was supposing you meant; <>.log:semantics means "the formula of this file"

21:47:09 * ndw struggles to wrap his head around this

21:48:06 <DanC> wow... I just got 404 report about the "Gentle Introduction to SGML" link from http://www.w3.org/MarkUp/SGML/

21:49:25 <sbp> DanC: http://web.archive.org/web/19990117033936/http://www-tei.uic.edu/orgs/tei/sgml/teip3sg/SG.htm

21:49:57 <DanC> it seems to have a new home at http://www-sul.stanford.edu/tools/tutorials/html2.0/gentle.html

21:50:06 <sbp> ugh

21:50:18 <DanC> and http://www.isgmlug.org/sgmlhelp/g-index.htm

21:50:24 <ndw> Trying to copy the log:conjunction example we spent so much time talking about...

21:50:28 <ndw> { ( </share/websites/norman/knows/topics.n3>.log:semantics ) log:conjunction :f .

21:50:28 <ndw> :f log:conclusion :g .

21:50:28 <ndw> :g log:notIncludes { t:talkback "no" }

21:50:28 <ndw> } => { t:talkback "yes" } .

21:50:35 <ndw> that goes bang badly in cwm...because?

21:51:15 <sbp> { t:talkback "no" } is not valid syntax

21:51:26 <sbp> 't:talkback "no"' is not a tripl

21:51:31 <ndw> uh right

21:51:34 <DanC> needs a subject.

21:51:37 <sbp> ...e. (same with talkback yes)

21:52:19 <sbp> also, there's not much point in finding the conjunction of one formula. perhaps you're planning to add more later, though

21:52:27 <DanC> are there rules in topics.n3? if not, you don't need to use log:conclusion. and if topics.n3 is the only source of data you're consulting, you don't need to form a conjunction

21:52:52 <DanC> jynx

21:54:19 <sbp> does jynx really apply when sbp.statement C DanC.statement? :-)

21:55:07 <ndw> No, there are no rules and I won't be adding any more files

21:55:27 <ndw> But I can't figure out how to add a variable to range over all the resources defined in topincs.n3 for the talkback tests

21:56:12 <DanC> outside all the {}s, put: this log:forAll :ESSAY.

21:56:30 <DanC> hmm... then...

21:57:01 <ndw> I tried effectively that. It still goes bang.

21:57:09 <ndw> ... assert not isinstance(pred, Formula) or pred.canonical is pred, "pred Should be closed"+`pred`

21:57:09 <ndw> AssertionError: pred Should be closed0_work

21:57:10 <ndw> if that helps

21:57:37 <DanC> ew. that smells like a cwm bug. but I think there's a straightforward (necessary, in fact) work-around.

21:57:45 <DanC> we need a shared emacs buffer.

21:57:54 <DanC> hmm... maybe esw CwmTips...

21:58:10 <ndw> Uh...

21:59:11 <DanC> I need a pattern I can use to match all essay-resources.

21:59:25 <ndw> a t:Article

21:59:33 <DanC> ok. explicitly?

21:59:38 <sbp> ?ESSAY log:includes { ?s ?p ?o }?

21:59:40 <ndw> (Yeah, I know what I said about essays, but...)

22:00:03 <ndw> Uh, actually the a t:Article part comes from somewhere else

22:00:20 <ndw> You can match all resources in topics.n3?

22:00:40 <ndw> (Unconditionally, I mean, all the resources in there are t:Article s

22:01:41 <DanC> I typed it in http://esw.w3.org/topic/CwmTips under 2.1 Mar 26 fiddling

22:02:41 * sbp "oh"s

22:02:42 <ndw> That goes bang with the thing that smells like a cwm bug

22:02:54 <DanC> hmm... lemme test locally...

22:04:23 <ndw> Uhm.

22:05:02 <ndw> I must not have saved something

22:05:33 <DanC> WorksForMe(tm)

22:06:03 <ndw> Right. It runs, but it produces only the rule on output

22:06:34 <DanC> hmm...

22:06:54 <DanC> reload http://esw.w3.org/topic/CwmTips ? that's straight from my emacs buffer. it works as expected.

22:08:06 * DanC wonders about cwm versions.

22:08:17 <DanC> $Id: cwm.py,v 1.147 2004/03/09 23:07:05 connolly Exp $

22:08:55 <sbp> logger, pointer?

22:08:55 <sbp> See http://www.ilrt.bris.ac.uk/discovery/chatlogs/rdfig/2004-03-26#T22-08-55

22:09:41 * DanC feels like a flashbulb just went off. :)

22:11:01 <sbp> "everybody move in a bit, please. in a bit. bit more. come on, I haven't got my wide-angle lens today... squish together. more!"

22:13:41 * ndw blushes. Forgot --think

22:13:58 <DanC> heh.

22:15:41 <ndw> Ok. Thank you once again. I think I can make it work now.

22:16:04 <DanC> you're winning now... good.

22:16:15 <ndw> I am, yes.

22:18:44 * DanC wonders if that's the same dehora that sends to www-tag on occasion... e.g. http://lists.w3.org/Archives/Public/www-tag/2003Jan/0301.html

22:19:10 <sbp> and that used to frequent the HTML lists many moons ago

22:19:24 <DanC> "eircom.net" in both places.

22:19:36 <sbp> or xml-dev

22:20:04 <sbp> also, apparently, an RDF Core member: http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001May/0037

22:20:33 * ndw remembers not to try to hide his identity around here

22:22:30 <DanC> bill? zat you? or is that some piece of software desparately trying to maintain a connection to the channel in the face of net glitches?

22:22:42 <sbp> it's possible. I've had a handful of people on IRC whose identities have been very difficult to trace (recently, a person was almost taunting us to find his identity, but neither a grep through #wikipedia.enrc logs nor a Google for diff -31 words chanlogs would help; still not 100% sure who it was)

22:24:12 * DanC wonders if sean is engaged in the creative commons review of GRDDL...

22:24:57 <DanC> to wit: http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2004Mar/0014.html

22:25:05 <sbp> nope; I'm not affiliated with Creative Commons. oh crap, and I had been meaning to write up that huge background document on RDF-in-X?HTML--I've got a ~10KB document of notes and links lying around

22:25:44 <DanC> Ben Adida is quite quick. he proposes a solution. I wonder if you'd look it over with me.

22:25:50 <sbp> certainly

22:26:05 <DanC> take a look at 0014 there then

22:26:20 <sbp> I don't think rel="http://web.resource.org/cc/rels/license" is valid

22:26:25 <sbp> NMTOKENS, IIRC

22:27:39 <sbp> ah, no. CDATA according to http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd

22:28:33 <DanC> that's greymarket... but not a show-stopper.

22:30:45 <sbp> ah, it changed in XHTML 1.1: "<!ENTITY % LinkTypes.datatype "NMTOKENS" >" - http://www.w3.org/TR/xhtml-modularization/DTD/xhtml-datatypes-1.mod

22:31:31 <sbp> notable, though; if you don't mind about validation, you could just throw the rdf:RDF in

22:32:34 * ndw calls it a day. Waves g'night!

22:32:44 <DanC> well, just throwing rdf:RDF in assumes HTML is "rdf-transparent"

22:32:46 <DanC> night, ndw

22:33:12 <sbp> "there's only one possible profile for the whole page" in Adida's email is false, too--it's a space separated list of URIs, no? though there appears to be some confusion over that, IIRC

22:33:16 <sbp> 'night ndw

22:33:21 <DanC> that is: a consumer that just hunts for rdf:RDF wherever it finds it is going to lose in many of the XSLT documents I've written.

22:33:42 <sbp> point

22:35:24 * DanC sent interim reply to ben

22:36:20 <DanC> my main issue is whether the N3 extraction he did for that license is OK in general.

22:38:14 <sbp> I doubt it. <a rel="next" href="./page.html">...</a> would come out as { <> <next> <./page.html> }

22:39:10 <DanC> well, one could limit it only to absolute URIs in the rel attribute. (hold your nose!)

22:39:10 <sbp> you'd have to match against some simplified absolute URI regexp, but then you're imposing retroactive semantics on @rel that isn't mandated by the HTML specification suite

22:39:17 <sbp> jynx!

22:40:29 <sbp> I note that HyperRDF uses colons in @rel, but only in html:link: "<link id='c' rel='rel:classes' href='http://www.w3.org/2000/07/hs78#' />" - http://www.w3.org/2000/07/hs78/

22:42:04 <DanC> is it really retroactive? or are we just extending HTML a bit? HyperRDF was in the grey zone; it accepts the risk that rel:classes will be specified to mean something else, to some extent.

22:42:08 <sbp> it's fun, incidentally, reading the pubrixt mail archives: I did the entire lot in one night and found a tremendous change in direction when the chair changed from jmr to dom :-)

22:42:46 <DanC> heh

22:44:20 <sbp> I'm not sure I'd want to be held to having made the legal statement <> <arbitrary:thing> <URI> if I said <a rel="arbitrary:thing" href="URI">...</a> in an HTML document somewhere

22:44:25 <DanC> I suppose extending HTML is missing the whole point. If you're gonna extend HTML, you might as well go whole hog like Birbeck.

22:44:52 <sbp> what'd Birbeck do?

22:45:25 <DanC> he did http://www.w3.org/MarkUp/2004/02/xhtml-rdf.html

22:45:30 <sbp> oh right

22:45:46 <sbp> nice for X2...

22:46:43 <sbp> funny how roughly every five months, someone comes up with a proposal to put meta/link where it wasn't intended, and believes that they've come up with a novel solution

22:47:18 <DanC> the challenge of community memory.

22:47:33 <sbp> heh, yes

22:48:39 * sbp thinks about limiting the @rel hack to terms starting with "http://"

22:49:25 <DanC> ew ew ew.

22:49:59 <sbp> well, ew ew ew whatever you do, really, but I'd feel more comfortable saying that if you use some string starting "http://" then you're talking about an HTTP URI

22:50:43 <DanC> I'm convinced that BenA's idea is not one I want to persue/endorse.

22:51:06 <sbp> i.e. I think that if I wrote "the author of http://example.org/ is a [slander slander slander]" on a mailing list, you could hold me to saying that because I've used that string starting with http://, it's socially obvious that I'm intending it as an HTTP URI

22:51:12 <sbp> okay

22:51:25 <sbp> I don't disagree :-)

23:41:04 <golbeck_> golbeck_ is now known as golbeck

23:50:26 <Meem_> Meem_ is now known as Meem