Semantic Web Interest Group IRC Chat Logs for 2003-08-27

00:04:58 <ndw> ndw is now known as ndw^afk

00:09:13 <danbri>http://www.driverdf.org/articles/DriveandRDFGraphsDiggingi.html

00:09:13 <dc_rdfig> A: http://www.driverdf.org/articles/DriveandRDFGraphsDiggingi.html from danbri

00:09:24 <danbri> A:|Drive and RDF Graphs: Digging in, by Rahul Singh

00:09:25 <dc_rdfig> Titled item A.

00:10:19 <danbri> A:Walks through some sample FOAF-handling code with Drive, a C# RDF parser + API.

00:10:20 <dc_rdfig> Added comment A1.

00:17:57 <JimJibber> JimJibber is now known as JibberJim

00:26:48 <datum> whoops.

01:37:56 <JibberJim> /ping hh

02:01:53 <sniffles_out> sniffles_out is now known as sniffles

02:07:55 * GNUWookie waves

02:09:49 * GabeW waves back

02:10:32 <GNUWookie> milkshakes?

02:10:46 <GNUWookie> :)

02:11:19 <GabeW> its not my machine

02:13:38 <GabeW> its a friend's machine

02:13:56 <GabeW> a virtual friend

02:14:07 * GabeW is repairing my winxp install

02:15:28 * GNUWookie feels somewhat sorry for GabeW, i never installed winxp...

02:15:46 <GabeW> yah, usb stuff has been spotty recently

02:15:54 <GNUWookie> last windows I installed was win2k professional 1 year ago

02:50:04 <GNUWookie> GNUWookie is now known as Wookiee-ZZZZZ

04:03:46 <WillWare> I have been studying danbri's FOAF doc, and near the top I see << xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" >>, and then later I see "rdf:about" used as an XML attribute. I'd expected that 22-rdf-syntax-ns would offer some machine-readable definition of "about", but have not been able to find it. Is it defined elsewhere? Am I suffering some dire misunderstanding of XML namespaces?

04:22:01 <jql> umm... dire misunderstanding

04:22:59 <jql> rdf:Description, rdf:about and rdf:resource are just tags for markup, along with things like xml:lang

04:23:58 <jql> so, since they're markup, they don't have a meaning that is described in the syntax doc

04:37:34 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

06:34:02 <jql> jql is now known as jql-zzz

08:19:37 <cedrickiss> cedrickiss is now known as cedric

08:57:04 <nwetters> morning campers

09:59:01 <jeen> hi libby

09:59:20 <jeen> got my mail?

10:00:12 <libby> yep thanks Jeen, very kind of you :)

10:01:08 <libby> (just in case he comes hack: WillWare, not a dire misunderstanding - many schemas will have RDF at the end of the namespace - but it is not compulsory)

10:01:17 <jeen> ah no trouble, it was 5 minutes work, half of which consisted of switching my environment to jdk 1.3...

10:01:28 <libby> sorry to be a pain

10:01:40 <libby> I can move over to 1.4, maybe I should do.

10:02:04 <jeen> no you're quite right, I think, if it's not *necessary* our code shouldn't be 1.4-only.

10:02:56 <libby> well I was wondering if in 1.4 there's an xml parser built in isn't there?

10:03:17 <jeen> but if you do switch, make sure you get 1.4.2. 1.4.1 has a memory leak in its StringBuffer implementation. Very nasty in combination with parsing large XML files...

10:03:47 <jeen> yeah there is, but we have had some problems with it. we still just use Xerces instead.

10:04:36 <dajobe> hi jeen

10:04:48 <jeen> hello DAve

10:04:55 <jeen> s,A,a

10:05:06 <dajobe> s/j/J/

10:05:10 <jeen> heh

10:05:27 <jeen> oh no, that's a company thing, we do all our names without capitals :)

10:05:42 <dajobe> very egalitarian.

10:05:57 <libby> ok, thanks jeen

10:06:03 <jeen> it looks great on a logo, but is heck if you prepare a powerpoint presentation...

10:06:46 <jeen> libby, if you have any further comments or bugreports on the parser, we'd love to hear them of course :)

10:07:06 <dajobe> you got a RIO user? well done :)

10:07:52 <libby> yep, seems excellent.

10:08:48 <jeen> dajobe, hey now. We got more than one. At least five that I know of :)

10:08:54 <dajobe> w00t

10:09:15 * dajobe hides from HPlabs people

10:12:03 <jeen> heh

10:13:53 <jeen> I think we're seeing their revenge already - sourceforge seems to be unreachable...

10:14:38 <ndw^afk> ndw^afk is now known as ndw

10:19:58 <arnarl> hi

10:44:50 <tav|offline> tav|offline is now known as tav

11:35:30 <JimJibber> JimJibber is now known as JibberJim

12:10:09 <sandro> Arg. No connectivity between DSL.net and W3.org.

12:28:53 <sandro> hello.... hello.... ?

12:29:24 <JibberJim> Hi Sandro, we can see you here...

12:33:04 <danbri> who're you talking to, Jim?

12:33:25 * danbri waves to sandro

12:42:03 <JibberJim> How long before the Image Description chat?

12:43:00 <shellac> 1430 UTC?

12:43:02 <danbri> its 3.30...

12:43:07 <danbri> uk time i mean

12:43:19 <shellac> BST

12:43:23 <Schuyler_> two hours?

12:43:31 <JibberJim> yep 2 hours, cheers.

12:43:42 <libby> .time utc

12:43:59 <libby> hm

12:44:12 <shellac> 12:44pm

12:44:29 <shellac> ask me another...

12:44:32 <libby> :)

12:46:19 <datum> 2003-08-27T12:46:19Z

12:47:18 <JibberJim> see shellac, you were wrong... it was 12:46 not 44 :-)

12:48:26 <shellac> I'm two minutes fast, apparently.

12:49:13 <shellac> faster.

12:49:16 <shellac> doh

12:52:17 <dajobe> logger's always at UTC, you can always ask it

12:52:27 <dajobe> logger, bookmark

12:52:27 <dajobe> See http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T12-52-27

13:09:40 <sandro> So I can reach lots of the web, and lots of the web can reach MIT, but I can't reach MIT. :-(

13:11:49 <bijan> Hmm. can you reach the lots of web that can reach MIT?

13:12:13 <bijan> I mean, do we have a reaching transitivity failure, or are you just not hooked up to the right place?

13:12:27 * bijan hope he's respondin gto a current post rather than logged hallucinations

13:13:03 * ndw wanders off to fire up VPN

13:18:09 <sandro> Yes, Bijan -- I'm talking to people on IRC who can reach MIT. So no, the Internet is not being transitive today.

13:18:25 <bijan> You changed it.

13:18:36 <bijan> From "the web" to "the internet"

13:18:55 <sandro> Yes, and I smacked myself, too.

13:19:32 <bijan> Hmm. if i can reach MIT, I could paste bits of web pages foryou :)

13:19:42 * sandro has a flashback of old dislike for the web. :)

13:19:56 <sandro> Can you CVS commit for me? :)

13:19:58 * bijan has that every day in every way

13:20:06 <bijan> Plus NEW dislikes :)

13:20:09 <sandro> I know, I know.

13:20:13 <bijan> Hmm.

13:20:16 <bijan> No.

13:20:24 <bijan> But i can barely cvs commit for myself

13:21:01 * sandro tries to avoid making a stupid joke about committment.

13:21:09 <sandro> and being committed

13:42:13 <nmg>http://www.ht03.org/papers/pdfs/7.pdf

13:42:14 <dc_rdfig> B: http://www.ht03.org/papers/pdfs/7.pdf from nmg

13:42:36 <nmg> B:| Which Semantic Web? [paper at Hypertext'03]

13:42:36 <dc_rdfig> Titled item B.

13:43:50 <nmg> B: an examination of the promises made about the SW in the research literature - slightly critical of the SW in places

13:43:50 <dc_rdfig> Added comment B1.

13:59:41 <shellac_> shellac_ is now known as shellac

14:12:02 <danbri_dna>http://lists.w3.org/Archives/Public/www-rdf-interest/2003Aug/0129.html

14:12:02 <dc_rdfig> C: http://lists.w3.org/Archives/Public/www-rdf-interest/2003Aug/0129.html from danbri_dna

14:12:28 <danbri_dna> C:|Image Description 'show and tell' IRC chat

14:12:28 <dc_rdfig> Titled item C.

14:12:50 <danbri_dna> C:Here, now(ish; 2.30UTC), for 1hr.

14:13:42 <danbri_dna> C:Timezone conversions: [http://www.timeanddate.com/worldclock/fixedtime.html?day=27&month=8&year=2003&hour=14&min=30&sec=0&p1=0|2003-08-27 2.30UTC].

14:14:02 <dc_rdfig> Added comment C1.

14:14:03 <dc_rdfig> Added comment C2.

14:14:21 <danbri_dna> C:Basic idea is show'n'tell for folk building image metadata apps, esp. w.r.t RDF+SVG

14:14:21 <dc_rdfig> Added comment C3.

14:14:52 <danbri_dna> C:See [http://esw.w3.org/topic/ImageDescription|ImageDescription] in the ESW Wiki for more context, link to prev chat logs (re Exim) and draft agenda.

14:14:52 <dc_rdfig> Added comment C4.

14:15:14 <danbri_dna> C:Agenda summary: 45 mins of show/tell, 15 mins discussion and followup.

14:15:14 <dc_rdfig> Added comment C5.

14:15:43 * sandro is very torn about attending, considers image work a hobby, not his job right now.

14:15:50 * DanC is very torn too

14:16:43 <danbri_dna> there'll be other times...

14:21:26 <GregElin> Hello

14:21:33 <JibberJim> Hi Greg

14:21:34 <gen> hi

14:21:39 <GregElin> Hi Gen.

14:21:42 <tav> tav is now known as tav|offline

14:23:01 <danbri_dna> hi all

14:23:08 <mortenf> 'lo

14:23:09 <gen> hi

14:23:10 <danbri_dna> see http://rdfig.xmlhack.com/ for links to agenda, context etc

14:23:43 <_joshua> Hey Greg

14:24:22 <GregElin> Hey Josh. Glad you could make it. Will you be willing to say a few words re geoURL? Seems relevant.

14:25:00 <danbri_dna> are people using geourl markup to tag locations?

14:25:05 <danbri_dna> of images i mean...

14:25:27 <danbri_dna> hi jo

14:25:35 <zool> re danbri

14:25:57 <ndw_> ndw_ is now known as ndw

14:26:01 <ircleuser> Hey Greg, I'd be happy to talk about the NewsML initiative within Reuters and other interests

14:26:12 * ndw turns his attention to #rdfig on command of his Palm

14:26:14 <ircleuser> ircleuser is now known as bptheo

14:27:12 * DanC imagines palm viruses, where zillions of people are suddenly scheduled to buy viagra

14:27:17 <GregElin> Hi Brian! That would be great.

14:27:28 <danbri_dna> lol

14:27:37 <ndw> I'm not sure I'd fall for that, DanC

14:27:38 <GregElin> While we wait for others to arrive...why don't we set up a little cue for our show and tell guides?

14:27:49 <danbri_dna> cue or queue?

14:28:11 <GregElin> The tone we are hoping to set today is less beauty contest or deep discussion, but more a review of what is out there, who's doing what.

14:28:18 <ndw> A queue I presume, so we can cue it later

14:28:29 * danbri_dna has a list of would-be showers in the wiki page... anyone to add? I see 'bptheo/Brian on NewsML/reuters'...

14:28:33 <JibberJim> Is there an attendance list?

14:28:34 <GregElin> (Uh...yeah...queue...)

14:28:57 <GregElin> Add Joshua for GeoURL.

14:29:19 <JibberJim> and move me to the end after everyone else, mine's been done...

14:29:24 <GregElin> I don't know if anyone is here from University of maryland, but photofinder is worth mentioning.

14:29:37 * danbri_dna adds to wiki

14:29:54 <JibberJim> Imarea is very interesting, but I've not heard from the folks after I gave them a heads-up.

14:29:58 <gerald> danbri, I could blab about my stuff, not sure if it's on topic

14:29:58 <mortenf> .time

14:29:59 <datum> Wed, 27 Aug 2003 14:29:59 GMT

14:29:59 <_joshua> I'm only 7% here

14:30:10 <mortenf> 'lo marc!

14:30:16 <_joshua> perhaps less

14:30:17 <marccanter> West Coast is in the house

14:30:23 <danbri_dna> change of mind, wiki too fiddly / switching to browser etc., so let's do it in irc/rdfig weblog

14:30:33 <marccanter> thoise of us who HAVEN'T gone to Burning Man

14:30:35 <danbri_dna> BLURB:Agenda

14:30:35 <dc_rdfig> D: Agenda from danbri_dna

14:30:40 <danbri_dna> hi marc!

14:30:45 <marccanter> howdy all

14:30:51 <GregElin> Howdy marc.

14:30:55 <_joshua> hey greg, did i show you http://burri.to/~joshua/map.png yet

14:30:55 <marccanter> :-)

14:31:03 <_joshua> The mapping technology swiftly moves along.

14:31:06 <danbri_dna> D:45 mins of 'show and tell' divided amongst anyone who wants to give a brief tour of their work, ideas, tools etc.

14:31:06 <dc_rdfig> Added comment D1.

14:31:17 <danbri_dna> D:then 15 mins of discussion

14:31:17 <dc_rdfig> Added comment D2.

14:31:32 <danbri_dna> D:Transcribing list of show'n'tellers from the Wiki:

14:31:32 <dc_rdfig> Added comment D3.

14:31:45 <danbri_dna> D:DanBri (very brief item on wordnet)

14:31:45 <dc_rdfig> Added comment D4.

14:31:51 <danbri_dna> D:Libby Miller

14:31:51 <dc_rdfig> Added comment D5.

14:31:54 <danbri_dna> D:Greg Elin

14:31:54 <dc_rdfig> Added comment D6.

14:32:01 <danbri_dna> D:Matt Biddulph

14:32:01 <dc_rdfig> Added comment D7.

14:32:20 <danbri_dna> D:bptheo/Brian on NewsML/reuters

14:32:20 <dc_rdfig> Added comment D8.

14:32:29 <danbri_dna> D:Joshua on GeoURL

14:32:30 <dc_rdfig> Added comment D9.

14:32:32 <danbri_dna> D:Jim Ley

14:32:32 <dc_rdfig> Added comment D10.

14:32:40 <danbri_dna> D:Gerald

14:32:40 <dc_rdfig> Added comment D11.

14:32:49 * danbri_dna counts on his fingers

14:33:11 <GregElin> And Bptheo needs to be added, after Gerald?

14:33:26 <JibberJim> Imarea/rdfpic not mentioned but also probably should be shown.

14:33:45 * danbri_dna had added bptheo between Matt and Joshua

14:33:59 <danbri_dna> OK, let's get started.

14:34:50 <danbri_dna> Greg and I instigated this and have decided to try co-chairing... So I'll say a few words, he'll say a few words, and then let's try cranking through the show'n'tells, in < 5 mins each...

14:34:59 <danbri_dna> Greg, that OK by you?

14:35:02 <GregElin> Let's put Imarea/rdfpic, photofinder, into an others category to do last unless there is someone here who can speak directly for them. but let's make sure they get mentioned.

14:35:16 <danbri_dna> yup, Jim maybe you could chump the urls?

14:35:29 <GregElin> Sounds, great Danbri. I think we are ready to start!

14:35:33 <JibberJim> Will do.

14:35:36 <marccanter> go 4 it

14:35:56 <danbri_dna> Basic idea here was to get a bunch of likeminded folks together, squeeze them into ridiculously small timeslots and try to hoard as many links to interesting image description apps as possible.

14:36:25 <GregElin> Welcome everyone to a show and tell on image description related meta data and all that's fun happenning there.

14:36:46 <danbri_dna> There's a bias towards RDF/XML, SVG etc technologies but I want to be inclusive, the important thing is showing off what we'vee been building and starting to explore how these apps can fit together...

14:36:49 <marccanter> and that's why we're using text - right?

14:37:10 <danbri_dna> feel free to share webcam urls... ;)

14:37:31 <danbri_dna> OK, does everyone have enough context? We're btw logged publically to the Web at http://rdfig.xmlhack.com/

14:37:47 <marccanter> folks should know that SVG ain;t working very well out here in the real world. I've tried several times ot install and it DON'T work - thus I've never seen the legendary FOAFnaut on action.

14:38:04 <GregElin> I think so, we have a question period at the end, so Danbri, I think we can started.

14:38:04 <danbri_dna> I've put myself at the top of the show'n'tell list, and I think I can do this in a one liner...

14:38:10 <danbri_dna> BLURB:danbri show'n'tell

14:38:10 <dc_rdfig> E: danbri show'n'tell from danbri_dna

14:38:16 <danbri_dna> E:Mostly tell...

14:38:16 <dc_rdfig> Added comment E1.

14:38:44 <danbri_dna> OK, getting started let me try be quick. I just wanted to put in a plug for the use of the Wordnet lexical database when describing image content.

14:39:06 <marccanter> what's that?

14:39:17 <danbri_dna> Wordnet is a freely available database of english language terms, and their inter-relationships. Several of us in the FOAF community in particular have been using Wordnet noun terms to classify images, and parts of images.

14:39:20 * danbri_dna grabs url

14:39:42 <marccanter> gloval ontology? OpenOntology?

14:39:47 <marccanter> global

14:39:49 <danbri_dna> E:See [http://rdfweb.org/topic/WhyWordnetIsCool|why wordnet is cool] in foaf wiki

14:39:49 <dc_rdfig> Added comment E2.

14:40:17 <danbri_dna> yeah, it is rough and ready, scruffy etc. but contains 50,000 categories of thing, and including image description markup in foaf rdf files is really easy.

14:40:40 <GregElin> (cooool....)

14:40:42 <danbri_dna> MattB, Jim, Libby might have more to say. I'll add more links too later, but wordnetwordnetwordnet...

14:40:45 <marccanter> coolio

14:40:47 <danbri_dna> OK... next?

14:40:56 <danbri_dna> Do we have Libby?

14:41:01 <libby> yep, hello

14:41:11 <danbri_dna> tell us something about image metadata in < 5 mins!

14:41:17 <libby> ok :)

14:41:25 <danbri_dna> urls welcome... (and make a BLURB for yourself as above, for the record)

14:41:34 <libby> image metadta is neat by difficult to autor (s/difficult/annoying)

14:41:42 <libby> author rather....

14:41:54 <danbri_dna> E:More info on [http://xmlns.com/2001/08/wordnet/|my wordnet rdf vocab].

14:41:54 <dc_rdfig> Added comment E3.

14:42:10 <libby> I'vv eben working on athoring interfaces and also query based interfaces to databases of image metdata - people, wordnet thing ewtc etc.

14:42:44 <danbri_dna> query in sense of human user, or machine api? (or both?)

14:42:49 <libby> basically 2 things to show: (1) crufty-looking but handy authoring tool using webserice-style access to remote datasources

14:43:04 <libby> I was talking about the former, but both

14:43:44 <danbri_dna> ok

14:43:52 <libby> here's that: http://swordfish.rdfweb.org/discovery/2003/06/codjsform/shell.html

14:44:10 <danbri_dna> logger, bookmark?

14:44:10 <danbri_dna> See http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T14-44-10

14:44:16 <GregElin> Libby can you say a bit more about what you mean by "interfaces" - GUI? API? Database interfacing?

14:44:27 <marccanter> pretty crufty

14:44:34 <marccanter> I won;t ask who does your UI :-)

14:44:38 <libby> so what happens is, you type stuff in the form; for some form entries, it goes off and searches a remote datasource, returns RDF, and then you pick the bits you want to craete an rdf file that describes a picture.

14:44:45 <danbri_dna> BLURB:libby show and tell

14:44:45 <dc_rdfig> F: libby show and tell from danbri_dna

14:44:58 <libby> e.g. search for dan, get a bunch of people, pick the one that's the real dan to add to a picture

14:45:04 <danbri_dna> F:See [http://swordfish.rdfweb.org/discovery/2003/06/codjsform/shell.html|js cataloguing ui]

14:45:05 <dc_rdfig> Added comment F1.

14:45:24 * libby does my ui; the key here is to show that remote services are useful for cataloguing

14:45:24 <marccanter> so for each search criterria - there's a huge database of stuff

14:45:33 <marccanter> then how do you combine together all this stuff?

14:45:36 <danbri_dna> F:And [http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T14-44-10|irc transcript]

14:45:36 <dc_rdfig> Added comment F2.

14:46:00 <marccanter> they're just keywords?

14:46:03 <libby> there areseveral dbs: the wordnet one for keyword, one I runb for people (though could be one of many); another for aiprtorts data. could add e.g. spacenamespace data

14:46:12 <libby> keywords are wordnet terms

14:46:23 <marccanter> OH - so you DO believe in centralized databases for people!

14:46:31 <danbri_dna> Libby, is the that having controlled vocab for the who/where/what/when of image description basically means having link to a big database of known entities, places etc?

14:46:48 <danbri_dna> ...which is scooped up from the web by harvesting etc?

14:46:53 <libby> it's combined in the javascript...people who ask me nicely get a passowrd and can upload stuff, or you can add to database by hand

14:47:22 <marccanter> this is for any kind of rdf - no? Not just FOAF...... correct?

14:47:31 <GregElin> Libby, is your primary goal investigating the interface? creating the database?

14:47:42 <libby> yep the key things are: decentralization, multiple sources of data; ease and speed of use; not storing everything on home computer; being able to switch betweenb data sources (e.f. peopsonal addressbokk.

14:48:03 <libby> interface, no....primary aim to get something useful for me to cataloguephotos :)

14:48:12 <danbri_dna> lib, you're over the 5 minutes... can i rush you a bit?

14:48:18 <marccanter> how would - say = an element IN a photo work?

14:48:27 <GregElin> Catalogue your photos and other peoples? can photos reside other locations?

14:48:31 <libby> sure yep. searchable photos db is here: http://swordfish.rdfweb.org/discovery/2001/08/codepict/

14:48:34 <libby> that's it really

14:48:40 <danbri_dna> good to have it driven by real need, even if a selfish need :)

14:48:42 <libby> photos can reside anywhere

14:48:49 <GregElin> cool!

14:48:53 <libby> whather works dpends on the form of thumbnail pages etc

14:48:59 <marccanter> add ne fields for new things to track?

14:49:06 <libby> yep....

14:49:18 <marccanter> lots of DIFFERENT people databases?

14:49:24 <libby> realize interface crufty...proof of concept is all....

14:49:38 <libby> I'm done, cheers for listening :)

14:49:43 <zool> :)

14:49:49 <marccanter> so no format-standard per se - just RDF

14:50:00 <danbri_dna> i guess the client app doesn't care which rdf store it talks to, and apart from issue of having to proxy when connecting to diff servers, it could talk to several at once

14:50:04 <danbri_dna> OK cool, thanks lib.

14:50:06 <danbri_dna> greg?

14:50:17 <GregElin> Thanks.

14:50:17 <marccanter> gotcha - coolio - ist

14:50:36 <GregElin> I'm into proofs of concepts, too, like Libby.

14:50:36 <danbri_dna> BLURB:greg show and tell

14:50:36 <dc_rdfig> G: greg show and tell from danbri_dna

14:50:41 <danbri_dna> logger, bookmark?

14:50:41 <danbri_dna> See http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T14-50-41

14:50:57 <danbri_dna> G:See [http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T14-50-41|IRC transcript].

14:50:57 <dc_rdfig> Added comment G1.

14:51:03 <GregElin> In 2000, I started prototyping a way to manage images as a collection of objects.

14:51:58 <GregElin> Basically, after working with relational databases and seeing some limitations, I thought it would be easier to work with image data if the computer could tread the image as a metaphor of the objects represented.

14:52:02 <GregElin> Just like we people do.

14:52:25 <GregElin> Here is link to an example image:

14:52:26 <GregElin>http://www.fotonotes.net/index.php?action=m&p=demo%2FDSC00041.JPG

14:52:26 <dc_rdfig> H: http://www.fotonotes.net/index.php?action=m&p=demo%2FDSC00041.JPG from GregElin

14:52:45 <GregElin> Which my mother annotated. Roll your mouse over the numbers to see annoations come up.

14:53:07 <danbri_dna> I'm not quite sure I follow re "metaphor", could you expand briefly?

14:53:38 <gerald> nice!

14:53:47 <marccanter> the image itself is teh metaphor

14:53:56 <marccanter> test is sub-ordinate to the image

14:53:57 <GregElin> When we talk about photos, we flip between talking about the photo as an IMAGE (composition, size, print quality, lens...)

14:54:26 * danbri_dna admires the demo

14:54:27 <GregElin> And talking about what is inside the image as if we were talking about the actual object. "This is my motorcyle."

14:54:39 <danbri_dna> Ah, right...

14:54:46 <GregElin> "This is my granmother. She gave me that hat. I still have it."

14:55:03 <danbri_dna> yup, it's easy to treat the technology transparently, like a window onto a world...

14:55:14 <marccanter> in this case - the image drives the metaphor - the text is the annotation (obviously)

14:55:23 <danbri_dna> (behind the scenes the computer still makes the distinction, i assume... ie. its a UI thing?)

14:55:28 <GregElin> But image software only understands pixels. And databases are lousy at treating data as metaphor (having more than one related meaning)

14:55:29 <marccanter> so meta-data and standards become crucial

14:55:34 <_joshua> Thanks, Rene Descartes

14:56:33 <marccanter> so the wordnet stuff gets stuffed into the fotonote annotations. No?

14:56:35 <danbri_dna> OK, so Greg can you say a few words about what's going on behind the scenes? what's the model associated with each picture?

14:56:36 <GregElin> I'm not sure it ultimately matters HOW the computer deals with things. What matters is that we have the experience and capability of working either with the IMAGE or working with what is in the IMAGE.

14:56:51 <danbri_dna> several areas/rectangles + text (plain, html?)

14:56:55 <GregElin> To be brief let me bullet point.

14:57:12 <GregElin> Data is stored directly in jpeg headers to allow data to travel WITH image.

14:57:13 <marccanter> embedded into JPEG

14:57:31 <gen> I love that part

14:57:35 <GregElin> Data is in a rough XML format (I'd like to figure out which standard to adopt)

14:57:48 <GregElin> Data can be kept externally, too, though.

14:57:54 <marccanter> have I introduced you to my RDF nerd friends?

14:58:00 * danbri_dna wonders if you've seen the adobe xmp stuff

14:58:19 <GregElin> Current implementation tracks overall narrative, and then multiple "selections". Each selection has story and region defined.

14:58:30 <ndw> Using RDF would make it compatible with other JPEG metadata tools, uhm, like my own :-)

14:58:30 <GregElin> Somewhat Danbri.

14:58:35 * danbri_dna was forgetting the clockwatching, we're at 6 mins now...

14:58:38 <marccanter> And I want to point out - Greg did this on his own - no government agency or sponsoring research group!

14:59:03 <GregElin> But the idea is to make these regions not just REGIONS...but addressable objects...so you can link to and work with and spawn and interrelated pieces from different source jpegs.

14:59:13 <GregElin> (thanks, marc)

14:59:21 <GregElin> Let me wrap up.

14:59:42 <marccanter> codepiction of French provencial chair to Napolean and Marie Antoinette

14:59:58 <GregElin> I decided that simple rectangles was okay for basic feature. Polygons can be advanced for those who want. I've tried to make the interface grandma easy. (I think the forms are an obstacle)

15:00:13 <marccanter> and that dildo ditting on the table to..........[insert here]

15:00:42 <marccanter> killer fucking killer killer killer UI - congrats again Greg!

15:00:54 * danbri_dna grins

15:00:56 <GregElin> Wordnet and wiki structures and Libby's stuff all suggest these tools could work together...and we end up with jpegs having rich data, some very formal, some informal. Just like the web

15:00:59 <danbri_dna> you like the UI then, marc? ;)

15:01:23 <marccanter> if only the whole world could be reduced to rectangles and overlay text annotations

15:01:27 <danbri_dna> MattB, you're up next...

15:01:33 <danbri_dna> what a world it'd be...

15:01:37 <mattb> ok

15:01:41 <GregElin> Last note: I have php and java libraries that read/write jpeg headers and build layers. I'd like to extend that code to work with this other stuff.

15:01:44 <GregElin> (done)

15:01:50 <danbri_dna> Cool, greg. thanks...

15:02:05 <bijan> logger, pointer

15:02:05 <bijan> See http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T15-02-05

15:02:08 <danbri_dna> ...if you've more URLs, post them here and folk can rummage around the details

15:02:11 <GregElin> (oops...http://www.fotonotes.net/dev for the record)

15:02:18 <danbri_dna> BLURB:mattb show'n'tell

15:02:18 <dc_rdfig> I: mattb show'n'tell from danbri_dna

15:02:20 * mattb suffering a bit of lag today, apologies if there are problems...

15:02:29 <mattb> I:[my slides and demos|http://www.hackdiary.com/archives/000034.html]

15:02:30 <dc_rdfig> Added comment I1.

15:02:37 <danbri_dna> I:See [http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T15-02-05|IRC transcript]

15:02:37 <dc_rdfig> Added comment I2.

15:02:39 <bptheo> Greg, I can't seem to get this link to work without logging into fotonotes.net, but it would be cool to show: http://www.fotonotes.net//index.php?action=m&p=bptheo/2003-06-03T215730Z_01_KAB12D_RTRIDSP_1_AFGHANISTAN.jpg

15:02:45 <mattb> i've been annotating my photos for a while

15:02:53 <mattb> inspired by libby's photo codepictions demo of a while back

15:03:18 <mattb> not doing anything as nice as actual polygon definition of areas, just saying <photo> <foaf:depicts> <wordnet:Dog> etc

15:03:32 <marccanter> that link is broken here

15:03:34 <mattb> i'm using RSS 1.0, and making it the backbone of a blog-ish site at http://www.picdiary.com

15:03:36 * bijan doing the polgon defintion etc.

15:03:51 <bijan> With a drawing/markup tool nearing beta, i believe

15:04:02 <bijan> (Not me personally, though i did do an early mockup)

15:04:03 <JibberJim> Ooh add BIjan to the agenda please.

15:04:33 <mattb> i spent some time building search engines and things that use the linkages within the wordnet ontology to make better search

15:04:42 <mattb> so search for animal finds pictures marked up as cat, dog, etc

15:04:47 <zool> do you have that little semantic search demo mattb

15:04:56 <mattb> since they have a hypernym/hyponym relation within wordnet

15:04:59 <libby> mattb has done a bunch of very cool stuff with wordnet data

15:05:04 <mattb> eg http://www.picdiary.com/cgi-bin/search.pl?word=hotel versus http://www.picdiary.com/cgi-bin/search.pl?word=building

15:05:06 <marccanter> so I click on Kentish town and all I see is the photo ID

15:05:10 <mattb> there is only one hotel, but many buildings

15:05:26 <mattb> the hotel shows up in the search for building, but only has <foaf:depicts> <wordnet:Hotel> in its metadata

15:05:39 <danbri_dna> example of wordnet plus FOAF markup: (using Greg's photo) <foaf:Image rdf:about="http://www.fotonotes.net/photographs/demo/DSC00041.JPG"><foaf:depicts><wordnet:Parchment/></foaf:depicts></foaf:Image>

15:06:04 * danbri_dna recommends folk rummage through mattb's online slides, a lot more detail there than we can get in 5 mins here...

15:06:11 <bijan> is anyone doing assertion propegation from whole photo to subregions, and from either to content?

15:06:18 <danbri_dna> mattb, how did you find wordnet's coverage, w.r.t. your photo collection?

15:06:34 <mattb> pretty good, apart from the american-centric aspect

15:06:37 <danbri_dna> I'd hate to suggest it is the last word in big general purpose ontologies, but it does seem to be a useful base...

15:06:45 <mattb> eg having to say "Bar room" rather than "pub"

15:06:55 <libby> yeah like wn:ics: american like ice in their drinks!

15:06:57 <danbri_dna> interesting

15:06:57 <bijan> E.g., photo takenAt 12:03 --> subregion takenAt 12:03 --> event depicted by photo happeningAt 12:03

15:06:59 <bijan> Etc.

15:07:00 <libby> wn:ice rather

15:07:16 <libby> pub's in there isn;t it?

15:07:18 * DanC has been noodling on that sort of thing, bijan...

15:07:19 <danbri_dna> bijan, that's interesting. similar need w.r.t. batches of photos.

15:07:35 <mattb> perhaps bad example... there was no word for the inside of a pub

15:07:38 <mattb> only pub as a building

15:07:40 <DanC> but I haven't done anything because of the "don't try to capture data you don't know how to exploit" principle.

15:07:41 <bijan> This is big on our list

15:07:45 <mattb> as opposed to "people in a pub"

15:07:45 <bijan> With video too

15:07:54 <danbri_dna> mattb, where are you going with it next?

15:07:54 <marccanter> so do you folks call that an exstension of an RDF vocabulary? More RDF verbs and such?

15:08:09 <danbri_dna> call which an extension?

15:08:10 <mattb> was interesting building the semantic search stuff - i used a wordnet API directly rather than an expression of wordnet ontology in some RDF-compatible form

15:08:21 <GregElin> missed that...what's big on your list.

15:08:22 <marccanter> this whole thing that Matt is doing?

15:08:23 * DanC relurks

15:08:24 <danbri_dna> we call RDF vocabs ontologies sometimes; i generally stick to 'vocabulary'

15:08:25 * ndw has to wander away for another telcon

15:08:32 <mattb> an important next thing is editing interfaces

15:08:44 <GregElin> Very cool, mattb. And nice to look at.

15:08:45 <mattb> at the moment i hand-edit RSS files, simple as that

15:08:47 <danbri_dna> it's more an application of an few rdf vocabs... mattb, did you have to actually invent any new rdf vocab?

15:08:57 <mattb> good point danbri_dna

15:09:02 <mattb> i used almost no new vocab at all

15:09:05 <mattb> that was a major aim

15:09:09 <mattb> to be lazy, get reuse

15:09:10 <danbri_dna> that's good, right? :)

15:09:15 <mattb> very good

15:09:23 <mattb> i used dublin core, FOAF, zool's spacenamespace, wordnet

15:09:26 <marccanter> me like too

15:09:30 <mattb> and rss 1.0 for structure

15:09:39 <marccanter> good old zool who's that?

15:09:43 <danbri_dna> in effect you create a document format without inventing any new namespaces/vocab... (similar to syntactic profile thread we're having on rdfweb-dev foaf list currently)

15:09:51 <mattb> i think i invented one URI for a single property, "picdiary:highlight" to indicate which pic in a collection is the highlight, the one to promote on the homepage

15:09:53 <danbri_dna> zool==jo walsh (hi jo!)

15:09:57 <danbri_dna> ok timecheck...

15:10:01 * mattb done

15:10:01 <danbri_dna> mattb, anything more to add?

15:10:14 <marccanter> coolio

15:10:21 <GregElin> Mattb, where is the data for each image stored?

15:10:24 <zool> ello

15:10:32 <danbri_dna> we have 20 mins left, 4 people left...

15:10:42 <danbri_dna> ...not much time within the 1hr for discussion, but let's crack on...

15:10:44 <danbri_dna> next: bptheo/Brian on NewsML/reuters

15:10:45 <mattb> GregElin: each picture collection, eg http://www.picdiary.com/highwalk , has an associated http://www.picdiary.com/rss/highwalk.rss

15:10:55 <mattb> these are aggregated together into a redland store for building pages, query, etc

15:11:00 * bijan really only lurking, didn't realize it was scheduledChat

15:11:07 <GregElin> So a folder structure?

15:11:16 <mattb> yeah, but only for storage purposes

15:11:22 <bptheo> btw, the link I posted a while ago on fotonotes.net should work now

15:11:31 <mattb> i deal with the info entirely in the rdf world, filesystem independent

15:11:34 <danbri_dna> (bijan, it was only belatedly advertised on www-rdf-interest, my fault + mail screwups last week)

15:11:36 <bptheo> It was a sample I tried connecting 2 Reuters photos together

15:11:48 <GregElin> cool. good approach.

15:11:50 <mattb> so rather than N newest files to make my homepage, i find N newest dc:date triples

15:11:54 <danbri_dna> BLURB:bptheo's show'n'tell

15:11:54 <dc_rdfig> J: bptheo's show'n'tell from danbri_dna

15:11:59 <danbri_dna> logger, bookmark?

15:11:59 <danbri_dna> See http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T15-11-59

15:12:00 <mattb> and navigate the graph from there

15:12:02 * mattb really done now :)

15:12:06 <marccanter> gotta annotate those guns and WHO they've killed recently

15:12:13 <danbri_dna> J:See [http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T15-11-59|irc transcript]

15:12:13 <dc_rdfig> Added comment J1.

15:12:14 <bptheo> I work for Reuters in their Innovation Studio (part of CTO group)

15:12:18 <marccanter> NOW for NewsML!

15:12:23 <GregElin> Here's brian's link again:

15:12:24 <GregElin>http://www.fotonotes.net/index.php?action=m&p=bptheo%2F2003-06-03T215730Z_01_KAB12D_RTRIDSP_1_AFGHANISTAN.jpg

15:12:24 <dc_rdfig> K: http://www.fotonotes.net/index.php?action=m&p=bptheo%2F2003-06-03T215730Z_01_KAB12D_RTRIDSP_1_AFGHANISTAN.jpg from GregElin

15:12:25 <bptheo> we're responsible for applying new technology to our business

15:12:51 <bptheo> A couple of years ago, Reuters wanted to "modernize" the way we prodice content as well as the format in which we distribute it

15:12:57 * danbri_dna nods

15:13:03 <marccanter> here here

15:13:22 <bptheo> NewsML was the answer... a very heavy-weight XML vocab for capturing both the editorial workflow as well as news + metadata

15:13:35 <bptheo> a not-so-interesting-but-public demo is here: http://about.reuters.com/newsml/newsmldemo.asp

15:13:47 <marccanter> and I got all wound up in a proposal to Tom Glocer to do something with it :-)

15:14:11 <bptheo> The interesting bits are less about the lengthy XML, but in how we're using it

15:14:27 <GregElin> Do we need to apply to get a password?

15:14:38 <bptheo> no

15:14:46 <bptheo> it should say what the login/pass are in the page

15:15:07 <marccanter> nope

15:15:07 <danbri_dna> Note: this site is password protected, to obtain a password please contact your local Reuters sales representative."

15:15:10 <danbri_dna> would that be you? ;)

15:15:11 <deltab> it does: newsml_/showcase

15:15:27 <bptheo> The interesting bits are how we're using NewsML internally... we built a new editorial publishing system around it

15:15:57 <bptheo> Given our size, we've taken the approach of working with the IPTC (iptc.org) to develop an ontology of "topic" codes for news stories

15:16:11 <marccanter> oh so YOU were the one who built the.......

15:16:22 <GregElin> oh, very cool...briefing books!

15:16:22 <danbri_dna> is that (going to be) available for others to use?

15:16:40 <bptheo> and from the journalists contribution tool leveraged NLP (natural language processing) by InXight to do auto-categorizing as the journalist types

15:16:53 <GregElin> Rock on!

15:17:11 <bptheo> for example, auto-coding the story for topic codes based on the ontology ( and training sets) and auto-tagging companies, locations, etc.

15:17:22 <GregElin> Brian, it seems Reuters has done the industrial version of what we could do with wordnet and the tools we are discussing here.

15:17:33 <danbri_dna> Is NewsML equally well suited to image description as to textual news articles? Or do you need to squeeze/extend it to fit with image classification...?

15:17:35 <bptheo> we're not doing much specifically around RDF at the moment

15:17:59 <_joshua> But the NewsML-encoded stories don

15:18:06 <_joshua> But the NewsML-encoded stories don't escape Reuters, right?

15:18:15 * danbri_dna nods; wordnet's interesting in large part because it is free, so likely to get planet-wide use... But creating these classifications is expensive... not everyone can give their topic trees away...

15:18:42 <bptheo> yes, there's always 2 schools of thought around ontologies... unfortunately, for us it is a necessary evil we must do. Most of our customers are financial professionals (not readers of Reuters.com), therefore doing complex metadata is pretty important

15:18:50 <marccanter> NewsML was "given" to the IETF - correct?

15:18:57 <GregElin> Right, danbri. But not all web pages are open to the public anyway. I think we all have our private and public ontologies...

15:19:37 * danbri_dna nods; giving away the 'upper layers' is something that appeals to some (eg. Cyc did that for a while, not sure if they still keep some private)

15:19:52 <bptheo> Once a journalist has written them story, their editor/bureau head starts linking other assets to the story using NewsML... So, in effect we use NewsML more as a container suitable for abstracting a rendering engine to display it

15:20:07 <GregElin> I think we should make a list at the end of meeting of the different very relevant *ML's.

15:20:10 <danbri_dna> bptheo, where's this work headed now?

15:20:14 <bptheo> we'd include things like there being 3 related images, what there formats are, links to thumbnails, web-friendly size, and full size

15:20:20 <bptheo> associated audio/video clips, etc.

15:20:22 <danbri_dna> thanks for volunteering, Greg ;)

15:20:45 <danbri_dna> does NewsML help you do that, bptheo, or are those image-oriented extensions your own?

15:20:46 * laughs

15:21:11 <bptheo> danbri- Well, we're just rolling out the News2Web editorial system internally and starting to integrate NewsML with out apps

15:21:15 <bptheo> this is where I've come in

15:21:30 <bptheo> I've been looking at fotonotes as a possible enhancment to the level of metadata we do around photos

15:21:49 <danbri_dna> ah right, cool

15:21:59 <marccanter> speaking of Rendering NewsML - you mean like this? http://broadbandmechanics.com/media/bizpres/ReutersResearcher.jpg

15:22:04 <GregElin> (I like that, Brian!)

15:22:07 <bptheo> unfortunately, there's a long cultural history at work here... like, the photojournalists are so trained to crop the photos way down such that the single description is relevant and contextual

15:22:09 <danbri_dna> I'm interested to hear more (prolly not right now as we're overrunning) about extensibility of NewsML for these purposes

15:22:11 <marccanter> just a ditty from our past

15:22:24 <bptheo> (thus, my little demo on fotonotes.net I posted a bit ago)

15:22:31 <danbri_dna> cool

15:22:36 <marccanter> bptheo - let's stay in touch!

15:22:43 <danbri_dna> bptheo, anything to say in wrapping up? this is cool stuff btw :)

15:22:57 <marccanter> life has a strange way of.....

15:23:01 <GregElin> Brian's point is an excellent one. Professional photos have less ambigious background material to annotate. ironies of ironies.

15:23:04 <bptheo> We're very much interested in finding new ways to leverage the amazing assets we have among our 2000 journalists around the globe... many of who are tech-savy

15:23:06 * danbri_dna looks at the http://rdfig.xmlhack.com/ lineup... still to come: Joshua, Jim, Gerald...

15:23:27 <marccanter> Ben Hammerlsey needs some leveraging (and good luck) right now

15:23:41 <marccanter> I gotta go - all - I vote YES to everything

15:23:42 <danbri_dna> bptheo, how best can we track this work, participate etc? is there a developer site, mailing list or web page we should keep an eye on?

15:23:44 <GregElin> I'm keen to go until we hear from everyone. I'm finding this very useful.

15:23:45 <marccanter> and when will it be ready?

15:23:51 <danbri_dna> heh ok marc, thanks for coming...

15:23:54 <marccanter> :-)

15:24:02 * danbri_dna happy to extend, this isn't super formal anyway...

15:24:11 <GregElin> I think we should start one. And I might volunteer for that, too.

15:24:18 <bptheo> danbri -- I belive most of it is on iptc.org

15:24:25 <danbri_dna> ok, that's good to know.

15:24:36 <bptheo> being a "good citizen" we handed our role in NewsML we created over to a consortium

15:24:49 <danbri_dna> ah, right

15:24:50 <bptheo> I belive there's is a open-source NewsML toolkit available we createdx

15:24:53 <GregElin> Brian, very quickly, how does NewsML and IPTC stuff work together?

15:24:55 * danbri_dna will take a look around iptc.org...

15:25:06 <bptheo> to facilitate easy creation/inspection of newsML

15:25:28 <bptheo> Greg, we created the first version several years ago, then worked with IPTC to standariize it

15:25:44 <GregElin> Is IPTC or AP proposing or working with any image annotation meta data stuff as a standard?

15:25:45 <bptheo> 1.0 was ratified about 2 years ago and has tweaked ever since

15:26:03 <bptheo> Greg, that's a good question, one I will find out

15:26:04 <GregElin> Or is it the IPTC standard NewsML?

15:26:08 <GregElin> Thanks.

15:26:19 <bptheo> that'

15:26:23 <GregElin> That was terrific Brian, stuff I don't think many of us were aware.

15:26:25 <bptheo> that's abou it for my 5 minutes

15:26:40 <danbri_dna> OK thanks, lots of stuff to rummage around in afterwards...

15:26:56 <danbri_dna> Joshua, you got time to give us a brief headsup on GeoURL?

15:27:09 <_joshua> Hmm. I'm not sure what to say about it.

15:27:19 <_joshua> Not a lot has changed lately

15:27:29 <danbri_dna> BLURB:Joshua's show'n'tell

15:27:29 <dc_rdfig> L: Joshua's show'n'tell from danbri_dna

15:27:31 <_joshua> For those that don't know, it's a way to geo-encode web pages

15:27:37 <JibberJim> hey I'll be talking about something I've not touched in over a year...

15:27:37 <danbri_dna> logger, pointer?

15:27:37 <danbri_dna> See http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T15-27-37-1

15:27:52 <gen> How many are in the geoURL db?

15:27:56 <_joshua> 40k items now

15:28:01 <_joshua> grows by 1-200 a day

15:28:13 <danbri_dna> L:see [http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T15-27-37-1|irc transcript]

15:28:14 <dc_rdfig> Added comment L1.

15:28:22 <_joshua> I've been tinkering with maps lately

15:28:22 <GregElin> You can look up a location and then see webpages that at geocoded to be related to that location.

15:28:38 <_joshua> Right. It's a geo-location reverse directory

15:28:55 <danbri_dna> so at moment its only implicitly working w/ photos, images... eg. you could geourl-tag an image archive page

15:29:06 <danbri_dna> (there is a nice rdf view of the geourl aggregate, too)

15:29:07 <_joshua> you put coordinates on a page, and then tell geourl about the page, and it indexes it.

15:29:14 <GregElin> First one, I think. Or at least first one to provide a simple model on how to add the geoURL code to your page and get it in a database for searching.

15:29:44 <danbri_dna> yeah, very impressive takeup -- from keeping it simple and easy and having nice tool that consumes the data...

15:29:45 <_joshua> yes, i've been thinking about building a variant of geourl that lets you index the geotags in EXIF or whatever

15:29:56 <_joshua> i believe nym is working on this as well

15:29:57 <GregElin> I think the next step is to geocode the photos, and then add geocodes to objects in the phone (like born in Indiana)

15:30:21 <_joshua> Of course, in some sense, geourl treats the relationship as inverted.

15:30:31 <danbri_dna> yup, there are lots of relationships that connect places to the things in photos, not just 'this photo was taken in ...'

15:30:32 <_joshua> The web page is metadata about the location, not the other way around.

15:30:38 <danbri_dna> eg. 'this chair was made in ...'

15:30:39 <GregElin> Josh is being his typical modest self about his accomplishment.

15:30:39 <_joshua> people have trouble following this distinction

15:30:49 <danbri_dna> he knows we love geourl already ;)

15:30:55 <GregElin> right, danbri.

15:31:04 <GregElin> We all love you and geoURL, Josh.

15:31:18 <_joshua> anyway, i've been struggling with far more mundane problems: label placement on maps

15:31:22 <_joshua> which is also image annotation :)

15:31:30 <GregElin> :-)

15:31:51 <_joshua> for example, http://burri.to/~joshua/map.png

15:31:55 <danbri_dna> images are annotations of the things they depict...

15:31:55 <_joshua> Jim was helping me with it

15:32:05 <zool> joshua, you looking at any of the openGIS.org web map standards - WFS and WMS services?

15:32:05 <_joshua> danbri: not necessarily

15:32:15 <danbri_dna> SVG again, interesting...

15:32:18 <_joshua> zool: yes. GeoURL should be a WFS.

15:32:30 <_joshua> WFS = web feature service, WMS = web map service

15:32:34 <_joshua> so vector and raster, basically

15:32:37 <GregElin> I think someone needs to undertake a "render engine" that allows supports different data backends and allows for variations of presenting annotions on, above, next to, etc. with the photos.

15:32:52 <_joshua> WFS is basically a restful interface to a geographic dataset, that emits GML

15:32:59 <_joshua> Except way more complicated.

15:33:15 <danbri_dna> is there a simple/obvious subset or profile?

15:33:16 <_joshua> Theoretically you could aggregate various WFS services together. This would be Really Fucking Cool.

15:33:26 <_joshua> GML is painfully complex

15:33:36 <GregElin> sorry, WFS??

15:33:42 <_joshua> I got some xslt to turn GML to SVG; it was about a megabyte of XML. Eugh

15:33:47 <_joshua> WFS = Web Feature Service

15:33:50 <zool> istr they propose big standard ontologies for web features

15:33:54 <danbri_dna> I like that geourl exposes itself as a service, ie. RDF from urls like: http://geourl.org/rdf/?lat=37.75&lon=-122.45

15:34:00 * GregElin thanks

15:34:06 <_joshua> query a webserver about a point, get back some GML

15:34:19 <_joshua> Anyway.

15:34:59 <danbri_dna> cool stuff :)

15:35:03 <danbri_dna> shall we move on?

15:35:08 <_joshua> Ok.

15:35:11 <zool> there could so so much potential in an RDF subset of GML features, sort of thing that mapbureau.com were proposing

15:35:12 <danbri_dna> Jim and then Gerald, if you're both still here...

15:35:14 <GregElin> Sounds good. Thank you joshua.

15:35:26 <danbri_dna> GML was originally in RDF, and so shares a good chunk of RDFish design...

15:35:29 * gerald still here

15:35:32 * zool sorry sidetracking

15:35:32 <JibberJim> yep, I'm here too.

15:35:46 <danbri_dna> sidetracking, no prob... that's what irc's for ;)

15:36:02 <zool> it seems to connect to the 'making meaningful statements about depicted things in the world' theme though

15:36:24 * mortenf also fiddling with gml/rdf

15:36:25 <danbri_dna> Jim, tell us about some old stuff you did that you've nearly forgotten about but which in fact is really cool and should be revisited and known about by many...

15:36:39 <JibberJim>http://jibbering.com/svg/AnnotateImage.html

15:36:40 <dc_rdfig> M: http://jibbering.com/svg/AnnotateImage.html from JibberJim

15:36:53 <danbri_dna> M:Jim's show'n'tell (SVG image annotation)

15:36:53 <dc_rdfig> Added comment M1.

15:36:54 <GregElin> I think that might be the overarching theme, zool. "Making meaningful statements about depicted things."

15:37:08 <JibberJim> M:Creates RDF which can later be crawled by Libby's codepiction and image search widgy. Also there are lots of SVG clients I did such as [http://jibbering.com/2002/8/img-desc.svg|SVG Image description] and [http://jibbering.com/rdfsvg/search.1?noun=Beer|Wordnet Search (beer in this case)].

15:37:10 <dc_rdfig> Added comment M2.

15:37:20 <JibberJim> M: For those of you without SVG - [http://jibbering.com/2003/8/svg-annotator.jpg|Annotator screenshot] and [http://jibbering.com/2003/8/img-desc.jpg|Image Description screenshot].

15:37:20 <dc_rdfig> Added comment M3.

15:37:25 * JibberJim prepared!

15:37:30 <danbri_dna> jim++

15:37:46 <JibberJim> M:General idea is that you can draw regions of the photo within the page, like Fotonotes but not limited to rectangles, and then attach a wordnet class, a description, and if it's a Person a name and mbox_sha1sum of them.

15:37:46 <dc_rdfig> Added comment M4.

15:38:16 <danbri_dna> M:A bit like HTML imagemaps, but cooler cos you get to say what the bit of the image depicts...

15:38:17 <dc_rdfig> Added comment M5.

15:38:40 <JibberJim> I played with lots of interfaces to the system, there was even a version which talked the description of the image out using a TTS engine.

15:38:56 <JibberJim> This was all inspired by danbri's and chaals image map thingy.

15:39:24 <danbri_dna> M:[+http://rdfweb.org/2002/01/photo/svg/charles-danbri-amaya-outline.jpg][+http://rdfweb.org/2002/01/photo/svg/charles-danbri-amaya-sm.jpg] (charles and me having the original vision ;)

15:39:24 <dc_rdfig> Added comment M6.

15:39:38 <GregElin> very cool Jibber.

15:39:43 <gen> Good night from Tokyo, all.

15:39:44 <danbri_dna> M6:+[http://rdfweb.org/2002/01/photo/svg/charles-danbri-amaya-outline.jpg]+[http://rdfweb.org/2002/01/photo/svg/charles-danbri-amaya-sm.jpg] (charles and me having the original vision ;)

15:39:44 <dc_rdfig> Replaced comment M6.

15:40:08 <JibberJim> but for various reasons I got distracted into the world of foafnaut instead in my spare hacking time so it's gone on the back-burner.

15:40:10 <danbri_dna> What is interesting about this, imho, is that these techniques turn the web into a big clipart database.

15:40:13 <GregElin> (Fotonotes also had a function which exploded the image like this. I like the way you did and describing the relationship of the object where it is in the image.)

15:40:49 <GregElin> YES! Righto, danbri.

15:40:54 <_adam_home> _adam_home is now known as adamhill

15:41:05 <GregElin> That's part of what I've always been interested in.

15:41:09 <danbri_dna> M:Related writeup: [http://rdfweb.org/2002/01/photo/|co-depiction, foaf etc] and work w/ charles on doing this in [http://www.w3.org/2001/08/rdfweb/svg-foaf.html|Amaya], w3c's browser.

15:41:09 <dc_rdfig> Added comment M7.

15:41:16 <JibberJim> M:Nadia Heninger did a new version of [http://jigsaw.w3.org/rdfpic/|RDFPic], but I don't know where it is, the RDF it produced was compatible with mine.

15:41:16 <dc_rdfig> Added comment M8.

15:41:36 <danbri_dna> M:I'll chase that up w/ Bert... I know he was working on it lately.

15:41:37 <dc_rdfig> Added comment M9.

15:42:17 <GregElin> JibberJim: is there a name space model in your implementation that makes the objects addressable via URL?

15:42:21 <danbri_dna> Jim, is the code available for others to use? Anything else needed except the SVG and .js they can see?

15:42:38 * ndw has sent patches to Bert for the RDF handling in rdfpic, but never heard a reply

15:43:16 <danbri_dna> M:See also [http://www.w3.org/2001/08/rdfweb/foafwho/imagemap/|experiments with generating same markup via HTML Imagemap authoring tools].

15:43:16 <dc_rdfig> Added comment M10.

15:43:34 <danbri_dna> I have a hunch that html imagemaps might be the key to this stuff going mainstream...

15:43:52 <JibberJim> Yes, everything in the picture is given a URI, if you look at some example RDF: HTTP://jibbering.com/rdf/example.rdf (which is the shepherds) you can see the polygons and subregions are addressable.

15:44:10 <danbri_dna> IRC spam coming up:

15:44:11 <danbri_dna> [[

15:44:11 <danbri_dna> <area shape="POLY"

15:44:12 <danbri_dna> coords="81,472,81,427,67,419,57,385,49,396,45,390,30,334,32,324,70,267,69,255,66,242,86,223,95,220,102,220,102,226,105,175,114,159,129,148,148,148,160,159,164,177,179,217,201,238,213,251,206,335,191,465,191,476,140,476,87,476,82,475"

15:44:13 <danbri_dna> alt="Coralie Mercier" href="mailto:coralie@w3.org" />

15:44:15 <danbri_dna> ]]

15:44:21 <danbri_dna> ...snippet of ordinary html imagemap markup.

15:44:31 <danbri_dna> It's *nearly* there in terms of giving something to hang the semantics off of.

15:44:45 <danbri_dna> Jim's markup puts chunks of this sort of thing, as SVG fragments, inside the RDF graph.

15:44:51 <JibberJim> So the RDF says http://jibbering.com/rdfsvg/#1029939144970_0

15:45:12 <JibberJim> is a part of a picture, and that part depicts a Shepherd.

15:45:36 <GregElin> And the usual question -- where is the data stored?

15:45:37 <danbri_dna> M:See also [http://rdfweb.org/2002/svgsemantics/picsng-demo.html|a PICS-NG demo] I built using Jim's tools.

15:45:37 <dc_rdfig> Added comment M11.

15:45:55 <jql-zzz> jql-zzz is now known as jql-work

15:46:06 <JibberJim> RDF documents currently just dumped into the filesystem. I also looked at dumping it into Annotea, and had that working, but it doesn't anymore.

15:46:27 <danbri_dna> Having them in the filesystem is nice and simple...

15:46:43 <GregElin> May I ask a self-interested question at this junction?

15:46:47 <JibberJim> Generally though it's RDF which Libby scutters up for her system, and hopefully should be able to search, but not checked recently that it actually works.

15:46:57 <GregElin> It can also keep to the end...

15:47:10 <danbri_dna> ('scutter' is rdf/foaf talk for harvest/index/collect, like a search engine crawler...)

15:47:31 <danbri_dna> gerald, are you pressed for time or can we have greg's q now?

15:47:42 <gerald> go ahead

15:48:50 <danbri_dna> greg?

15:48:55 <GregElin> thanks. If I shared the fotonotes libraries that read/write data to the JPEG headers, do you think you all might adopt it to include this different data into the jpeg rather than separate?

15:49:07 <GregElin> Or do people like keeping it separate?

15:49:21 <mortenf> separate please, lighter transfers etc.

15:49:40 <chrisGoad> However, for some apps, inserting into the jpg would be great

15:49:43 <masaka> put rdfs:seeAlso in JPEG ?

15:49:46 <danbri_dna> I like doing it both ways depending on context. opensource libs for stuffing metadata inside JPEG would certainly be useful. There seem to be several styles of doing it, XMP is getting a bit of traction...

15:49:53 <JibberJim> Oh yeah, sorry forgot to mention that there is a version of the search which when given a jpeg uses a service from danbri to extract XMP packeted RDF from the photo and displays on that.

15:50:03 <gerald> I prefer separate

15:50:09 * GregElin inclines steeply toward open source...

15:50:25 <JibberJim> I'd certainly prefer to store it with the jpeg sometimes, but this being a client-side web-app that really wasn't an option.

15:50:25 <danbri_dna> I generally manage it separately, but sometimes for packaging/shipping, copying it inside the file useful too

15:50:43 <GregElin> I agree, Dan.

15:50:46 <danbri_dna> self-describing images definitely a nifty idea

15:50:55 <gerald> I'd like tools to be able to handle internal or external metadata

15:51:02 <danbri_dna> I did a bit of work w/ dave beckett before he came to ilrt, on this w.r.t. PNG meta

15:51:07 <mortenf> i guess "both" is a nice answer :)

15:51:08 <GregElin> Yes, Gerald.

15:51:15 <JibberJim> I'd like it inside, and I've been trying to press dino and the SVG WG folks to give me access to XMP packets in jpegs, please bug him too if you want this...

15:51:16 <danbri_dna>http://www.tasi.ac.uk/2000/09/rdfmeta/

15:51:16 <dc_rdfig> N: http://www.tasi.ac.uk/2000/09/rdfmeta/ from danbri_dna

15:51:21 <danbri_dna> N:RDF for self-describing images

15:51:22 <dc_rdfig> Added comment N1.

15:51:39 <dino> we're adding it Jim

15:51:39 <edd> danbri_dna: N:| ?

15:51:41 <danbri_dna> N:writeup of RDF inside PNG experiments using Dave Beckett's pngmeta library

15:51:41 <dc_rdfig> Added comment N2.

15:51:48 <dino> problem will be getting someone to implement it

15:51:48 * JibberJim cheers!

15:51:52 <danbri_dna> N:|RDF for self-describing images

15:51:53 <dc_rdfig> Titled item N.

15:51:56 <GregElin> self describing images great, too, for visually impaired, listening to photos on your telephone...

15:52:22 <danbri_dna> dino, extracting XMP is quite easy (nasty hack will do it).

15:52:23 <GregElin> dino - talk to me (hint hint)

15:52:42 <danbri_dna> yup, big accessibility arguments for this stuff --- that's how i got talking to charles about it

15:52:56 <masaka> I want to find metadata of an image by itself, but reference seems ok

15:53:08 <GregElin> Cool. Thank you. Shall we move onto Gerald?

15:53:11 * danbri_dna looks at clock, ok we should hear from Gerald, and then drift into general chitchat...

15:53:18 <danbri_dna> yup, what greg said

15:53:21 <JibberJim> - http://jibbering.com/2002/9/talking.html is my talking version of the shepherds photo.

15:53:35 <JibberJim> some heavy win32 dependancies though.

15:54:19 * danbri_dna gets 'alert - please use IE on Windows', http://jibbering.com/2002/9/img-desc-talk.svg (from Safari/ASV under macosx :)

15:54:35 <danbri_dna> gerald, tell us aboot your image meta stuff...

15:54:39 <JibberJim> 'cor it works - I get the same with ASV6 on windows though, so not that well.

15:54:45 <gerald> :)

15:55:00 <danbri_dna> BLURB:Gerald's show'n'tell

15:55:00 <dc_rdfig> O: Gerald's show'n'tell from danbri_dna

15:55:05 <danbri_dna> logger, pointer?

15:55:05 <danbri_dna> See http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T15-55-05

15:55:17 <gerald> I am writing some software to help me manage my massive collection of bad digital photos: http://impressive.net/software/photo/

15:55:22 <danbri_dna> O:see [http://ilrt.org/discovery/chatlogs/rdfig/2003-08-27#T15-55-05|irc transcript].

15:55:22 <dc_rdfig> Added comment O1.

15:55:39 <gerald> Sample output from a recent version: http://impressive.net/people/gerald/2003/02/16/12-58-26-sm.html

15:56:05 <gerald> It is currently optimized for Unix nerds like me; not much effort has gone into usability or UI work yet.

15:56:19 <gerald> I have mainly been trying to streamline the process to help me publish photo galleries quickly.

15:56:44 <gerald> To use it I run "captivate *.jpg" and it prompts me for titles, descriptions, locations, access control info, quality ratings,

15:56:55 <GregElin> GUI on looking at images and data is nice, though. Obvious. Understated. Clean.

15:56:56 <gerald> and stores the results in RDF and generates XHTML for each photo and the collection.

15:57:11 <danbri_dna> rdf view of the image: http://impressive.net/people/gerald/2003/02/16/12-58-26.rdf

15:57:27 <gerald> My design goals:

15:58:39 * GregElin thinking we are building the different pieces that together could be almost make a new operating system for digital photos...

15:59:03 <GregElin> Very cool

15:59:41 <danbri_dna> If you added this: <foaf:depicts xmlns:foaf='http://xmlns.com/foaf/0.1/' xmlns:wordnet="http://xmlns.com/wordnet/1.6/"><wordnet:Frisbee/></foaf:depicts>

15:59:50 <danbri_dna> ...you'd be ever more pedanticweb compatible ;)

16:00:11 <danbri_dna> I'm so happy the stuff like "Frisbee" is in wordnet. It beats trying to come up with my own set of categories...

16:00:21 <gerald> I want to edit my image metadata using the same text editor I use for everything else (in my case, vim, but you could use emacs or anything else)

16:00:21 <gerald> Everything is stored in plain HTML/RDF/JPEG files, so they're cacheable, cvsable, CDROMable: no PHP or anything else in the way.

16:00:22 <gerald> I want to store metadata externally from the image, because I use CVS to manage my site and I don't want a new revision of the whole file whenever I change a few words in the description.

16:00:23 <bptheo> (not sure what that change of topic name was, must have been a missed keystroke somewhere in the wrong window... sorry)

16:00:29 <gerald> Also, I think there will naturally be many sources of metadata related to any given image (ratings/comments on friends' sites, etc), so you'll need to handle ex

16:00:32 <gerald> ternal metadata anyway.

16:00:38 <gerald> So far this works fairly well for my needs, but it probably has bugs in the way it uses RDF and URIs, feedback welcome.

16:00:48 <gerald> I originally started from http://www.w3.org/TR/photo-rdf/ except I store my metadata externally, and added a few things (access control, quality rating)

16:00:50 <danbri_dna> yup, there'll always be more tan one place to look for yr metadata...

16:01:09 <danbri_dna> the access control and quality ratings side is interesting, I've not looked at that all yet.

16:01:28 * JibberJim notes the example RDF the rdf:about seems to point to the wrong image - not enough ../../ in there.

16:01:30 <danbri_dna> Do you feel the urge to make it any more controlled-value based, eg. for retriveal?

16:02:01 * gerald thinks there are several outstanding bugs in its handling of ../../

16:02:27 <gerald> danbri, not sure I understand. You mean wn:frisbee vs "Frisbee"?

16:03:02 <danbri_dna> yeah, i mean markup that says "this photo depicts a thing that is a member of the class 'wn:Frisbee'

16:03:21 <danbri_dna> so that people can build UIs to distributed dataset based on wordnet categories

16:03:26 <danbri_dna> greg, re *GregElin thinking we are building the different pieces that together could be almost make a new operating system for digital photos...

16:03:50 <gerald> not really. I'm glad others are working on that so I can use it later, but I mainly want a way to publish travel photos etc quickly

16:03:52 <danbri_dna> ...one of the things i ended up realising, having come to metadata via image stuff in large part, was that most of the needs of image metadata aren't really about the image...

16:04:06 <danbri_dna> ...they're about the things in the image, so that image-specific standards were only a small part of the story

16:04:19 <danbri_dna> same thing maybe goes for 'new operating system'...

16:04:40 <danbri_dna> rdf and xml are basis for a new way of representing information... out of which fall some nifty ways of managing images...

16:05:38 * danbri_dna belatedly welcomes masaka to this (somewhat overrun) chat

16:06:04 <danbri_dna> masaka, i had an action from the last (exif-oriented) chat to tell you that your EXIF schema won our beauty contest...

16:06:14 <danbri_dna> it was the most complete/polished in the view of people at the chat...

16:06:30 * danbri_dna remembers gerald using exif metadata markup too...

16:06:37 <danbri_dna> gerald, any lessons from your exif experience?

16:06:40 * mortenf is away: meeting...

16:06:44 <GregElin> Gerald, stuff is great. Anything more to add?

16:07:10 <gerald> it takes me long enough to do the who/what/where part of a few hundred pics when I get back from a trip

16:07:15 <gerald> hmm... I don't do much with exif except slurp it out of the photos and store some of it as RDF

16:07:17 <gerald> greg, that's about it for me

16:07:18 <ndw> ndw is now known as ndw^afk

16:07:36 <bptheo> gerald, probably already seen, but this discussion at Mark Pilgram's blog was interesting: http://diveintomark.org/archives/2003/06/17/rdf_photo_gallery

16:07:38 <danbri_dna> thanks gerald :)

16:07:39 <GregElin> We have a last minute addition to show and tell, Masaka...

16:07:47 <masaka> danbri, yes, I got an email from Karl Dubost

16:08:20 <danbri_dna> masaka, show'n'tell would be great :)

16:08:30 <masaka> Ah, yes, I have short description of my roud-trip idea

16:08:38 <gerald> bptheo, yes, thanks

16:08:42 <masaka>http://kanzaki.com/docs/sw/img-descr20030827.html

16:08:43 <dc_rdfig> P: http://kanzaki.com/docs/sw/img-descr20030827.html from masaka

16:09:12 <masaka> I think embedding metadata in Jpeg is fine, but sometimes hard to maintain

16:09:23 <danbri_dna> P:|Masaka's show'n'tell

16:09:23 <dc_rdfig> Titled item P.

16:09:35 * mattb agrees, finds external storage of metadata easier for hacking up with tools to hand

16:09:39 <JibberJim> ooh so close to my RDF syntax...

16:09:45 <masaka> so, I want to find a standard way to connect image and metadata both way

16:09:46 * GregElin nods

16:09:49 <mattb> nice to dynamically embed metadata into a jpeg on download perhaps

16:09:55 <karlcow> :) I have to reply to the positive answer of masaka. :)

16:10:11 <danbri_dna> +1 re external/tools

16:10:38 <gerald> find . -name *rdf | xargs egrep -i location.*ottawa

16:12:08 * danbri_dna finds the mouseovers in http://kanzaki.com/courier/xs/works/2003/imagedesc/0827.rdf.html -- nice work :)

16:12:11 <GregElin> How far are you with your coding, masaka?

16:12:27 <danbri_dna> and again the data structures here seem pretty close to the svg/imagemap stuff, and to greg's tools

16:12:36 <GregElin> And what is your day job, too?

16:13:29 <danbri_dna> whose day job? masaka's?

16:13:37 <masaka> I'm not a professinoal programmer, rather just interested in metadata

16:13:40 <GregElin> re masaka's rollovers: another case for having a tool so end-user can choose how they want to present descriptions with image.

16:14:04 <GregElin> yes, Masaka's. This is very nice.

16:14:20 * danbri_dna assumes nobody minds this chat running on and on... i'm always happy to hear about image metadata stuff (so assume the same true of everyone ;)

16:14:24 * danbri_dna remembers timezones though...

16:14:35 <danbri_dna> masaka, what markup do you use for regions?

16:14:37 * JibberJim will just add a similar thing I found...

16:14:39 <JibberJim> Digimon/Imarea - http://db.cs.helsinki.fi/~msisalmi/twiki/bin/view/Digimon/WebHome an University of Helsinki software project that seems to be creating an Annotator with RDF/SVG.

16:14:42 <GregElin> Masaka, can you say one or two more words on why you think roundtrip is important, and anything you learned or challenges you had doing it?

16:14:56 <libby> very nice

16:15:20 <masaka> In this example, I use xhtml imagemap, but can easily translated into SVG

16:15:33 <danbri_dna> Jim, could you chump that url so we can find it again later?

16:15:42 <danbri_dna> masaka, ah cool, i didn't realise!

16:16:08 <JibberJim> Do you think you could work with a similar RDF to http://jibbering.com/rdf/example.rdf masaka? so we can unify to a single vocab?

16:16:11 <danbri_dna> I have an aspiration to go try to persuade the XHTML Working Group to improve imagemaps to allow for such apps to attach metadata to regions...

16:16:22 <danbri_dna> +1 on single vocab... if at all possible

16:16:26 <JibberJim>http://db.cs.helsinki.fi/~msisalmi/twiki/bin/view/Digimon/WebHome

16:16:26 <dc_rdfig> Q: http://db.cs.helsinki.fi/~msisalmi/twiki/bin/view/Digimon/WebHome from JibberJim

16:16:36 <JibberJim> Q:=Digimon/Imarea

16:16:36 <dc_rdfig> Replacement must be a valid URL.

16:16:53 <JibberJim> Q:A University of Helsinki software project that seems to be creating an Image Annotator with RDF/SVG.

16:16:53 <dc_rdfig> Added comment Q1.

16:16:56 <JibberJim> Q:|Digimon/Imarea

16:16:56 <dc_rdfig> Titled item Q.

16:17:23 * gerald 's connection is/was very laggy btw, sorry for delayed responses earlier

16:18:14 <danbri_dna> masaka, nice demo... do you have any plans for extending etc? where next with it?

16:18:59 <masaka> This is my current interest and will work for a while

16:19:20 <danbri_dna> do keep us posted (eg. msg to www-rdf-interest or irc ...)

16:19:23 <danbri_dna> we should wrap up...

16:19:34 <masaka> Yep

16:19:41 <danbri_dna> if folks have any closing thoughts, other image-meta URLs to share, now would be a good time...

16:19:49 * danbri_dna also hoping edd can sneak in a quick foafbot demo

16:20:11 <GregElin> I would like to quickly mention another tool: photofinder http://www.cs.umd.edu/hcil/photolib/

16:20:20 * danbri_dna remembers something else to plug, beyhond Wordnet, as a bulk taxonomy for depictions

16:20:25 <edd> can you throw the swtches danbri? i'm busy working elsewhere right now

16:20:30 <GregElin> Should we do foafbot now?

16:20:41 <danbri_dna> I'll try, edd.

16:20:42 <danbri_dna>http://tap.stanford.edu/tap/tapkb.html

16:20:42 <dc_rdfig> R: http://tap.stanford.edu/tap/tapkb.html from danbri_dna

16:21:01 <danbri_dna> R:|TAP Knowledge Base -- another db could be used for depictions. Is a big list of well known entities...

16:21:02 <dc_rdfig> Titled item R.

16:21:18 <GregElin> And can people list the relevant ML's here:

16:21:18 <danbri_dna> OK, thanks everyone for coming! Sorry we overran, clearly too much stuff to squeeze into an hour...

16:21:21 <GregElin> newsML,

16:21:25 <GregElin> Dublin core

16:21:27 <zool> foafbot, picture of edd and danbri

16:21:40 <foafbot> Sorry, couldn't find any matching pictures.

16:21:49 <danbri_dna> FOAF, Wordnet, RSS, Exif-...

16:21:50 <GregElin> That's way cool.

16:21:51 <edd> it's true. we've never been codpicted.

16:21:54 <zool> foafbot, picture of edd and zool

16:22:01 <foafbot> Picture at http://matthau.yoz.com/cam/20020609/bots_talk4.jpg -- according to Edd Dumbill

16:22:01 <foafbot> Caption: From left to right: Edd, Yoz, Matt Webb, Jo Walsh, Celia Romaniuk

16:22:24 <GregElin> Oh...this is very cool. I have an app for that...

16:22:38 <danbri_dna> foafbot is edd's RDF IRC bot, it crawls a set of FOAF files and other linked RDF metadata, merges it into a big database, but (very interestingly) keeps track of who-said-what metadata

16:22:58 <zool> foafbot, edd's name

16:23:00 <foafbot> edd's name is Edd Dumbill, according to Dave Beckett, Edd Dumbill, Jason Bell, Jo Walsh, Matt Biddulph, Schuyler Erle; and anonymous sources Anon10, Anon1008, Anon1038, Anon1052,

16:23:00 <foafbot> Anon1105, Anon1153, Anon1179, Anon1196, Anon1227, Anon1228, Anon1283, Anon1352, Anon1586, Anon1689, Anon1734, Anon1789, Anon1837, Anon236, Anon29, Anon34, Anon390, Anon415,

16:23:00 <foafbot> Anon419, Anon475, Anon617, Anon63, Anon86, Anon9, Anon90

16:23:09 <danbri_dna> Also is UTF-8 happy, so if your IRC client groks UTF-8, Japanese etc characters show up

16:23:55 <_joshua> provenance is necessary, of course

16:24:00 <danbri_dna> foafbot, nori's name?

16:24:00 <foafbot> nori's name is のり, according to Anon197

16:24:08 <danbri_dna> foafbot, picture of nori?

16:24:13 <foafbot> Picture at http://diary.noasobi.net/img/nori_pank.jpg -- according to Anon198

16:24:13 <foafbot> Caption: いつの写真だ :p)

16:24:13 <foafbot> Picture at http://diary.noasobi.net/img/solo2.jpg -- according to Anon197

16:24:27 * danbri_dna sees caption in Japanese; wonders who else can...

16:24:47 <GregElin> This is *so* much better than google image search.

16:25:01 <danbri_dna> anyway, just more to demo that image metadata is worthwhile for non-visual apps too

16:25:04 <danbri_dna> yes, defintely :)

16:25:11 <danbri_dna> and the markup is really pretty simple...

16:25:17 <_joshua> Greg: alltheweb's image search is way better than google's

16:25:21 <danbri_dna> Greg, we should wrap up. Anything more to add?

16:25:29 <_joshua> Unfortunately, Edd's bot is way susceptible to spammers.

16:25:39 <JibberJim> and the markip is generated by mine, rdfpic, and libby's image annotators at the moment.

16:25:40 <GregElin> So we could extend Foafbot to read other image description data, too?

16:25:45 <zool> _joshua, in what sense?

16:26:41 <danbri_dna> greg, yes. (i) get the image descriptions into RDF (ii) write python code to map from irc queries in natural language into RDF API calls

16:26:41 <_joshua> It's just the general spam problem

16:26:41 <edd> zool: in the same way all of FOAF is

16:26:47 <JibberJim> It's all RDF GregElin, so it's just what it's been taught to query. foafbot is backed by an RDF store.

16:26:49 <danbri_dna> edd, got an url for foafbot src?

16:26:53 <_joshua> People could make junk FOAF data. Just like the normal web thing

16:27:12 <_joshua> Clearly we need an irc bot that speaks n3

16:27:13 <danbri_dna> foafbot's one of my fave rdf apps btw... :)

16:27:16 <_joshua> also humans that speak n3

16:27:17 <JibberJim> Yep _joshua, which is when you'll start wanting to do more with the GPG stuff.

16:27:18 <GregElin> Well, I've been thinking about how p2p filesharing technology might be used to move your image meta data to a database...then people could search and get images from your machine. Name spaces or internal data would control public/private, etc.

16:27:26 <edd> danbri: http://svn.usefulinc.com/svn/repos/branches/foafbot/foafbot-twisted/src/

16:27:34 * edd mumbles n3 over my dead rotting body

16:28:05 * JibberJim decides a living Edd more important than an N3 foafbot.

16:28:22 * edd adds disclaimers about that URL being live subversion repository

16:28:28 <GregElin> I just want to make an observation and pat this group's collective and individual backs...

16:28:38 <Morbus> RDF::Simple requires Template Toolkit?

16:28:38 <dc_rdfig> Label RDF not found.

16:28:43 <Morbus> Wow, that rules it out for me immediately

16:28:51 <Morbus> Template Toolkit is HUUUUGE, and not crossplatform.

16:28:55 <danbri_dna> edd, could you link http://svn.usefulinc.com/svn/repos/branches/foafbot/foafbot-twisted/src/ from foafbot homepage?

16:28:59 <edd> FOAFbot homepage is http://usefulinc.com/foaf/foafbot - released source availble from there

16:29:07 <edd> danbri_dna: reluctantluy

16:29:10 * danbri_dna nods to Greg, some really impressive work going on...

16:29:13 <zool> TT is insanely useful

16:29:16 <Morbus> oh, just the serialiser needs it.

16:29:21 * danbri_dna promises not to critique your indenting style

16:29:24 <Morbus> sure, its useful, but its too big.

16:29:25 <_joshua> TT isn't crossplatform?

16:29:25 <zool> first i've heard of it not being crossplatform

16:29:28 <Morbus> text::template is handier.

16:29:35 <Morbus> _joshua: doesn't work on mac classic, last i knew.

16:29:40 <Morbus> contains XS code that hasn't been ported.

16:29:45 <_joshua> Ah.

16:29:57 <_joshua> Lemme go blame Chris Nandor

16:29:57 <GregElin> I think including images with this data, or better put "intertwingling" images with this data, really brings together some long sought goals of computer human interface.

16:30:00 <Morbus> likewise, it bulks up installation necessities for end user programs.

16:30:05 <Morbus> (/me is thinking of amphetadesk)

16:30:11 * zool wrinkles nose

16:30:25 <Morbus> apparently, its only the serializer though.

16:30:31 <Morbus> the parser for the array buckets is only sax.

16:30:35 <Morbus> so i could build my own serializer.

16:30:44 <zool> is a good point, and am not making heavy or interesting use of TT ... just expectations of endusers having it installed, where i thikn of endusers as people running webservers, bots etc

16:30:57 * danbri_dna declares #rdfig officially re-intertwingled... image chat leftovers can mix with usual goings on...

16:30:59 <_joshua> I should look at TT some more

16:31:00 <wware> Belated thanks to libby and jql for explanation of "rdf:about"

16:31:00 <Morbus> zool: yeah, i can't think like that, unfortunately.

16:31:02 <GregElin> Frankly, when I see that foafbot demonstration, I really feel the potential of RDF and the semantic web. Getting a picture back is far more meaningful than a stock quote.

16:31:02 <danbri_dna> thanks again all! cheers...

16:31:10 <Morbus> 99% of my amphetadesk users don't even know what perl is, so i can't assume they have it installed.

16:31:11 <_joshua> I'm almost tempted to use TT and Class::DBI instead of HTML::Mason for some stuff

16:31:12 <GregElin> Whole different level of experience.

16:31:14 <Morbus> which makes things interesting ;)

16:31:32 * danbri_dna has to head off, but agrees with Greg... I'm not sure why but i do find the imaeg demos v compelling

16:31:54 <danbri_dna> like codepiction paths through foaf, more interesting than 'knows' paths...

16:31:59 <zool> you have XML::RSS to do the amphetadesk job, though

16:32:04 <Morbus> xml::Rss sucks.

16:32:08 <Morbus> its four times slower than xml::simple.

16:32:09 <GregElin> Because, Dan...the photos are more of a metaphor for people and things than text...

16:32:17 <Morbus> i can't use it in enduser programs.

16:32:22 * zool nods

16:32:26 <GregElin> Yes...thank you danbri for helping pull this together.

16:32:26 <Morbus> far too slow for 40+ feed parsing at once.

16:32:39 <Morbus> but yes, i've looked into it ;)

16:32:43 <GregElin> Uh...Morbus...could we wrap up the image discussion?

16:32:50 <zool> i'd happily consider losing the TT dependencies... but tbh the sax-based parser is not that fast either

16:32:54 <GregElin> Any last minute thoughts? Anyone?

16:32:59 <zool> sorry GregElin

16:32:59 <Morbus> i use xml::simple to parse all flavors of rss, all scriptingNews, and all OPML.

16:33:08 <Morbus> sorry GregElin: didn't know there was a planned discussion.

16:33:25 <JibberJim> GregElin could you export RDF?

16:33:29 <GregElin> Well, we are wrapping up and returning to regularly scheduled arguments...

16:33:34 <danbri_dna> greg, and you yours.

16:33:43 <danbri_dna> Next time I'll do a better job w/ pre-announcements etc :(

16:33:47 <zool> Morbus, i'd be curious to see benchmarks ;)

16:33:58 <Morbus> zool: i had some a year ago.

16:34:03 * danbri_dna tries to blame last weeks mail outage but it was 90% human flakyness (mine)

16:34:08 <GregElin> Same, here, Dan. But I think the turn out was great and the show and tell great. I'm going to go play with everything.

16:34:25 <danbri_dna> greg, how do you feel w.r.t. extending your stuff to use full polygons, as well as rectangles?

16:34:39 <danbri_dna> yup, we squeezed a lot in...

16:34:40 <GregElin> I'd love to do it. No problem.

16:34:48 <JibberJim> and RDF?

16:34:50 <danbri_dna> cool, i'd love to see that

16:34:51 <GregElin> Sure.

16:34:54 <danbri_dna> and yeah, rdf please!

16:34:57 <danbri_dna> ;)

16:34:58 <JibberJim> cool.

16:35:21 <JibberJim> We've got no action items, and no schedule for another chat?

16:35:21 <GregElin> I'd like to see a lightweight syntax and RDF so both could be used as the market shakes out.

16:35:47 <danbri_dna> Oh, a reminder. See http://rdfig.xmlhack.com/ for links to IRC logs and the URL annotations we made during the meeting.

16:35:53 <GregElin> I don't think so. Jibber. I think we digest. I think September is busy for everyone. I think I start an email list or something.

16:36:00 <danbri_dna> RDF subset could be a lightweight syntax...

16:36:15 <Wookiee-ZZZZZ> Wookiee-ZZZZZ is now known as mdupont

16:36:18 <GregElin> Email list sound like a good idea, or should this all just be part of an existing list?

16:36:26 <JibberJim> not that I want any action items, but I'd like to see someone looking at the RDF vocabs currently out there for doing this, and we need to schedule a meet. I've not got any mailing list room left, so you'll have to news.gmane.org if you want me :-(

16:36:46 * danbri_dna was looking into new list possibilities at w3.org re geo etc., but it got bogged down by waiting on plans for next phase of Interest Group etc

16:36:51 <GregElin> Well how about a late september or early october meeting?

16:37:07 <danbri_dna> I'd love to host an image description sublist of RDF IG at w3c, but would need a bit of thought re scoping etc...

16:37:14 <GregElin> Dan, I'm happy to start something and then see it move elsewhere later.

16:37:23 <danbri_dna> I think there's clearly interest / expertise around...

16:37:41 <GregElin> I'll start a wiki/blog on this topic. And we can go from there.

16:38:13 <danbri_dna> Feel free to use http://esw.w3.org/topic/ImageDescription and nearby for wiki; if you make a blog, would be cool to link it from there

16:38:36 <GregElin> I'm overcommitted the next 2 weeks, but I'd like to reparse this discussion, see who doing what...and pull something together, some overview.

16:39:07 <GregElin> I think the most important thing is that we all continue to work on overlapping pieces with particular focus we like and see what emerges naturally.

16:39:14 <danbri_dna> OK, let's sync back up in irc when you've got time. I'm normally here and in #foaf...

16:39:41 <GregElin> Sounds great, Danbri. Thanks again. I think this was a milestone.

16:39:44 <danbri_dna> next for me is chase up exif thing w/ masaka and w3.org hosted namespace. A mailing list could quite naturally fall out of that, though no promises yet...

16:39:58 * danbri_dna grins, nods, waves and rests his typing wrists for a while...

16:40:13 * GregElin smiles back at dan's waving.

16:40:34 <GregElin> I'm closing out here, but will be in FOAF for a bit...

16:40:40 <ndw^afk> ndw^afk is now known as ndw

16:40:59 <GregElin> Bye all...

16:41:09 <danbri_dna> 'k cu

16:41:40 <adamhill> adamhill is now known as _adam_lunch

16:42:44 <wware> wware is now known as WillWare

17:25:58 <shellac_> shellac_ is now known as shellac

17:57:00 <mdupont> mdupont@introspector:~/cwm$ python xml2rdf.py

17:57:00 <mdupont> Traceback (most recent call last):

17:57:00 <mdupont> File "xml2rdf.py", line 26, in ?

17:57:00 <mdupont> RESOURCE = notation3.RESOURCE

17:57:01 <mdupont> <AttributeError: 'module' object has no attribute 'RESOURCE'

17:57:06 <mdupont> DanC: you have an idea?

17:57:35 <mdupont> rdfx2kif.py:#introduced SYMBOL to replace notation3.RESOURCE

17:57:44 <mdupont> maybe i need to get a different version

17:58:24 <mdupont> RCS file: /sources/public/2000/10/swap/xml2rdf.py,v

17:58:25 <mdupont> Working file: xml2rdf.py

17:58:25 <mdupont> head: 1.20

18:06:32 <WillWare> When people add underscores to their nicks, does that mean they're away from their computers?

18:14:34 <sandro> No, WillWare. It means the nick they want is taken, probably by another session of theirs.

18:23:31 <WillWare> I finally managed to write some Python code to translate RDF/XML to triples. Partially broken, I think, but it's a start. I'd like to get far enough along to fix the broken parts of the code, merge a couple of graphs into a larger graph, and do some queries on the graph. I picked up an "intro to graph theory" book and maybe it'll give me ideas.

18:24:17 <WillWare>http://groups.google.com/groups?selm=3F4C304A.283DAB0B%40alum.mit.edu

18:24:17 <dc_rdfig> S: http://groups.google.com/groups?selm=3F4C304A.283DAB0B%40alum.mit.edu from WillWare

18:35:05 <_adam_lunch> _adam_lunch is now known as adamhill

18:36:54 <sandro> WillWare, is any particular reason you're rolling your own instead of, maybe, using redlib?

18:37:08 * sandro is a big fan of learning by rolling his own.

18:43:30 <WillWare> There seems to be a lot of very confusing "tutorial" material about RDF and the semweb, so this is an exercise in self-education. If I had an actual app in mind, I'd try to learn about existing things.

18:46:16 * sandro nods understandingly

18:53:23 <WillWare> My employer is a cellphone firmware developer (www.savaJe.com). It would be interesting to do something semweb-ish and location-aware in cellphones, like querying for nearby restaurants, hotels, gas stations, etc, or maybe coordinating meetings with people, and negotiating preferences like restaurant cuisine and food allergies.

18:53:39 <sandro> Very much so!

18:55:18 <danbri> re restaurants, see http://esw.w3.org/topic/RestaurantRecommendation for points to datasets, works in progress etc

18:55:36 <WillWare> Here in the Boston area we have a few little neighborhood 802.11 hotspots. I've read a little bit about the mudlondon thing and it'd be kinda cool to have something like that around here.

18:55:50 <GabeW> WillWare: heh.. we did stuff like that at Airflash without semweb stuff - just big databases

18:56:20 <GabeW> but it would have been nice to plug into the web as a data source ;-)

18:56:48 <WillWare> Well, yeah, you don't need semweb for that. But maybe RDF would be less of a kludge, and maybe the semweb will take off big-time, etc.

18:57:09 <GabeW> and we did this in 1999 when there were no phones that were actually location aware - OOPS

19:00:35 <WillWare> The hardware we are working with now is not location-aware, unless we dig into the protocol stack and see which cell we're connected to. Then we'd have to go to some website to find the cell's lat and long, and it's still only approximate. But it's way better than complete ignorance, esp if you've driven thru upstate NY looking for a hotel at 1AM when some unknown event has sucked up all the hotel rooms in a 50-mile radius.

19:01:03 <danbri> if you can get the cellid / location db, sounds worthwhile

19:02:25 <sandro> All the phones I can get from Verizon right now (I was just looking -- I need one) are location aware. I'm not sure exactly what that means, though.

19:04:40 <WillWare> Is that in the USA? I just upgraded my Verizon phone to the new LG phone with a camera, and I don't think it can tell me anything about location. The guy at the Verizon store was saying that location awareness is a year or two away, IIRC.

19:10:38 * ndw mutters location aware maybe, useful overseas no.

19:10:40 * ndw pouts

19:10:53 * ndw wishes his dratted Verizon contract expired before March of next year

19:17:19 <WillWare> Yup, the phones we tinker with are GSM (Verizon is CDMA). There are a couple GSM providers in the US, T-Mobile and AT&T, but Verizon has much better coverage. T-Mobile is a little tempting because then maybe I could do 802.11 in Starbucks. But the coverage is too big a draw, at least for the present.

19:22:32 <ndw> I want the TMobile sidekick thingy DanC has :-)

19:27:18 <WillWare> The sidekick thing does look like fun, but I'm hoping for something I could plug a laptop into. I dunno if the sidekick has an RJ45 plug on the side.

19:29:00 * ndw wonders if any of the assembled company has some exim MTA experience to field an OT question...

19:31:26 * ndw shrugs. Oh, well.

19:32:20 <JibberJim> WillWare, I thought the 911 cellphone thing forced providers to know location of phones, therefore you could get it back from the provider?

19:32:37 * GabeW laughs out loud at JibberJim

19:32:56 <JibberJim> Most of the providers offer it here in the UK...

19:32:57 <GabeW> providers have been fighting E911 tooth and nail

19:33:11 <GabeW> here in the US at least

19:33:38 <JibberJim> Yes, odd that in the rest of the world it's the providers adding it so as to provide services, nothing to do with ES.

19:33:56 <GabeW> American CellCos are back-asswards to the extreme

19:33:57 <WillWare> I dunno too much about politics/policy. I only know two ways to get location: GPS, and lookup cellid in a database. GPS is AFAIK not in use yet. I don't know which phones are doing anything with cellid lookups.

19:34:09 <GabeW> there are also triangulation technologies

19:34:20 <GabeW> basically using cell site data

19:34:42 <JibberJim> That's the provider, not the phone that can do that though GabeW...

19:35:10 <GabeW> right

19:35:15 <WillWare> You can triangulate maybe by signal strength from the cells, though that might be a little bogus in an urban environment with buildings and reflections.

19:35:29 * GabeW doesn't pay attention to LBS any more

19:35:49 <JibberJim> When ChrisGoad showed us at a mini-foaf-meet it was pretty accurate in London...

19:36:08 <WillWare> Yeah, I'm just conjecturing, no real world experience.

19:36:28 <JibberJim> Mines very limited...

19:41:02 <danbri> cellid data seems too hard to come by

19:41:39 <_joshua> didn't geowanking HAVE this discussion already?

19:44:27 <WillWare>http://www.emtel.etsi.org/Workshop/Non-presented_papers/E112%20White%20paper.pdf

19:44:27 <dc_rdfig> T: http://www.emtel.etsi.org/Workshop/Non-presented_papers/E112%20White%20paper.pdf from WillWare

19:45:49 <WillWare> Such location DBs exist in Europe apparently, but they might only be available within the provider's internal network.

19:47:57 <danbri> yup, they certainly exist... those companies need to remember where they stashed all their funny antennas...

19:48:03 <danbri> but public don't get to see the detail

19:54:23 <DanC> I brought my car into the shop with squeaky breaks. The guy just called to say it needs a new EGR cylinder (~$300), a rear main seal replaced (~$700), and 120,000mi service (tuneup, breaks, etc., didn't write down the price).

19:54:38 <ndw> ouch

19:54:38 <DanC> I wouldn't know an EGR cylinder if it bit me in the ass.

19:55:10 <DanC> But my brother-in-law services cars for a living, and when we've called him to double-check prices from this dealer, it's always checked out in the past.

19:55:50 <DanC> I'm *not* going to make the mistake of having the timing belt replaced twice within 6 months just because I didn't review my records, though...

19:56:02 <DanC> that is: not going to make that mistake *again*.

19:56:21 <ndw> oh my.

19:56:45 <ndw> at 120,000mi, is it still worth > $1000?

19:57:01 <ndw> I bailed on my truck when it needed $800 of repair and it had a blue book value of $700

20:01:47 <DanC> it's still a nice car. probably can't replace it with something that costs less than $1000/yr.

20:03:18 <ndw> Yep. That makes sense, too.

20:03:19 <DanC> I just feel silly spending ~$1500 on the say so of some guy I hardly know from adam.

20:03:37 <DanC> I'm an engineer, for chrissakes. I'm used to knowing how things work.

20:03:49 <ndw> Just so long as he doesn't suggest you replace the timing belt again ;-)

20:07:44 <adamhill> adamhill is now known as _adam_espresso

20:52:47 <JonB> I once tried to get a guy to replace a $100 part on a car, and he refused... said that I'd soon need $3000 of parts at which point the car would have been worth $3000 ... I leased a Honda

20:53:58 <ndw> OT: what's the wget magic to set the accepts: header?

20:53:58 <dc_rdfig> Label OT not found.

20:54:05 <ndw> oops

20:56:49 <sbp> (-A or --accepts)

20:57:34 <sbp> (hmm, no, wait. that's for recursive gets, matches file name. argh)

20:58:27 <sbp> (ah, --header='Accept: ...')

21:01:01 <ndw> Thank you.

21:01:14 <ndw> I confirmed my fears.

21:01:19 <sbp> np

21:01:19 * ndw goes off to have a good scream and shout

21:01:35 <sbp> aw. problem? :-)

21:07:52 <ndw> ndw is now known as ndw^having-a-goo

21:08:02 <ndw^having-a-goo> ndw^having-a-goo is now known as ndw^ranting

21:11:40 <JonB> ndw: once you're done ranting :-) take a look at this:

21:12:06 * JibberJim is still waiting for ndw's rant.

21:12:21 <JonB>http://www.openhealth.org/RDF/parser/RELAXNGccRDF.rng

21:12:21 <dc_rdfig> U: http://www.openhealth.org/RDF/parser/RELAXNGccRDF.rng from JonB

21:12:39 <JonB> U|A RELAXNGCC RDF parser

21:12:54 <JonB> U:|A RELAXNGCC RDF parser

21:12:54 <dc_rdfig> Titled item U.

21:14:07 <JonB> although relaxngcc barfs on the rules for property attributes :-(

21:14:47 <JonB> as they're attribute * - (local:*|rdf:ID ...) ... too complicated for current implementation

21:15:14 <JonB> in any case the idea is to annotate a schema with stuff that emits triples

21:17:09 <zool> hi raelity

21:17:22 <raelity> hey zool

21:17:29 <raelity> just logging on to find you, actually

21:17:40 * zool ocwers

21:17:44 <zool> good to see you :)

21:18:00 <_adam_espresso> _adam_espresso is now known as adamhill

21:18:18 <JibberJim> another zool stalker - you seem to attract them zool, always people coming looking for you.

21:19:04 <_joshua> raelity: are you on the board for ETcon next year?

21:20:08 <adamhill> adamhill is now known as hillbob

21:20:43 * zool denies everything

21:21:05 <hillbob> hillbob is now known as strongbob

21:25:17 <raelity> _joshua: on board? you mean as program chair? indeed!

21:33:04 <dngor> Same raelity as blosxom? I've got an irc bot that manages a blosxom's file tree...

21:35:13 <raelity> dngor: aye, that'd be me

21:35:19 <raelity> dngor: do tell

21:35:40 <dngor> It's unreleased, and the pilot version's not running. The SourceForge project name is "ircxom".

21:36:06 <raelity> so what's it do?

21:36:07 <dajobe> yeek, how the heck do you pronounce that?

21:36:17 <raelity> hey morbs

21:36:41 <dngor> Anything addressed to it goes on the web. http://poe.perl.org/plog/ is the stagnant location of the pilot blosxom.

21:36:46 <JibberJim> like Erksome for me dajobe...

21:37:05 <dngor> It also knows commands like title, mv, and rm for manipulating categories and things.

21:37:09 <raelity> chump blosxom, basically?

21:37:19 <dngor> I guess yes.

21:37:24 <raelity> groovy

21:37:27 <raelity> i'd love to try it out

21:37:32 <raelity> any chance i could run it somewhere?

21:37:34 <raelity> or you could?

21:37:36 <raelity> #blosxom?

21:37:53 <Morbus> raelity!

21:37:54 <dngor> Let's see if I have one set up on this machine.

21:39:01 <raelity> hey, i know another morbus who's quite an arachnoid

21:39:04 <raelity> ;-)

21:39:17 <Morbus> heh, heh.

21:39:47 <strongbob> strongbob is now known as adamhill

21:42:41 <dngor> ircxom's web presence is at http://poe.dyndns.org/~troc/plog/

21:43:00 <dngor> ircxom: La la la, test test test.

21:43:01 <ircxom> Logged as #118.

21:43:45 <dngor> ircxom: rm #118

21:43:45 <ircxom> Entry #118 removed.

21:46:05 <Morbus> how in lord's name do you pronounce ircxom?

21:46:45 <GabeW> erk-zom

21:46:56 <GabeW> ?

21:47:05 <GabeW> irksome

21:47:06 <Morbus> irksome? heh, heh.

21:47:37 <zool> dngn[and|or]++ # omnipresent, metathematic

21:47:55 * mortenf is back (gone 05:41:15)

21:49:18 <dngor> zool: :)

21:57:13 <ndw> DanC, you posted a bit about SSH1 vulnerabilities a while back. Got a pointer to the incantations that forbid SSH1?

22:03:11 <ndw> DanC, Nevermind. Found it http://rdfig.xmlhack.com/2003/07/02/2003-07-02.html

22:06:05 <_joshua> dngor: anyone do an IMAP backend for POE?

22:06:31 <dngor> Not yet. crab may be working on one soon.

22:06:51 <_joshua> fascinating

22:07:23 <_joshua> that'd be hella useful

22:16:57 <ndw> ndw is now known as ndw^afk

22:25:55 <mdupont> mdupont is now known as Wookiee-ZZZ