Semantic Web Interest Group IRC Chat Logs for 2003-12-10

00:33:36 <md-777> md-777 is now known as md-zzz

00:46:54 <D[a]vey> D[a]vey is now known as Davey

02:06:48 <Davey> Davey is now known as D[a]vey

02:31:25 <D[a]vey> D[a]vey is now known as Davey

04:40:53 <DanC_jam> .time

04:40:53 <datum> Wed, 10 Dec 2003 04:40:53 GMT

05:07:40 <DanC_jam> .time

05:07:40 <datum> Wed, 10 Dec 2003 05:07:40 GMT

06:04:58 <D[a]vey> D[a]vey is now known as Davey

06:59:44 <Davey> Davey is now known as D[a]vey

08:35:46 <arnarl> hi

09:30:45 <AnkhGone> AnkhGone is now known as Ankh

10:29:00 <md-zzz> md-zzz is now known as mdupont

11:18:55 <dajobe> dajobe has changed the topic to: RDF, Semantic Web and XML 2003 Conference Philadelphia day 2. Scratchpad http://rdfig.xmlhack.com/ logs http://www.ilrt.bris.ac.uk/discovery/chatlogs/rdfig/

11:28:53 <dajobe>http://www.xmlconference.org/xmlusa/2003/wednesday.asp

11:28:53 <dc_rdfig> A: http://www.xmlconference.org/xmlusa/2003/wednesday.asp from dajobe

11:29:00 <dajobe> A:|XML 2003 Wednesday 10 December

11:29:00 <dc_rdfig> Titled item A.

11:29:33 <dajobe> A:their [http://www.xmlconference.org/xmlusa/2003/blogroll.asp|blog roll] isn't growing very fast

11:29:33 <dc_rdfig> Added comment A1.

11:29:33 * Ankh waves to Dave

11:29:38 <dajobe> hiya

11:29:46 <Ankh> how's things??

11:29:49 <dajobe> good good

11:29:59 <Ankh> is good conference so far?

11:30:02 <dajobe> I'm in Bristol

11:30:10 <Ankh> oh! heh

11:30:17 * Ankh working on converting his image metadata to something closer to RDF

11:30:38 <dajobe> yeah, it's a trend ;)

11:31:29 <Ankh> heh

11:32:18 <Ankh> I search it with xml query right now, trying not to make the rdf too complex for my searching skills!

11:33:00 <dajobe> did I see that you mentioned using Galax?

11:34:40 <Ankh> you might have done :)

11:34:44 <Ankh> I do, that and saxon

11:35:00 <Ankh> Galax has some strong typing support though, which helps me

11:36:08 * dajobe watches Ankh shift into a vendor-neutral stance

11:36:29 <dajobe> saxon's in java?

11:36:33 <Ankh> yes

11:37:02 <dajobe> which is correct: XML Query or XQuery ?

11:37:19 <Ankh> XML Query, although people use XQuery for short, that's oK too

11:38:29 * Ankh contemplates dublin core for location and decides it can wait for next revision

11:38:42 <Ankh> "read my lips: no new namespaces"

11:38:57 <dajobe> is that you saying it? or quoting?

11:39:17 <Ankh> me saying it parodying a foreign politician.

11:39:54 <danbri> "read my lips: no nude taxis"

11:40:46 <Ankh> er

11:40:48 <Ankh> darn! :-(

11:40:58 * mdupont looks for nude taxis

11:42:07 <mdupont> http://www.sptimes.com/2002/10/29/Columns/Singing_the_digital_b.shtml

11:42:41 * Ankh worries

11:42:54 <mdupont> ahhhh

11:43:06 <mdupont> nude taxis =~ new taxes

11:43:10 <mdupont> my gawd

11:43:35 * mdupont is not upbeat enought for this channel

11:43:44 * mdupont makes new coffee

11:47:47 <danbri> DanC, did you speak w/ PatH re RDF Semantics WD at all since rdfcore on friday?

12:03:43 <sandro> Any W3C'ers using homer right now? It kicked me off and is asking for a password when it shouldn't.

12:04:02 <danbri> known problem :(

12:04:29 <sandro> someone's workin' on it?

12:04:48 <danbri> my shell session died eventually, after some errors, but ssh tunnel stayed up. you can probably tunnel to another machine and rejoin us in team irc.

12:04:53 * Ankh is still on homer

12:04:57 <danbri> failing that, cu in #sysreq

12:05:13 <Ankh> (using ssh)

12:05:44 <Ankh> oops, but then I told mutt to save and got disconnected :-X

12:09:51 <sandro> yeah -- I sent some rdfig msgs earlier that fell into the homer abyss:

12:10:05 <sandro> <sandro> Do you know the state of the art in RDF image metadata? (I don't)

12:10:10 <sandro> <sandro> I'm still kinda stuck on whether to use the same URI for the abstract (resolution/format independent) photograph and a web page about the photograph. I can get the right behaviors with a redirect in there, but I don't have a solid argument about why you NEED the redirect.

12:10:14 <sandro> <sandro> (and of course "image" is not necessarily the same as "photograph". I'm focussing on photographs.)

12:11:30 <danbri> yeah, I've lately glossed over all those work/manifestation/item/copy distinctions, just used foaf:Image

12:11:48 <danbri> usually giving instances of that as URIs for JPEGs etc

12:12:06 <danbri> but it doesn't feel right, if you want to talk clearly about alternate versions of the same image, etc.

12:12:17 <danbri> thumbnail, orig high res version, etc.

12:15:36 <mdupont> D[a]vey: Hey, this works! thanks for the support http://asimov.thefrontnetworks.net/rdf/rdfapi-php/test/custom_test.php

12:15:48 <mdupont> php, here i come

12:16:01 <D[a]vey> haha

12:16:19 <D[a]vey> mdupont: toSting ?

12:17:06 <D[a]vey> mdupont: also, Warning: Call-time pass-by-reference has been deprecated - argument passed by value; If you would like to pass it by reference, modify the declaration of [runtime function name](). If you would like to enable call-time pass-by-reference, you can set allow_call_time_pass_reference to true in your INI file. However, future versions may not support this any longer. in /usr/local/www/data-dist/rdf/rdfapi-php/api/rdql/RdqlMemEngine.php on l

12:19:27 <mdupont> but is works

12:19:50 <mdupont> i guess is should turn that option on

12:20:16 <mdupont> now all i need a way to run cwm

12:20:18 <mdupont> from php

12:20:22 <mdupont> ... :)

12:20:48 <mdupont> is there a cwm xmlrpc support?

12:20:52 <mdupont> or soap

12:27:31 <DanC_jam> no, danbri, haven't contacted Pat. nor Brian, for that matter.

12:27:56 <Ankh> danbri - for the image stuff I've been doing I have multiple versions/representations of every image, and am determined to keep them separate

12:28:24 <D[a]vey> what is "cwm" mdupont?

12:28:33 <DanC_jam> .google cwm timbl

12:28:34 <datum> cwm timbl: http://infomesh.net/2001/cwm/

12:28:39 <DanC_jam> odd

12:28:43 <DanC_jam> .google cwm rdf

12:28:44 <datum> cwm rdf: http://www.w3.org/2002/Talks/0508-swad/slide5-0.html

12:28:58 <DanC_jam> well, those are pretty close

12:30:15 <DanC_jam> cwm has http GET support (log:contents). but nothing beyond that. I dunno how you'd do POST in a reasoner.

12:31:41 <mdupont> DanC_jam: i would post to a filter operation that takes the source as an uri, the filter as an uri

12:31:49 <mdupont> if not specified, the result is returned

12:32:04 <mdupont> otherwise it would be posted to the specified uri

12:32:15 <DanC_jam> why POST?

12:32:20 <mdupont> errr

12:32:24 <mdupont> thats right

12:32:36 <mdupont> well, you might want to post the results to another server

12:32:43 <mdupont> but i guess you could write the file

12:32:56 <DanC_jam> you better not be POSTing questions. cf. whenToUseGet7

12:32:57 <mdupont> and get the other server passing it the file you wrote

12:33:03 <mdupont> so your right

12:33:07 <mdupont> we can ignore post

12:33:10 <mdupont> :)

12:33:14 <DanC_jam> good.

12:33:23 * mdupont sees a beam of light though the clouds

12:33:38 <Ankh> Ankh is now known as AnkhGone

12:59:17 <dajobe> it's timbl http://news.bbc.co.uk/1/hi/technology/3306767.stm

13:14:49 <D[a]vey> I wonder what .stm *is*

13:16:13 <mattb> apache server-side includes

13:17:03 <D[a]vey> no, thats .shtm by default.

13:17:27 <mattb> yes, you ae correct

13:17:34 <mattb> and the bbc have their server configured to run .stm as ssi too

13:17:35 <JimJibber> but it is also stm.

13:18:00 <D[a]vey> heh, I find it hard to believe they do all their site with SSI, honestly.

13:18:04 <mattb> believe me

13:18:06 <mattb> i know

13:18:11 * mattb sighs

13:18:22 <D[a]vey> Oh?

13:18:36 * JimJibber finds it amazing that bbc users have never come across broken SSI includes, I do all the time...

13:18:52 <D[a]vey> I never have. heh

13:18:53 <mattb> d[a]vey: /whois me: ~mattb@gateb.thls.bbc.co.uk

13:18:57 <mattb> :)

13:19:04 <D[a]vey> aaah... I see :D

13:19:06 <eaon|zZz> eaon|zZz is now known as eaon

13:19:20 <D[a]vey> mattb: bet that looks good on the C.V. ;)

13:19:48 <D[a]vey> D[a]vey is now known as Davey

13:20:15 <deltab> maybe the SSI executes code that dies when you access it

13:20:53 <JimJibber> Ah, you think the BBC use their advanced SemWeb knowledge to realise it's me accessing and censor the news?

13:22:56 <Davey> mattb: eew.

13:23:12 * Davey points BBC to PHP

13:23:27 <dajobe> hahaha

13:23:43 <mattb> what dajobe said

13:23:49 <Davey> pfft.

13:23:59 <Davey> what do you think would be better? Even Yahoo! uses PHP :D

13:24:06 * JimJibber points it at an all javascript backend!

13:24:11 <deltab> Google knows of about 7620 pages affected

13:24:12 <dajobe> it works darn well the way it is, imho

13:24:13 <mattb> you would not believe the internal political shift the bbc would need to move to anything different whatsoever

13:24:15 <mattb> or perhaps you would

13:24:31 <mattb> it's mostly static html anywya

13:24:42 <mattb> the ssis are just used to pull in standard nav furniture and so on

13:24:43 <deltab> out of 515000 pages, that's not bad

13:24:50 <dajobe> deltab: exactly

13:24:55 <dajobe> gotta run

13:24:56 <Davey> mattb: static... HTML. Seperate files?!

13:25:16 * Davey would hate to see the BBC try to re-design

13:26:35 <JimJibber> You can see the many redesignd of the BBC news site by looking at their PDA/digital TV etc. versions, they're all available.

13:28:27 <mattb> yep, just cos static files are used to serve doesn't mean they're the source of the content

13:28:32 <mattb> the site's baked out of a cms

13:28:43 <mattb> static files are the most efficient way to serve at request-time

13:29:10 <JimJibber> Gas Mark 7 for 30 minutes type thing/

13:29:11 <Davey> mattb: I see :)

13:29:39 <Davey> quite interesting to see how such an enourmous site does its thing.

13:29:47 <Davey> mattb: what language is that CMS written in?

13:30:55 <mattb> java, i believe, but i work in a different part of the bbc so i've not had direct access to it

13:31:24 * JimJibber recommends php as the first step away...

13:31:39 <Davey> java? wow.

13:31:53 <Davey> I guess its GUI clients and such.

13:31:59 * mattb runs away from language/systems advocacy

13:32:31 <Davey> I'm not going to get into that, as dajobe said, it works well :)

13:32:51 <Davey> it would be interesting to see how PHP would bear up in such a situation though.

14:17:55 <kendallclark> BLURB:Adam Bosworth beats the "XML is self-describing" horse hard

14:17:56 <dc_rdfig> B: Adam Bosworth beats the "XML is self-describing" horse hard from kendallclark

14:18:10 <kendallclark> B:(in his XML Conf, 2003, keynote)

14:18:11 <dc_rdfig> Added comment B1.

14:23:05 <kendallclark> B:"Programmers don't think about things like SAX and DOM, but rather things like 'purchase order'." -- Huh?

14:23:05 <dc_rdfig> Added comment B2.

14:23:22 <kendallclark> B:That's rather a bizarre claim, I think.

14:23:22 <dc_rdfig> Added comment B3.

14:25:13 <kendallclark> B:"There's not a single standard 'out there' which says 'call that service before this one'."

14:25:13 <dc_rdfig> Added comment B4.

14:25:43 <sandro> RDFCore folks? I'm confused about which drafts are the intended ones. Is it names like http://www.w3.org/2001/sw/RDFCore/TR/WD-rdf-primer-20030117/Overview.html or is it the ones in ./staging/ or something?

14:26:23 <DanC_jam> the former, sandro; see last week's minutes for very explicit details

14:27:01 <DanC_jam> i.e. http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Dec/0015.html

14:27:21 <kendallclark> wow, all things being equal (even if not), i'd rather be at the world info summit with TBL!

14:27:58 * DanC_jam is booting up kinda slowly again today...

14:28:23 * kendallclark finds it hard to stay in bed late in philly

14:28:56 <DanC_jam> well, I'm up and I went to breakfast with EricP, but now I'm back in my hotel room

14:29:59 <kendallclark> good choice

14:30:11 <kendallclark> though it's quite full today

14:33:46 <kendallclark> B:He's seriously slagging really complex web services; sounds like a REST evangelist, though I'm not sure he means to sound like one.

14:33:46 <dc_rdfig> Added comment B5.

14:34:46 <kendallclark> B:This is actually quite interesting, since BEA is as responsible as any other institution in complexifying web services, particularly the whole EJB thing, which he just slagged hard.

14:34:46 <dc_rdfig> Added comment B6.

14:43:54 <kendallclark> B:"My service-provider customers want a 'soap cookie', not fancy coordination/transaction stuff."

14:43:54 <dc_rdfig> Added comment B7.

14:47:39 <kendallclark> B:Pushing hard on "synchronization"; suggests SyncML needs updating.

14:47:40 <dc_rdfig> Added comment B8.

14:49:33 * DanC_jam notes WSDL 2.0 @ 2pm in my gizmo calendar. http://www.xmlconference.org/xmlusa/2003/wednesday.asp#13

14:50:26 <kendallclark> B:argh, he pimps weblogs ("there are these things called blog readers") hard, which is bizarre.

14:50:27 <dc_rdfig> Added comment B9.

14:52:09 <DanC_jam> hmm... SyncML... I wonder if there's a real chance to update that.

14:52:38 <kendallclark> he suggested current version is too, well, solipsistic.

14:52:41 <kendallclark> my word

14:52:51 <DanC_jam> wtf is up with internet service from the marriott? I'm getting IIS error messages; they're clearly intercepting port 80 and running it thru a proxy.

14:52:56 <kendallclark> too much focused on *my* data, in other words

14:52:58 <edd> he's not wrong, SyncML is a complete interop disaster

14:53:03 <edd> DanC_jam: yeah. it's a pain

14:53:32 <kendallclark> well, my question is more to his point about synchronization being the underlying goal of a new kind of data model

14:53:36 <kendallclark> the much larger, stronger claim

14:53:41 * DanC_jam is glad to hear SyncML isn't going to do to RDF sync what <font> did to stylesheets

14:55:25 <ericP> danc, set up a proxy channel: `ssh -L3128:squid.lcs.mit.edu:3128 homer.w3.org` and tell your browser to use proxy localhost:3128

14:56:35 * DanC_jam calls hotel operator to make sure calling the 800 number for net support won't incur the $1 local call charge

14:57:23 <DanC_jam> I'm not paying $10/day for crap service, ericp

14:58:22 * DanC_jam listens to DHCP clues by voice; fairly well done

14:59:09 * DanC_jam waits on hold for STSN technician. tel:+1-800-571-9462

14:59:27 <DanC_jam> jake

14:59:56 <kendallclark> 'net in my hotel is $10 a day but rather good service

15:01:03 <ericP> danc, paying $10/day for crap service, neither am i

15:01:38 * DanC_jam asks for the port 80 shenanigans to be turned off; "what? huh? proxy?"

15:02:29 <ericP> i asked if the room came with internet access. they said "yes". i arrived, discovered it didn't, and that they policed stuff that looked like HTTP (not just on 80) and opted for a tunnel

15:03:40 <edd> so i had an idea for a new tech called RSS. really simple synchronisation

15:03:45 <DanC_jam> Jake claims they don't use proxies. I'm pretty certain he's wrong. I see "ISA Server: S-UTL01-NYNOC.stsn.com Via: Time: 12/10/2003 2:52:57 PM GMT" in this message.

15:04:51 * DanC_jam waits on hold while Jake asks around

15:05:50 <edd> there's no way they don't have proxies, i've seen IIS pages and Redhat error pages where i should have had 404s

15:06:32 <DanC_jam> Jake apologizes for misinforming me, tweaks something; I'm winning now.

15:07:03 <edd> Heh, that's been irritating me for days. Guess I should have phoned.

15:07:08 <edd> Did they say what the trouble was?

15:07:20 <DanC_jam> he said something about turning off "proxy detect"

15:07:34 <MarkBaker> MarkBaker is now known as MarkBa

15:07:56 <DanC_jam> I occasionally gripe to timbl about these port 80 shenanigans, and he encourages me to press the matter. That's (one of the reasons) why I bothered to call tech support.

15:08:28 <DanC_jam> Jake said that no, they can't turn the proxy off, but he turned "proxy detect" off and my problems seem to have subsided.

15:08:52 <DanC_jam> phpht. I still can't get to http://www.advogato.org/

15:09:13 <danbri>http://web.archive.org/web/20000816002508/http://www.biztalk.org/Resources/canonical.asp

15:09:13 <dc_rdfig> C: http://web.archive.org/web/20000816002508/http://www.biztalk.org/Resources/canonical.asp from danbri

15:09:33 <danbri> C:|Serializing Graphs of Data in XML, Bosworth, Layman, Rys (2000)

15:09:33 <dc_rdfig> Titled item C.

15:09:49 <danbri> C:This dropped off the Web; I just dug it out via archive.org.

15:09:49 <dc_rdfig> Added comment C1.

15:10:46 <DanC_jam> A:has anybody asked that the rdfig blog be added?

15:10:46 <dc_rdfig> Added comment A2.

15:10:54 <danbri> C:Adam Bosworth presented similar position at QL'98 in Boston; see associated [http://www.w3.org/TandS/QL/QL98/pp/microsoft-serializing.html|paper] by Andrew Layman (then a member of the RDFS WG btw).

15:10:54 <dc_rdfig> Added comment C2.

15:14:33 <DanC_jam> edd? I get 404 @ http://usefulinc.com/edd/blog

15:14:35 <danbri> C:See [http://www.w3.org/TR/1998/WD-rdf-schema-19980814/|Aug 1998 RDF Schema Working Draft].

15:14:36 <dc_rdfig> Added comment C3.

15:14:52 <kendallclark> wow, this hp talk is killing me

15:15:09 <DanC_jam> is that a good thing or a bad thing, kendallclark?

15:15:15 <edd> DanC_jam: whoa. that proxy is killing things for you.

15:15:22 <kendallclark> a very bad thing

15:15:27 <DanC_jam> the 404 isn't from a proxy, ed

15:15:37 <DanC_jam> phpht. perhaps it is

15:15:39 <kendallclark> i mean, i'm trying to write a deviant column, and this keynoter won't shut up! :>

15:15:48 <edd> DanC_jam: i promise you it isn't from me, i run Zope on that site and I bet you're not seeing a zope error.

15:16:30 <DanC_jam> there error is from Apache; threw me for a minute, until I got the same error from another blog

15:16:43 * DanC_jam grumbles, considers calling STSN again

15:17:55 * DanC_jam calls STSN tech support again tel:+1-800-571-9462

15:18:09 * edd using the free wireless in the marriott lounge lobby, for some reason that seems to have fewer proxy problems. still useless DNS tho.

15:18:35 * DanC_jam wants to have lots of good justification for refusing to pay the marriott for net service

15:19:30 <kendallclark> hehe

15:19:57 <kendallclark> so, i'm wondering if danc ever drops the "i work for TBL" bomb on poor hotel tech support types?

15:20:32 * danbri wonders how much (CDROM/DVD?) space a cache of the entire DNS would need

15:20:58 <danbri> (hmm that's probably not a wellformed question, let alone a feasible datasource to acquire, ...)

15:21:27 * DanC_jam runs a local dns cache on my laptop; endorses djbdns

15:21:33 * DanC_jam gets jake again

15:22:02 <kendallclark> there's a local caching nameserver howto for linux; it's quite nice, as i recall

15:22:03 <ericP> feasible datasource, mason's friend mackler used to take yearly snapshots of the root name servers.

15:23:03 <danbri> interesting. i'm pretty ignorant of dns protocols...

15:25:34 * DanC_jam informs jake@STSN that if they can't turn the port 80 intercept off, I'm gonna tell marriott that they don't provide the service they advertise, so I don't feel obliged to pay

15:27:14 <kendallclark> oh my.

15:27:15 <ericP> danc, if you make them reconfigure and reboot, i'll lose a lot of state. ironic.

15:27:36 <kendallclark> tehy've just awarded the 'xml cup' and when lauren announced the names, this amazingly cheesy music started playhing, on cue.

15:27:42 * kendallclark shudders

15:28:01 <edd> oh boy

15:28:03 <edd> so who won?

15:28:11 <edd> that's prolly report-worthy, kendallclark :)

15:28:27 <danbri> adam bosworth and michael sperberg-mcqueen

15:29:01 <edd> can't wait for the RDF cup

15:29:20 <kendallclark> "the memories are a large part of what I remembere" -- well, there ya go, Mister Tautology!

15:29:41 * DanC_jam reaches front desk @ marriott; STSN hasn't charged me.

15:30:20 * DanC_jam regrets not being there to see MSM get his due

15:30:34 * DanC_jam packs up

15:31:44 <kendallclark> wow, bosworth's accceptance is 10x longer than MSM's.

15:42:23 <ericP> libby, is shellac around?

15:44:53 <libby> he was...I'm at HP Labs, he's at home. may have nipped out

15:44:53 <libby> hello danja

15:45:20 <danja> hi libby!

15:45:37 <ericP> a while ago i was pestering yhou for a TreeHugger example query.

15:45:54 * danja now trying kvirc, apologies for any junk

15:46:05 <ericP> he gave me one in xsl. i was wicked busy and forgot to ask him for the TreeHugger version.

15:46:36 <libby> um, that sorta is the treehugger version. do you mean the xquery version?

15:47:01 <ericP> yeah, i guess i'm confused by what the TreeHugger name applies to

15:47:09 <ericP> i thought it was the XQuery mapping

15:47:25 <libby> it started off as xslt

15:47:29 <libby> then xquery

15:47:36 <ericP> ahh

15:47:47 <ericP> so yeah, it was the XQuery i wanted

15:47:49 <libby> really most of the content is in xpath

15:47:49 <libby> - common to both

15:48:00 <libby> I'll bug him when I see him in a bit

15:48:23 <ericP> i got the impression that TreeHugger was using a less brittle interpretation of the xpath

15:48:42 <libby> they're not really xpaths, more like rdfpaths

15:48:44 <ericP> ie, resilient against different serializations of the same data

15:48:45 <ericP> ok

15:48:52 <libby> yeah

15:49:05 * shellac here now

15:49:37 <shellac> I can try the xquery - it will be the same paths, of course

15:49:46 <ericP> that would be great - tx!

15:50:13 <ericP> i'm putting it on a handout for an XQuery RDF Query BOF at XML2003

15:50:25 <shellac> ah - ok

15:51:29 * shellac rattles brain to stimulate what little XQuery knowledge I possess

15:52:05 <ericP> try pop rocks

15:52:28 <shellac> and coca-cola

16:04:57 <ericP> brb

16:06:04 <evlist>http://www.xmlconference.org/xmlusa/2003/wednesday.asp#2

16:06:04 <dc_rdfig> D: http://www.xmlconference.org/xmlusa/2003/wednesday.asp#2 from evlist

16:06:48 <evlist> D:|NLM's Public Domain DTDs: A 9-Month Update - Deborah A Lapeyre & Jeff Beck

16:06:48 <dc_rdfig> Titled item D.

16:07:50 <evlist> D:The plan is to create an active community beyond these DTDs.

16:07:50 <dc_rdfig> Added comment D1.

16:09:52 <evlist> D:The goal is to preserve the intellectual content of journal articles.

16:09:52 <dc_rdfig> Added comment D2.

16:10:47 <evlist> D:It's a modular DTD suite that can be used to define custom DTDs.

16:10:48 <dc_rdfig> Added comment D3.

16:12:39 <evlist> D:Borrows from existing DTDs (MathML, DocBook, ...) when possible.

16:12:39 <dc_rdfig> Added comment D4.

16:12:55 <danja> which reminds me of something I wanted to ask here...

16:13:51 <danja> has anyone had a crack at getting an RDF API to output RDF/XML that matches a DTD?

16:14:41 <danja> e.g. RSS 1.0 syntax (I think) is constrained enough for it to have a DTD

16:15:02 <danja> so it would be handy to take some arbitrary RDF

16:15:26 * danbri clutches new purchases

16:15:32 <danbri> .google "XQuery from the experts"

16:15:35 <datum> "XQuery from the experts": http://www.awprofessional.com/catalog/product.asp?product_id=%7BC4E463B7-6CD6-4EA1-A920-3617A2B453A1%7D

16:15:44 <MarkB>http://newsimg.bbc.co.uk/media/images/39619000/jpg/_39619051_annan_bernerslee203.jpg

16:15:44 <dc_rdfig> E: http://newsimg.bbc.co.uk/media/images/39619000/jpg/_39619051_annan_bernerslee203.jpg from MarkB

16:15:52 <MarkB> E:|Good co-depiction fodder

16:15:52 <dc_rdfig> Titled item E.

16:15:56 <danja> and have the appropriate bits spew out as DTD-valid XML

16:15:58 <danbri> .google xforms essentials dubinko

16:15:59 <kendallclark> heh, i got that book for free, for some reason, in the mail.

16:15:59 <datum> xforms essentials dubinko: http://dubinko.info/writing/xforms/

16:16:23 <MarkB> E:found [http://news.bbc.co.uk/2/hi/technology/3306767.stm|here]

16:16:23 <dc_rdfig> Added comment E1.

16:17:10 <edd> danbri: aw, i had a spare copy of that xquery book

16:17:20 <edd> danbri: also i have that cable for you... where are you now?

16:17:55 <evlist> D:The article DTD has been written as a superset of existing DTDs for articles.

16:17:55 <dc_rdfig> Added comment D5.

16:19:14 <evlist> D:An extensive documentation is available.

16:19:15 <dc_rdfig> Added comment D6.

16:20:13 <kendallclark> BLURB: Mike Champion's WSA Presentation

16:20:13 <dc_rdfig> F: Mike Champion's WSA Presentation from kendallclark

16:20:30 <kendallclark> F:Which I'm attending, but not churning -- it's very good.

16:20:30 <dc_rdfig> Added comment F1.

16:23:36 <evlist> D:DTDs have been used because W3C XML Schema do cause more interoperability issues and DTDs have been considered easier to modularize.

16:23:36 <dc_rdfig> Added comment D7.

16:24:57 <kendallclark> F:Mike's said good things about some of the semweb effort w/r/t the WSA work. OWL, for example, has "proven to be quite useful" in the WSA effort. I think this is stuff Bijan Parsia has been doing. At least, Evren Sirin of UMD has been working on it too.

16:24:57 <dc_rdfig> Added comment F2.

16:25:19 <kendallclark> F:"I've come away with a lot more respect for OWL...than I did a couple of years ago", Mike says.

16:25:19 <dc_rdfig> Added comment F3.

16:25:22 <kendallclark> F:Yay!

16:25:22 <dc_rdfig> Added comment F4.

16:25:28 <MarkB> F:[http://lists.w3.org/Archives/Public/www-ws-arch/2003Dec/0008.html|some notes] of his posted to www-ws-arch

16:25:29 <dc_rdfig> Added comment F5.

16:26:04 <evlist> D:But other schema languages are being considered (RELAX NG).

16:26:05 <dc_rdfig> Added comment D8.

16:26:15 <kendallclark> F:Ah, thanks, MarkB.

16:26:15 <dc_rdfig> Added comment F6.

16:26:28 <MarkB> no prob

16:26:47 <evlist> D:The adoption by the community is better than they had anticipated.

16:26:47 <dc_rdfig> Added comment D9.

16:27:05 <shellac> ericp: ok, this seems to work http://rdfweb.org/people/damian/2003/12/query_example.xq

16:27:25 <shellac> ericp: output http://rdfweb.org/people/damian/2003/12/query_example.out.xml

16:28:21 * MarkB was under the impression Mike liked OWL as a design-time tool rather than run-time. it's a start though.

16:28:31 <ericP> thanks a bunch, shellac

16:29:05 <danbri> +1 re design time liking.

16:29:10 <danbri> I think that's what he's saying

16:29:21 * danbri agrees with it w.r.t. FOAF experience

16:30:11 <shellac> xquery syntax has changed on me before, tho', so it might not be quite right

16:30:19 <MarkB> and of course, one person's design time is another's run time 8-)

16:34:34 <kendallclark> MarkB: yes, I hope I didn't suggest otherwise.

16:34:41 <kendallclark> i.e., that's what he's just said in his presentation

16:36:22 <MarkB> no no, you didn't suggest otherwise. i was just adding some context.

16:36:29 <evlist> D:The DTDs are not really open source (they want to retain control on the updates) but public domain (can grab them and use them as you like).

16:36:30 <dc_rdfig> Added comment D10.

16:37:53 * DanC_jam packs up...

16:39:29 <danbri> danc, where are you, where going?

16:39:40 <danbri> (re rdfcore pub discussion...)

16:39:42 <evlist> D:More info at http://dtd.nlm.nih.gov/

16:39:42 <dc_rdfig> Added comment D11.

16:45:59 <evlist>http://www.xmlconference.org/xmlusa/2003/wednesday.asp#11

16:45:59 <dc_rdfig> G: http://www.xmlconference.org/xmlusa/2003/wednesday.asp#11 from evlist

16:46:28 <evlist> G:|Incremental XML Parsing and Validation in a Text Editor - James Clark

16:46:28 <dc_rdfig> Titled item G.

16:48:18 <evlist> G:Requirements: free XML editor that supports RELAX NG that he would use.

16:48:19 <dc_rdfig> Added comment G1.

16:49:21 <evlist> G:Result: nXML mode for Emacs all written in Emacs Lisp (~15 000 lines).

16:49:22 <dc_rdfig> Added comment G2.

16:50:26 <evlist> G:If you don't use RELAX NG, you can use trang or msv to convert your schemas into RELAX NG.

16:50:26 <dc_rdfig> Added comment G3.

16:53:14 <evlist> G:Emacs Lisp is slow and better solutions than brute force had to be considered.

16:53:14 <dc_rdfig> Added comment G4.

16:55:49 <evlist> G:nXML is a text based XML editor as opposed to editors editing XML structures.

16:55:49 <dc_rdfig> Added comment G5.

16:57:12 <evlist> G:nXML doesn't require a schema, but when a schema is provided it detects invalidity contnuously and provides completion.

16:57:12 <dc_rdfig> Added comment G6.

16:57:29 <evlist> G6:nXML doesn't require a schema, but when a schema is provided it detects invalidity continuously and provides completion.

16:57:30 <dc_rdfig> Replaced comment G6.

17:00:56 <edd> G:Complete continuous revalidation impractical, so must find a way of incremental validation.

17:00:57 <dc_rdfig> Added comment G7.

17:01:17 <edd> argh, missed that slide.

17:02:22 <edd> G:Change within validated region moves end of validated region to behinning of change, region bdaries recomputed immediately after each change, validation is postponed uintil either editor is idle or editing operations require validation state.

17:02:22 <dc_rdfig> Added comment G8.

17:02:27 <evlist> G:Validation is postponed until either the editor is idle or the editing operation require to validate.

17:02:27 <dc_rdfig> Added comment G9.

17:02:32 <edd> oops

17:02:36 * edd hands scribe to evlist, sorry..

17:02:46 <evlist> no problem

17:03:18 <evlist> G9:""

17:03:18 <dc_rdfig> Deleted comment G9.

17:03:44 <evlist> edd, go ahead, you're typing faster than I do!

17:03:59 <DanC-AIM> I'm having wireless trouble from rm109 where james is talking. Can ping the gateway, 10.11.0.1, but not 18.29.0.200

17:04:13 <edd> DanC-AIM: you need to use your web bnrowser to re-enter the code

17:04:27 <edd> evlist: can you do this slide, i just missed it

17:05:02 <evlist> I've missed it too!

17:05:33 <DanC-AIM> Ah! Good idea, edd

17:05:43 <edd> G:Choosing the validation algo: must be efficient both to store the validation state, and compare it with a stored state.

17:05:44 <dc_rdfig> Added comment G9.

17:06:04 <edd> G:Derivative based algos satisfy this requirement and are commonly used for RELAX NG validation

17:06:04 <dc_rdfig> Added comment G10.

17:07:17 * DanC_jam thanks edd for tech support clue

17:07:31 <edd> G:XML parsing -- need to be able to start parsing from a position that has a stored validation state. Full XML parsing can only be done sequentyially. Solution is a different devision of responsibility between parser and validator.

17:07:31 <dc_rdfig> Added comment G11.

17:07:36 <edd> DanC_jam: np

17:07:48 <DanC-AIM> byebye

17:08:00 * DanC_jam chalks another one up for the sidekick and aimbridge.py

17:08:07 <edd> G:Parser responsibilities -- parse prolog as usual, divide instance into tokens, check all XML constraints that are internal to a single token, provide information about each token.

17:08:08 <dc_rdfig> Added comment G12.

17:08:28 * edd is scribing the slides literally, folk with illuminating comments please chip in, i likely can't scribe both

17:08:47 <DanC_jam> the need to parse backwards was argued in the SGML/XML WG, but folks weren't convinced. This is james's "I told you so" to the world 1/2 ;-)

17:09:21 <edd> G:Validator responsibilities -- check that tags balance, check that namespace prefixes are declared, check for duplicate attributes, check that the instance consists of a single elements

17:09:22 <dc_rdfig> Added comment G13.

17:09:35 <edd> G:Note some features traditionally associated with the parser have been moved to the validator.

17:09:36 <dc_rdfig> Added comment G14.

17:09:41 * DanC_jam reviews decision record (http://www.w3.org/XML/9712-reports.html) to see if this is recorded

17:10:38 <DanC_jam> I wonder if the things that nxml-mode doesn't support correspond to those things clark voted against in 9712-reports

17:10:39 <edd> G:Implications of these new responsibilities: we need a richer validation state to include info about open elements and inscope namespaces. This means the parser can start at any position in the instance, provided it is between tokens

17:10:40 <dc_rdfig> Added comment G15.

17:10:57 * DanC_jam wonders whether our chump notes are covered by the paper

17:11:04 <edd> G:Instance parsing is stateless, store validation state only between tokens, any change to internal subset requires complete reparsing.

17:11:04 <dc_rdfig> Added comment G16.

17:11:28 <edd> if you can find a URI for the paper, i'll stop :)

17:11:52 <DanC_jam> hmm... "Today, after discussion, we agreed that support for external text entities would be an optional feature of XML 1.0 (dissenting: Clark, Paoli, Sharpe)."

17:13:22 <DanC_jam> hmm... I don't see a paper after all

17:14:18 * AnkhGone notes xml core already produced a spec for describing open elements, inscope namespaces, open physical entties

17:14:24 <AnkhGone> 3 or 4 years ago

17:19:32 <evlist> G:As a summary, James mentions that this would not have work for SGML since he is taking advantage of the simplications of XML.

17:19:33 <dc_rdfig> Added comment G17.

17:19:47 <evlist> G:He insists that he never builds a tree.

17:19:48 <dc_rdfig> Added comment G18.

17:20:01 * DanC_jam hunts for the list of XML 1.0 doodads that nxml-mode doesn't grok

17:20:45 <evlist> G:The parsing is done twice (minimal without validation for syntax highlight) and full with validation.

17:20:46 <dc_rdfig> Added comment G19.

17:20:57 * DanC_jam can't find the nxml-mode tutorial in /usr/share/doc/nxml-mode on this debian box. odd!

17:21:13 <DanC_jam> G:Clark demos nxml-mode on an OWL/RDF/XML document

17:21:14 <dc_rdfig> Added comment G20.

17:21:42 <DanC_jam> G:quite relevant, since the XML Schema WG is on record objecting to RDF's syntax. Perhaps the bug is in XML Schema, not RDF.

17:21:43 <dc_rdfig> Added comment G21.

17:22:41 <kendallclark> heh, jclark loads the nci ontology in nxml ;>

17:22:42 <DanC_jam> G:Clark loads NCI's cancer ontology, while praying to the demo gods...

17:22:42 <dc_rdfig> Added comment G22.

17:23:24 <DanC_jam> ooh!

17:23:27 <sandro> the oncology ontology!

17:23:38 <kendallclark> and hendler and jen's names popped on the screen; yay!

17:23:48 <kendallclark> yep, the on(c,t)ology

17:23:50 <DanC_jam> G:syntax-highlighting the end of the document works even though only 22% is validated so far

17:23:50 <dc_rdfig> Added comment G23.

17:23:54 <danbri> mindswap++

17:24:13 * DanC_jam calls for group hug

17:24:23 <danbri> .google grouphug

17:24:25 <datum> grouphug: http://grouphug.us/

17:24:30 <mattb> hehe

17:24:40 * kendallclark gags, a bit :>

17:24:48 <danbri> careful what you wish for :)

17:25:27 <kendallclark> you might make someone gag!

17:26:13 * DanC_jam is sooo glad clark can't be bought and that the world gets his stuff for free

17:26:45 <kendallclark> it's like bonnie prince charles on XML/Lisp crack, with a layer of James Mason (the actor!) Cool

17:28:48 <kendallclark> funny; jclark should take up haskell.

17:29:19 <DanC_jam> . Q: do you have a formal proof that the incremental parsing algorithms work?

17:29:32 <kendallclark> i didn't think that such an insane question, actually.

17:29:36 * danbri -> lunch, then 'know what yr schemas mean' session (rm 105, xml2003)

17:29:42 <DanC_jam> . A: umm... at first I developed it and debugged ad-hoc; at this point, I have no formal proof, but I'm quite confident that it's right.

17:30:05 <evlist> kendallclark: he does describe his algorithms in haskell.

17:30:17 <DanC_jam> logger, chump G

17:30:17 <DanC_jam> G:See [http://ilrt.org/discovery/chatlogs/rdfig/2003-12-10#T17-30-17|discussion]

17:30:17 <dc_rdfig> Added comment G24.

17:30:24 <kendallclark> evlist: he does? where?

17:31:12 <evlist> the algorithms of RELAX NG and NRL are documented in haskell.

17:31:23 <kendallclark> ah, that's very cool. didn't know that.

17:31:25 <kendallclark> thx

17:31:34 <DanC_jam> reasons-to-study-haskell++

17:31:56 <kendallclark> (my point was just that haskellers often prove algorithmic and programmatic correctness, and it's a very practical language.)

17:32:26 <DanC_jam> where "very" is clearly a matter of taste

17:32:47 <kendallclark> you think?

17:33:00 <kendallclark> i rather doubt it's that subjective, danc.

17:33:03 <kendallclark> but -shrug-

17:34:19 <DanC_jam> for a language to qualify as practical, it should be easy to learn; haskell has a nice design, but learning haskell involves learning some very novel concepts with respect to pascal/C and even java/python

17:34:22 <kendallclark> lunch

17:34:27 <evlist> see http://www.thaiopensource.com/relaxng/implement.html and

17:35:25 <DanC_jam> and for a language to qualify as practical, there should be APIs for every database system, windowing system, network protocol, etc. Haskell has a ways to go there, AFAIK. And doing C extensions to haskell has got to be tricky.

17:37:15 <Arnia> DanC_jam: If you'd started from a functional language, you'd find imperative languages hard to learn. Don't confuse familiarity with true ease of understanding

17:37:53 <DanC_jam> what's "truth"? popularity correlates highly with practicality. It's easier to learn something when lots of people around know it.

17:37:54 <Arnia> DanC_jam: Are OO languages impractical because they require you to learn about objects?

17:38:22 <DanC_jam> OO langauges were impractical for that reason at one time, yes. Since then the knowledge has become nearly ubiquitous, so they're practical now.

17:38:43 * Arnia never had any problems with learning haskell nor with explaining it

17:39:03 <Arnia> I could explain functional code to a non-coder friend far quicker than the equivalent imperative stuff

17:39:19 <Arnia> It looked closer to the concepts for him

17:40:05 <DanC_jam> I think we agree more than we disagree; I appreciate functional languages. But claiming that haskell is practical, and that this claim is objectively true, seems like over-selling, to me.

17:40:25 * DanC_jam packs up for lunch

17:40:35 * Arnia goes to get his black tie on

17:40:53 <Arnia> Yeah... but I want to make haskell more 'practical'

18:18:18 <Davey> Davey is now known as D[a]vey

18:29:16 <D[a]vey> D[a]vey is now known as Davey

18:36:27 <iwaiAway> iwaiAway is now known as iwaim

18:36:33 <edd> hey grant.

18:37:16 <kendallclark> hey bill

18:38:47 <kendallclark> w'less flaky where i am

18:55:55 <Davey> Davey is now known as D[a]vey

19:02:43 <kurro> hi

19:09:53 <kurro> :)

19:10:05 <kurro> anyone there?

19:12:19 <AnkhGone> there, or here? :)

19:13:16 <kurro> anywhere man

19:13:54 <LotR> nope

19:15:25 <kendallclark>http://www.xml.com/pub/a/2003/12/10/deviant.html

19:15:26 <dc_rdfig> H: http://www.xml.com/pub/a/2003/12/10/deviant.html from kendallclark

19:15:29 <kurro> ok! I have a question about owl. is this place the correct place to place it or there are other places?

19:15:38 <kendallclark> H:|The Tag's Town Hall

19:15:38 <dc_rdfig> Titled item H.

19:15:59 <kendallclark> H:An Onsite XML Deviant (not really pubbed yet, but it's there for a look for the curious)

19:15:59 <dc_rdfig> Added comment H1.

19:16:00 <LotR> kurro: those aren't mutually exclusive

19:17:15 <kurro> LotR thanks

19:17:46 <MarkB> H:webarch has gone [http://lists.w3.org/Archives/Public/www-tag/2003Dec/0168.html|last call]

19:17:46 <dc_rdfig> Added comment H2.

19:18:09 <kendallclark> MarkB, the Grand Supplementer!

19:18:57 * DanC_jam looks around for bwm..

19:21:02 <kendallclark> hi edd

19:21:45 <edd> hi k.

19:24:08 <kurro> LotR may you help me with a couple concepts on OWL? (this is my first time using IRC)

19:24:11 <MarkB> i aim to please

19:26:31 <kurro> how can i define a class for a finite state automata whil all the constraints?

19:29:11 <kurro> how can i define a class with five sets with some set bein a subset of another?

19:30:12 <kurro> is it possible at all using OWL?

19:34:21 <MDubinko> DSDL is incredibly expansive

19:35:18 <kendallclark> seems v. ambitious

19:42:01 <libby> E:[http://swordfish.rdfweb.org/photos/genfiles/ilrt/5137388545953361.rdf|codepicted]

19:42:01 <dc_rdfig> Added comment E2.

19:42:08 <libby> !pic kofi

19:42:11 <whwhwhwh> Kofi Annan http://newsimg.bbc.co.uk/media/images/39619000/jpg/_39619051_annan_bernerslee203.jpg

19:44:32 <kendallclark> dajobe: hi

19:44:45 <dajobe> hi

19:45:03 * MDubinko doesn't hang out enough with RDF folks :-)

19:45:46 <kendallclark> they're a weird lot!

19:46:27 * kendallclark finding listening to sinatra more rewarding than a post-lunch session

19:50:01 <evlist>http://www.xmlconference.org/xmlusa/2003/wednesday.asp#21

19:50:02 <dc_rdfig> I: http://www.xmlconference.org/xmlusa/2003/wednesday.asp#21 from evlist

19:50:32 <MDubinko> nice job, eric

19:50:39 <evlist> I:|Combining Multiple Vocabularies Without Tears - Makoto Murata

19:50:39 <dc_rdfig> Titled item I.

19:55:39 <evlist> I:Murata Makoto shows examples of composite (XHTML + MathML, XForms + XHTML2 + user defined, XHTML 2, MathML, SVG, RDF, ...) documents

19:55:39 <dc_rdfig> Added comment I1.

19:58:12 <MDubinko> I:"Who has read all these documents: XHTML, RDF, SVG, MathML, XForms" -- one person raises hand

19:58:12 <dc_rdfig> Added comment I2.

19:59:19 <MDubinko> I:by necessity, mixed vocabulary handling needs to handle partial understanding

19:59:19 <dc_rdfig> Added comment I3.

20:01:57 <evlist> I:The intention of namespaces was to solve the problem.

20:01:57 <dc_rdfig> Added comment I4.

20:04:26 <evlist> I:We need to be able to use the schemas for the different namespaces without understanding them.

20:04:27 <dc_rdfig> Added comment I5.

20:06:46 <evlist> I:These namespaces may use different schema languages.

20:06:46 <dc_rdfig> Added comment I6.

20:07:35 <evlist> I:Special purpose schema languages will continue to appear: on a small scope they are always more convenient than general purpose schema languages.

20:07:36 <dc_rdfig> Added comment I7.

20:10:00 <kendallclark> danbri!

20:10:05 <danbri> hi

20:13:17 <LotR> hi Ankh

20:13:24 <Ankh> hi :)

20:13:51 <evlist> I:This divide & validate paradigm can be implemented as a SAX events dispatcher.

20:13:52 <dc_rdfig> Added comment I8.

20:13:54 <Ankh> assert(freenode::LotR == gnome::LoTR) ?

20:14:27 <danbri> Ankh, do you know any opensource engines that map incoming Xqueries into backends that are SQL-based? (ie. no angle-brackets...)

20:14:51 <Ankh> not offhand, there are some free-as-in-beer ones I think

20:15:04 <dajobe> ah, jclark showed nxml on rdf/xml. nice

20:15:21 <LotR> Ankh: true

20:15:27 <Ankh> :)

20:17:04 <evlist> I:If everything goes well, DSDL Part 4 should be a final ISO standard next year.

20:17:05 <dc_rdfig> Added comment I9.

20:24:25 <evlist> I:After namespaces in XML, this is "the second step for a non-monolithic WWW".

20:24:25 <dc_rdfig> Added comment I10.

20:33:05 <Ankh> danbri, Howard Katz might be an interesting person to talk to about the subject of open source xq engines and jdbc/odbc/sql back ends; Jerome,or Peter Funkhauser, might know more about what's out there, though, beyond the stuff listed on the W3C public XQuery page

20:35:48 <Ankh> also see http://davis.wpi.edu/~dsrg/rainbow/mainwindow.htm for an open source one using SQL databases

20:38:22 <danbri> doh, he was just here.

20:38:26 <danbri> where'd he go?!

20:38:37 <danbri> jerome's here, ericp's geeking with him

20:43:15 <Ankh> oh, good, I'm I'm glad they met

20:44:26 <DanC_jam> recommendations for the 4pm session?

20:45:01 <kendallclark> clark, nrl

20:45:09 <DanC_jam> hmm... Clark... I don't think I've ever heard of NRL. http://www.xmlconference.org/xmlusa/2003/wednesday.asp#27

20:45:48 <DanC_jam> as a TAG dude, I have some obligation to see clark talk about NRL, I guess.

20:46:54 <DanC_jam> oops; seems I missed Murata-san on namespace mixing.

20:47:07 * DanC_jam thanks MDubinko for the notes

20:47:16 <DanC_jam> and evlist

20:47:51 <evlist> DanC_jam: James Clark will speak on the same subject

20:48:26 <DanC_jam> er... that seems like a program-committee bug; are they really talking on the same thing?

20:48:54 <evlist> Murata Makoto gave an high level overview, James Clark will go into the details.

20:49:06 * DanC_jam looks up room#...

20:49:20 <evlist> It's room 111

20:49:22 <DanC_jam> 111

20:49:37 <DanC_jam> any power outlets available? 1/2 ;-)

20:49:44 <DanC_jam> I'm at 78% batt

20:50:05 <evlist> yes, there are some still available

20:50:28 <DanC_jam> ooh... NormW on caching at 4:45

20:50:37 <evlist> yes

20:50:51 <danbri> what room is clark in ? that's 111?

20:50:58 <evlist> yes, 111

20:51:28 * DanC_jam hopes norm will connect XML catalogs with HTTP caching

20:54:55 <evlist>http://www.xmlconference.org/xmlusa/2003/wednesday.asp#27

20:54:55 <dc_rdfig> J: http://www.xmlconference.org/xmlusa/2003/wednesday.asp#27 from evlist

20:55:20 <evlist> J:|Namespace Routing Language (NRL) - James Clark

20:55:20 <dc_rdfig> Titled item J.

20:56:31 <DanC_jam> A:I just sent a request; to appear in [http://lists.w3.org/Archives/Public/www-archive/2003Dec/thread.html|www-archive]

20:56:32 <dc_rdfig> Added comment A3.

21:03:16 <evlist> A:Namespaces have a significant cost. Benefit is to avoid ambiguity, especially valuable when boundaries are unpredictable.

21:03:16 <dc_rdfig> Added comment A4.

21:03:27 <evlist> J:Namespaces have a significant cost. Benefit is to avoid ambiguity, especially valuable when boundaries are unpredictable.

21:03:27 <dc_rdfig> Added comment J1.

21:03:34 <evlist> A4:""

21:03:35 <dc_rdfig> Deleted comment A4.

21:03:57 * DanC_jam settles down in 111, watching J

21:04:02 <kendallclark> J:"I don't want to have to choose between namespaces and validation. I want both."

21:04:02 <dc_rdfig> Added comment J2.

21:04:21 <evlist> J:Independently developed vocabularies have independent schemas.

21:04:21 <dc_rdfig> Added comment J3.

21:09:44 * Ankh glad J is for James :)

21:10:07 <Ankh> also, I always carry a power extension & multiway outlet to these conferences

21:10:18 <kendallclark> J:"I prefer multiple, small, complementary, single paradigm languages to large monolithiic multi-paradigm ones."

21:10:18 <dc_rdfig> Added comment J4.

21:10:39 <kendallclark> J:That's interesting, since I agree, but I feel rather diff when it comes to programming languages (that is, I like multiparadigm proggie langs).

21:10:39 <dc_rdfig> Added comment J5.

21:11:03 <Ankh> I used to do a lot of stuff in awk, other small languages, but after a while it gets too tiring to learn them all

21:11:42 * DanC_jam brought an extension cord this tiem

21:11:47 <Ankh> heh

21:12:52 <kendallclark> heh, more haskell influence on JJC

21:13:32 <kendallclark> F:He calls NRL a "combinator" schema language; more Haskell influence (at least at the naming level, maybe the algorithmic, too)

21:13:32 <dc_rdfig> Added comment F7.

21:15:32 <kendallclark> F:(FWIW, in Haskell, a combinator is a function or definition with no free variables.)

21:15:32 <dc_rdfig> Added comment F8.

21:16:23 <Ankh> combinators are part of church's work on lambda calculus iirc, predate Haskell by many years

21:16:36 <Ankh> and James is certainly familiar with LISP

21:16:45 <evlist> J:Element tree is divided into sections where all elements have the same namespace.

21:16:45 <dc_rdfig> Added comment J6.

21:16:51 * sandro wonders how folks pronounce combinator

21:17:28 <Ankh> bin as in dustbin

21:17:31 <kendallclark> Ankh: of course, i was just riffing about james and haskell in the earlier session, and got a bit stuck on it :>

21:17:36 <Ankh> ah ok

21:17:37 <evlist> J:An element is a section root if its has a different namespace than its parent.

21:17:37 <dc_rdfig> Added comment J7.

21:18:15 <sandro> which syl accented?

21:18:18 <kendallclark> someone asked about whether he'd proved his nxml algorithms, and evlist told me that james had expressed some of his relaxng stuff in haskell

21:18:30 <kendallclark> com-bi-NAY-tore

21:19:03 <kendallclark> (i don't like to miss chances to pimp haskell; it's frighteningly cool, imo)

21:19:19 <evlist> kendallclark: see http://www.thaiopensource.com/relaxng/implement.html

21:19:52 <kendallclark> thx, eric

21:21:25 <evlist> J:"NRL only cares about a section's namespace; it does not care about its internal structure (almost)."

21:21:25 <dc_rdfig> Added comment J8.

21:21:27 <kendallclark> so, james is really smart, obviously.

21:21:33 <Ankh> sandro, first, but I'm a bit English :)

21:21:53 <Ankh> kendallclark, I think that may be an understatement :)

21:22:01 <kendallclark> but he's a particular kind of smart, the kind i like best; he's the kind of smart that makes me feel smarter for being around him. i'm not sure i am smarter, but i feel it, and that's worth a lot.

21:22:19 <evlist> J:A NRL schema is a set of rules.

21:22:19 <dc_rdfig> Added comment J9.

21:22:26 <kendallclark> Ankh: didn't let me finish. :>

21:22:34 <Ankh> heh

21:22:47 <Ankh> yes, he is a bit like that

21:23:02 <kendallclark> he's like a good diss adviser in that way

21:23:06 <evlist> J:Rules can contain multiple actions.

21:23:07 <dc_rdfig> Added comment J10.

21:24:50 <evlist> J:Modes (similar to modes in XSLT) allow context dependent processing.

21:24:50 <dc_rdfig> Added comment J11.

21:27:23 <edd> heh, RDF seems to be JJC's standard pathological use case : )

21:27:38 <kendallclark> heh

21:27:41 <evlist> J:You can combine sections for validation.

21:27:41 <dc_rdfig> Added comment J12.

21:27:46 <kendallclark> that's funny, edd

21:28:01 <evlist> edd: Murata Makoto gave the example of RDF as well.

21:29:23 <evlist> J11:You can combine sections for validation through <attach/> elements.

21:29:23 <dc_rdfig> Replaced comment J11.

21:32:10 * DanC_jam isn't following Clark v. well; maybe my attention is too divided or I'm just tired. but I could use a better "story" or maybe some pictures

21:32:27 <evlist> J:Attributes are attached by default to their parent elements but may also be considered as attribute sections.

21:32:27 <dc_rdfig> Added comment J13.

21:33:41 <evlist> J:NRL supports interleaved namespaces (such as XHTML in XSLT transformations).

21:33:42 <dc_rdfig> Added comment J14.

21:35:27 <kendallclark> he's slightly terrifying, actually

21:35:38 <DanC_jam> well, google will give us the URL today, but it's kinda arrogant to expect that to apply for a long time. I like conference materials to last for a long time.

21:36:03 <kendallclark> arrogant or just dumb?

21:36:05 <evlist> why terrifying?

21:36:28 <DanC_jam> jjc... applies to both Jeremy Carroll and James Clark? huh.

21:36:30 <kendallclark> oh, he just tosses off these super impressive projects like I change clothes. It's rather intimidating.

21:36:45 <evlist> yes, I agree.

21:36:56 <kendallclark> terrifying good, not terrifying bad. :>

21:37:37 * evlist learns a lot to work with jjc, Murata Makoto and Rick Jelliffe (to name few) for DSDL.

21:38:43 <evlist> J:Bigger problem: can NRL help processing multi-namespace documents?

21:38:43 <dc_rdfig> Added comment J15.

21:38:58 <kendallclark> we need to tempt jjc over to semweb efforts, i think

21:39:04 * DanC_jam is curious about how the DSDL collaboration works; meeting logistics, email, phone, chairs... that sort of thing

21:39:25 <kendallclark> with both jjcs, our technique would be unstoppable

21:39:31 * DanC_jam is quite content to let jjc choose his own projects

21:39:53 <evlist> Basically it's meeting during Idealliance meetings and mailing lists hosted by myself.

21:40:31 <sandro> But it will be a good sign if Clark starts to adopt more SemWeb stuff, I think.

21:40:43 <DanC_jam> quite

21:41:09 <kendallclark> it's not clear he's temptable anyway, just an idle thought.

21:41:11 <DanC_jam> that he uses the nci ontology for his demos shows he's paying at least a little attention

21:41:24 <dajobe> true

21:41:34 <sandro> right

21:41:36 <kendallclark> and the several rdf references

21:41:46 <kendallclark> even if as canonical pathology

21:43:59 <DanC_jam> J:seems relevant to [http://www.w3.org/2001/tag/issues.html#mixedNamespaceMeaning-13|tag issue mixedNamespaceMeaning-13], which the TAG decided to reduce to 3 smaller issues

21:43:59 <dc_rdfig> Added comment J16.

21:44:46 <DanC_jam> J:though just because you can validate mixed-namespace documents, you don't necessarily grok their meaning

21:44:46 <dc_rdfig> Added comment J17.

21:45:02 <kendallclark> DanC: i can't believe i don't know this, but what's the deal w/ tag membership? does tim pick members?

21:45:33 <DanC_jam> tim picks 2 or 3, and the remainder are elected. it's quite an intricate process...

21:45:37 <DanC_jam> ... it seems to be working well.

21:45:42 <evlist> J:That's a feature: DSDL focus is on validation.

21:45:43 <dc_rdfig> Added comment J18.

21:46:01 <kendallclark> it's a good group, i was just wondering about turnover and future members and the like.

21:46:37 <DanC_jam> how is it a feature? it facilitates document composition, but does anybody want these composed documents? does anybody know how to consume them?

21:47:34 <DanC_jam> the 2nd tag election just confirmed the results from the 1st.

21:47:45 <DanC_jam> i.e. the same 9 stayed.

21:48:00 <kendallclark> ah. nice. i missed that.

21:48:35 <evlist>http://www.xmlconference.org/xmlusa/2003/wednesday.asp#33

21:48:35 <dc_rdfig> K: http://www.xmlconference.org/xmlusa/2003/wednesday.asp#33 from evlist

21:48:57 <evlist> K:|Caching in with Resolvers - Norm Walsh

21:48:58 <dc_rdfig> Titled item K.

21:49:14 <kendallclark> oh grumble!

21:49:22 <kendallclark> it gets worse and worse; clark is only 5 yrs older than me. -sigh-

21:49:30 <kendallclark> disgusting

21:51:45 <evlist> K:The real nice way to use URIs is to use absolute URIs on the network, but...

21:51:46 <dc_rdfig> Added comment K1.

21:52:28 <evlist> K:...networks go down, latency is sometimes significant.

21:52:28 <dc_rdfig> Added comment K2.

21:55:20 <evlist> K:Catalog based resolution is a solution.

21:55:20 <dc_rdfig> Added comment K3.

21:56:54 <evlist> K:XML Catalog (OASIS) takes the concept from SGML Open Catalog.

21:56:54 <dc_rdfig> Added comment K4.

21:56:59 <JeffSuttor> I build behind multiple proxies, firewalls, disconnected, etc. also sometimes want diff Public IDs to resolve differently, etc. catalogs are a survival solution

21:57:07 * DanC_jam hopes the topicmap folks grok what norm's presenting

21:58:55 <edd> DanC_jam: whyso?

22:02:29 <DanC_jam> as of Extreme Markup this year, they were complaining about URIs because the net isn't 100% reliable.

22:02:36 <DanC_jam> uris can be made arbitrarily reliable

22:02:47 * DanC_jam looks up notes...

22:03:05 <Ankh> only arbtratily close to reliable... same as everyting else

22:03:25 <edd> aha, i see/

22:03:38 <DanC_jam> 2003-08-05.txt:14:13:27 <DanC-AIM> Sigh.. Tommie is plugging xml catalogs. Do we really want to go back to hosts.txt files?

22:04:28 <Ankh> xml catalogs can also be viewed as a simplistic form of client-side managed cache

22:04:34 <DanC_jam> exactly.

22:04:48 <DanC_jam> catalogs are great, as long as they're keyed by URI

22:04:49 <Ankh> (assuming you don't get sucked in to the SGML "public identifer" nonsense)

22:04:54 <Ankh> yaeah

22:07:43 <DanC_jam> is norm going to point out that you can also use a generic http cache, like wwwoffle?

22:07:53 <evlist> suspense!

22:08:10 <danbri> Ankh, I meet Howard K, was good to talk to him...

22:08:57 <Ankh> danbri, cool

22:09:07 * DanC_jam q+ to note, re lack of catalog support in jing, that wwwoffle or the like, a generic http cache, would work

22:09:21 * DanC_jam q-

22:09:23 <edd> hehe

22:09:26 <evlist> K:Caching proxies are another solution.

22:09:26 <dc_rdfig> Added comment K5.

22:10:20 <evlist> K5:Caching proxies (such as WWWOffle) are another solution.

22:10:20 <dc_rdfig> Replaced comment K5.

22:10:51 <kendallclark> you gotta love a proxy that removes spam and adverts from schema files! :>

22:11:19 <DanC_jam> I'm pretty sure, though I can't confirm, that the "URIs can't work cuz the network breaks" came up during a topicmap patterns talk. http://rdfig.xmlhack.com/2003/08/05/2003-08-05.html#1060118110.032253

22:11:30 <evlist> or translate them to more friendly schema languages!

22:12:00 <DanC_jam> ooh... norm groks wwwoffle more deeply than I do... those Purge rules look really handy!

22:12:01 <kendallclark> transcoding intermediary schema switching language!

22:12:07 <kendallclark> tissl!

22:12:09 <evlist> yes.

22:13:58 <DanC_jam> (walsh misspoke; not a HEAD, but a GET/If-Modified-since)

22:15:47 * evlist apt-gets wwwoffle

22:15:59 <edd> i used to use it a lot before i got broadband

22:17:14 <kendallclark>http://uri.net/ddds-faq.html

22:17:14 <dc_rdfig> L: http://uri.net/ddds-faq.html from kendallclark

22:17:46 <evlist> K:Catalogs and cache work well together, especially if catalog delegation is used.

22:17:46 <dc_rdfig> Added comment K6.

22:17:53 <kendallclark> L:|Dynamic Delegation Discovery System

22:17:53 <dc_rdfig> Titled item L.

22:18:06 <kendallclark> (crap, shoulda put that into K:)

22:18:44 <DanC_jam> parting thoughts slide should say "they work together with URIs, but *not* with FPIs"

22:18:57 <kendallclark>http://www.ietf.org/rfc/rfc3401.txt

22:18:57 <dc_rdfig> M: http://www.ietf.org/rfc/rfc3401.txt from kendallclark

22:19:21 <kendallclark> M:|RFC 3401, DDDS, Pt 1: The Comprehensive DDDS

22:19:21 <dc_rdfig> Titled item M.

22:19:43 <kendallclark> MarkB: See also RFCs 3402 through 3405/06

22:19:50 <kendallclark> M:See also RFCs 3402 through 3405/06

22:19:50 <dc_rdfig> Added comment M1.

22:22:53 <edd> yay, i just notched up a score for sensible URI naming in Gossip.

22:23:09 <edd> (Gossip's a jabber client, was using the "gossip:" uri as a namespace for its log format)

22:23:09 <kendallclark> gossip?

22:23:58 <LotR> should it just use file: uris or what?

22:24:41 <edd> i suggested they use and http: uri and document their log format at that place.

22:25:17 <edd> inventing new uri schemes on a per app basis is a nasty habit that some gnome folk have inadvertently slipped into

22:26:37 <kendallclark> and they don't even x- preface them, ime

22:26:40 <LotR> yeah. witness their help browser

22:27:03 <LotR> info: man: help: eww.

22:27:21 <kendallclark> scrollkeeper? it's a mess in many ways.

22:28:25 * DanC_jam awards edd 350 points for preventing UriSpaceSquatting

22:28:26 * edd cashes them in for wn:Beer

22:28:37 <LotR> no, gnome-help-browser in 1.0. I forgot what they called the 2.0 one

22:28:52 <edd> yelp

22:28:52 <LotR> isn't scrollkeeper originally debian's

22:29:00 <kendallclark> no

22:29:10 <kendallclark> well, i don't believe so

22:29:11 <DanC_jam> this is pretty much a TAG meeting. 1/2 ;-)

22:29:29 <kendallclark> well, a dream tag meeting, tag + jjc :>

22:30:39 <edd> +1000 on packaging

22:30:47 <edd> adobe seemed to have invented a packaging format for what i saw with their xml document bundling stuff from the keynote

22:31:30 <DanC_jam> .googke wpk connolly

22:31:36 * edd -> reception

22:31:41 <DanC_jam> .google wpk connolly

22:31:42 <datum> wpk connolly: http://www.w3.org/2000/09/wpk43/Overview.html

22:31:51 <DanC_jam> ^ my proposal for packaging

22:32:34 <LotR> oh, heh. I was wondering how wpk would resolve to dan :)

22:51:36 <kurro> :)

23:03:00 <D[a]vey> D[a]vey is now known as Davey

23:09:51 * eikeon browses http://esw.w3.org/topic/DualUseUri

23:16:26 <mdupont> mdupont is now known as md-zzzz

23:22:40 <libby> !picpath kendall,sinatra

23:22:41 <whwhwhwh> Kendall Clark to Frank Sinatra via Edd Dumbill, Libby Miller, Dan Connolly, Tim Berners-Lee, Bill Clinton, John F. Kennedy

23:22:42 <whwhwhwh> in 7 steps

23:22:43 <whwhwhwh> see http://swordfish.rdfweb.org/discovery/2002/02/paths/byname.jsp?name=kendall&name2=sinatra

23:23:03 <eikeon> !picpath eikeon,libby

23:23:05 <whwhwhwh> no results found, sorry

23:23:07 <eikeon> :(

23:23:44 <eikeon> !picpath eikeon,kendall

23:23:46 <whwhwhwh> no results found, sorry