Semantic Web Interest Group IRC Chat Logs for 2003-11-03

00:04:06 <GeaDDon> GeaDDon is now known as eaon|zZz

00:09:53 <grrrrr> table bindings on conjunction? come again?

00:11:09 <grrrrr> the left side of the OR returns thing bob

00:11:25 <grrrrr> the right side returns fido

00:12:26 <grrrrr> 2 records. i didnt build a test case, the "proper" behavior is to return 2 thing records, of the 3 things correct?

00:13:08 <ericP> the ...FROM Things,Types WHERE Things.name="bob"... doesn't cause you to get the "bob" answer three times?

00:13:16 <grrrrr> the simple advent of changeing your sql to prepared statements might actually get smarter returns

00:14:05 <ericP> i don't believe it until you try it.

00:14:07 <grrrrr> ericP, your query might crash on ACCESS but would work on postgres for 2, now i want to test

00:21:08 <Davey> Davey is now known as D[a]vey

00:36:34 <grrrrr> oh well first off, the where clause leaks

00:37:39 <grrrrr> select * from thing,typ where typ.id=thing.typ and (name='bob' or class='dog')

00:38:05 <grrrrr> the OR in your clause derails the type.id=typ

00:42:49 <grrrrr> see http://www.hellaweb.com/hellabot/369

00:45:25 <ericP> i am jack's complete lack of surprise

00:45:39 <ericP> that's what i've been saying

00:46:42 <ericP> yes you can remap it, but it's hard work in the general case

00:46:42 <ericP> and that's what i'm doing

00:46:42 <ericP> and ORM does noting tohelp me

00:47:07 <ericP> sure it might change the syntax a littel, but that's hardly my problem

00:50:30 <grrrrr> is the parenthetical syntax somehow innaplicable to what you're attempting?

00:51:50 <grrrrr> there

00:51:57 <ericP> ain't gonna help unless i can do joins in parentheses

00:52:34 <grrrrr> subqueries?

00:55:00 <grrrrr> the query i pasted uses parens. its basic sql.

00:55:39 <grrrrr> with or without parens, the returned rows will all share the same number of columns, but without a join they'll throw in a combinatorial, and then add the OR anyways

00:57:49 <ericP> the best solution seems to be a mapping to outer joins

00:57:52 <ericP> see http://www.w3.org/2003/11/01-RDF-SQL-tmp/notes

00:58:05 <ericP> i assume you've already seen it

00:58:12 <ericP> do you remember the graph query example?

00:58:18 <ericP> and the SQL query mapping underneath?

00:58:30 <ericP> the amapping was done by hand. i want to automate it

00:58:58 <grrrrr> which figure is it?

00:58:59 <ericP> and it seems i just just buckle down and do it

00:59:18 <ericP> see Example 1: disjunctive query where left side has two terms and right side has three

01:00:25 <ericP> us RDF folks aren't so stupid after all

01:00:49 <ericP> just trying to solve hard problems

01:02:06 <grrrrr> the data and query are unitellegiable symbols, it looks like the output of a code obfuscator, you realize that, right?

04:40:42 <aszs>http://rx4rdf.liminalzone.org

04:40:42 <dc_rdfig> A: http://rx4rdf.liminalzone.org from aszs

04:41:20 <aszs> A: i just noticed you guys discovered this site a few days ago before it went live

04:41:20 <dc_rdfig> Added comment A1.

04:41:50 <aszs> A: its up now -- check it out

04:41:50 <dc_rdfig> Added comment A2.

04:45:12 <aszs> A: rx4rdf provides a deterministic mapping between the RDF abstract syntax to the XPath data model, allowing xpath, xslt and xupdate to be applied to a RDF graph

04:45:13 <dc_rdfig> Added comment A3.

04:47:35 <aszs> A: rhizome is a wiki-like content management and delivery system that is built on rx4rdf and stores everything in a rdf store

04:47:35 <dc_rdfig> Added comment A4.

04:48:38 <aszs> a: it lets you create abstract RDF resources the same way you'd create a wiki name

04:51:37 <aszs> a: it also introduces rhizml and rxml, wiki-like text formatting languages that provide alternative serializations for XML and RDF, respectively

04:51:58 <aszs> A: it lets you create abstract RDF resources the same way you'd create a wiki name

04:51:58 <dc_rdfig> Added comment A5.

04:52:12 <aszs> A: it also introduces rhizml and rxml, wiki-like text formatting languages that provide alternative serializations for XML and RDF, respectively

04:52:12 <dc_rdfig> Added comment A6.

04:58:40 <niq> sbp, you around/alive?

04:59:11 <sbp> yeah, hi niq

04:59:19 <sbp> sorry I haven't posted the thingy yet

04:59:23 <sbp> I'm still aware I've got to do it

04:59:34 <niq> I'm looking @ http://infomesh.net/2002/rdfinhtml/

04:59:49 <niq> (someone asked about it)

05:00:10 <sbp> it's a bit out-of-date now

05:00:44 <niq> You might perhaps add a mention of the difference between HTML and XHTML re: contents of <script> (since you suggest that as a possibility)

05:01:33 <sbp> yeah, point. there are quite a few things like that that I need to add, though. there was a taskforce on it quite recently that explored a number of approaches, and all the feedback from there needs to be reincorporated

05:01:52 <sbp> for example, Masayasu Ishikawa spoke about the <script> hack, so that needs to be put in

05:02:03 <sbp> but I think it'd be best to do a new document from sractch

05:02:06 <niq> - HTML makes it CDATA, so that "works" better

05:02:12 <sbp> er, scratch. I just don't have the time now

05:03:00 <niq> fairy nuff

05:03:08 <sbp>http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2003Jun/0035

05:03:08 <dc_rdfig> B: http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2003Jun/0035 from sbp

05:03:29 <sbp> B:|Re: Validating XHTML with embedded RDF, Masayasu Ishikawa to public-rdf-in-xhtml-tf

05:03:29 <dc_rdfig> Titled item B.

05:04:10 <sbp> B:Summary, the <script> hack doesn't work in XHTML... unless you use <![CDATA[]]>, but it doesn't come recommended

05:04:10 <dc_rdfig> Added comment B1.

05:25:22 <jql> hrm. that was indeed ugly

06:04:41 <jeremiah> I came up with a zero-knowledge way of storing triples, as might be useful in a DHT, can someone please read the description and tell me if it's already been done, I'm actively searching now as well

06:04:43 <jeremiah>http://www.kingprimate.com/weblog/archives/2003_11_02.html#09:30PM

06:04:43 <dc_rdfig> C: http://www.kingprimate.com/weblog/archives/2003_11_02.html#09:30PM from jeremiah

06:13:19 <jeremiah> C:|zero-knowledge triple storage

06:13:19 <dc_rdfig> Titled item C.

06:14:53 <jeremiah> C: This might be useful for an RDF DHT, it's probably already been done though.

06:14:53 <dc_rdfig> Added comment C1.

06:27:23 <sbp> DHT?

06:27:32 <sbp> reminds me a lot about semplesh...

06:35:19 <jeremiah> sbp I worked on semplesh

06:35:28 <jeremiah> but then nothing came of it, but it's been in my head

06:36:15 <sbp> yeah. I remember. me too

06:37:18 <jeremiah> is anyone working on RDF DHTs that you know of now?

06:37:54 <sbp> I don't even know what DHT stands for. Distributed Hash Table, at a guess

06:38:00 <jeremiah> yeah

06:38:08 <sbp> then nope. I don't... sorry

06:38:26 <sbp> haven't even heard the general concept repeated

06:38:34 <jeremiah> that's sad

06:38:39 <jeremiah> I think it'd be insanely useful

06:38:40 <sbp> one part P2P, one part RDF, one part SW...

06:38:47 <sbp> of course. but it's a major project, too

06:39:05 <sbp> needs some bram-ish coders to get involved. I think that's why Aaron didn't pursue it

06:39:06 <jeremiah> yeah

06:39:10 <sbp> because it was just too damn big

06:39:26 <jeremiah> I know some coders of that caliber though

06:40:00 <jeremiah> over freenet is also a possibility, but then it'd be so slow as to be useless

06:40:07 <sbp> hehe

06:40:19 <sbp> and it'd require Java, and not work at all on many systems

06:40:28 <jeremiah> right

06:48:19 <jeremiah> alright I'm out for the night

07:04:52 <jql> jql is now known as jql-zzz

09:02:15 <Davey> Davey is now known as D[a]vey

09:07:44 <grrrrr_> grrrrr_ is now known as grrrrr

09:07:50 <grrrrr> grrrrr is now known as grrrrrr

09:08:52 <grrrrrr> grrrrrr is now known as grrrrr

09:14:19 <mdupont> mdupont is now known as md-work

10:32:02 <D[a]vey> D[a]vey is now known as Davey

10:44:40 <eaon|zZz> eaon|zZz is now known as eaon

11:25:45 <Davey> Davey is now known as D[a]vey

11:36:40 <eaon> eaon is now known as eaon|out

12:05:07 <dajobe> A:|Rx4RDF

12:05:08 <dc_rdfig> Titled item A.

12:05:35 <dajobe> C:DHT=Distributed Hash Tables

12:05:35 <dc_rdfig> Added comment C2.

12:07:50 <Arnia22> Arnia22 is now known as Arnia

12:54:08 <arnarl> hi

14:37:07 <danbri>http://sesame.aidministrator.nl/publications/users/ch05s06.html

14:37:07 <dc_rdfig> D: http://sesame.aidministrator.nl/publications/users/ch05s06.html from danbri

14:37:30 <danbri> D:|"Construct" queries in Sesame's SeRQL query language

14:37:30 <dc_rdfig> Titled item D.

14:37:51 <danbri> D:Looks like a rule language to me. Can someone explain to me where rules stop and queries start?

14:37:52 <dc_rdfig> Added comment D1.

14:46:13 <AndyS> Does this allow for backward chaining over multiple rules?

14:50:03 <AndyS> Does this provide a stream of (small) graphs or one large, new graph?

14:51:28 <dajobe> I dunno, but FYI Jeen and Arjohn are keen to discuss SeRQL in Amsterdam

14:54:31 <AndyS> danbri - what's your take on rules vs query?

14:59:49 <GregElin> Morning all! Dan, Libby -- got a moment?

15:00:43 <shellac> libby's not here, dan might be in a telecon

15:01:00 <shellac> unless you wanted dan libby?

15:01:32 <GregElin> No...it was (Dan || Libby) || (Dan && Libby)

15:02:12 <GregElin> (thanks shellac)

15:02:17 <JibberJim> dan && libby would never be evaluated...

15:02:28 <JibberJim> well never to true anyway

15:02:53 <shellac> (dan ^ libby) || (dan && libby)?

15:03:02 <GregElin> Wouldn't that be Dan.add(Libby) and Libby.add(Dan)?

15:03:35 <GregElin> Okay...looking for either.

15:03:51 <GregElin> And Jim, still interested in a image metadata chat in Nov?

15:05:05 <JibberJim> Yep, Larry O'Brien (sp?) seemed to also be doing some interesting stuff in the area, on his tablet PC.

15:06:52 <GregElin> Ooo...tabletPC! How exciting. I keep thinking I want to get one and play.

15:07:20 <GregElin> Jim, do you usually attend the www200x conferences?

15:08:04 <JibberJim> I never have done before, and I feel it's unlikely I will next year

15:08:19 <GregElin> 2004 is in NYC...

15:08:50 <JibberJim> yeah, too much hassle at border...

15:19:06 <GregElin> ...and so ends the great experiment formerly known as America. Not with a bang from an enemy's hatred but with the internal whimper of national security.

15:21:49 <dajobe> several of us from here are likely to be at WWW2004 - myself, libby, danbri and some others

15:22:19 <GregElin> Terrific.

15:23:20 <GregElin> I've volunteered with the local hosts, NYUCAT and IBM, to coordinate digital photos for the event.

15:23:43 <dajobe> ooh

15:23:52 <dajobe> maybe we can do some foafy thing

15:24:03 <dajobe> we (edd & I) made a chump web site last year

15:24:06 <GregElin> That would be great!

15:24:11 <dajobe> see http://www2003.xmlhack.com/

15:24:17 <GregElin> let me check that out...

15:24:30 <dajobe> I was wondering about doing something more planned - have a blog, trackbacks, chump, irc, &c you name it

15:25:33 <GregElin> Okay. You'd think I'd know by now that Chimera can't read svg...let try again.

15:25:46 <dajobe> :)

15:25:55 <GregElin> I'm thinking something along the lines of a semantic photo archive would be very intersting...

15:25:57 <dajobe> no easy svg soltuions for us linux users either

15:25:58 <dajobe> yes

15:26:14 <dajobe> I think mattb & mortenf might be good to have along with that

15:26:30 * danbri catches up, sorry AndyS missed your comments; have been in telecon and have too many irc windows open!

15:26:31 * chrisc has still never see a svg rendered... :-(

15:26:32 <dajobe> look at vocabs and gps, geocodings

15:26:36 <GregElin> Gather photos at 2004, but also try to get people to put together an archive from past years...try and go back to 1995, 1994.

15:26:48 <dajobe> hmm

15:26:49 <danbri> <AndyS> danbri - what's your take on rules vs query?

15:27:18 <danbri> that they're really hard to scope, that there's arguments for keeping together and for separating, also overlaps w/ OWL, with XQuery etc distress me!

15:28:38 <GregElin> msg danbri Hey. Did you see the past couple of lines? I'm putting together a proposal for a semantic photo archive for www2004.

15:29:19 <AndyS> One factor is the timescale - how long it takes for a WG to get to rec. What thoughts there? Long, and complete vs short and steping stone?

15:29:26 <danbri> greg, sounds good. can't listen now am on phone.

15:30:47 <GregElin> I like www2004.xmlhack.com - be interesting if we could add a reference photo with the chumped information.

15:31:43 <dajobe> yeah

15:32:11 <dajobe> I'll talk to edd about it. we did www2003 <7 days before the conf; this time we can plan a bit more

15:33:22 <GregElin> Sounds just great, dajobe. I'll gather some more info on my end. I've never been to a www200x before, so I am not familiar with the protocal for what's official v. community and so on.

15:34:16 <GregElin> But we are trying to do some interesting things here to get NYC tech companies involved and have stuff going on in the city that would be of interest to the audience.

15:35:13 <dajobe> GregElin: it's never been official, just pretty handy since the wireless is now pretty much pervasive at the web conferences

15:35:25 <dajobe> and, ahem, not an OSX/i-tool only

15:36:57 <GregElin> so many interesting things to experiment with. One might think with wireless and other things...the community could document the whole conference quite nicely: chumpbots, wikis, blogs, photos, movies, audio files...

15:37:51 <dajobe> yes, but trying not to make it all too fragile

15:37:57 <dajobe> so using semweb tech to make that happen

15:38:11 <dajobe> foaf for people, rss 1.0 for feeds, and other vocabs for photo pics etc.

15:38:28 <GregElin> I like the sound of that.

15:38:32 <dajobe> applying the various tools we have to make news-style or phoot-style, or rdf-style views of the data

15:38:40 <dajobe> and as open data, no barriers

15:38:49 <GregElin> Right.

15:38:52 <dajobe> photo-style

15:39:05 <dajobe> it's almost as if there were a web of semantics!

15:39:24 <dajobe> maybe we should make a wiki page to start

15:39:29 <dajobe> is esw.w3.org up, hmm

15:39:38 <dajobe> yes

15:39:39 <GregElin> Plus, we could end up with a really rich photo archive (with a creative commons license) for ongoing experimentation.

15:39:48 <dajobe> BLURB:WWW2004 thoughts

15:39:48 <dc_rdfig> E: WWW2004 thoughts from dajobe

15:39:52 <dajobe> logger, chump E

15:39:52 <dajobe> E:See [http://ilrt.org/discovery/chatlogs/rdfig/2003-11-03#T15-39-52|discussion]

15:39:53 <dc_rdfig> Added comment E1.

15:41:06 <GregElin> Would like the photos to be more than just pictures of people (a) at podium (b) on panel and (c) eating and drinking in small circles

15:41:34 <dajobe> howso?

15:42:05 <dajobe> E:[http://esw.w3.org/topic/WWW2004|WWW2004 thoughts] wiki page

15:42:06 <dc_rdfig> Added comment E2.

15:42:13 <dajobe> E:please edit it, I suck at wiki text

15:42:13 <dc_rdfig> Added comment E3.

15:42:19 <dajobe> hey libby

15:42:21 <GregElin> We need photos of things, too: people's first servers, early computers, landmark "experiments"

15:43:05 <GregElin> Hi Libby.

15:44:45 <libby> hey greg

15:45:26 <GregElin> Just chatting with dajobe regarding putting together a semantic photo archive for the WWW2004 conference, which will be here in NYC in May.

15:45:47 <libby> yes, I just noticed. that'd be neat. could do it very easily too (in rdf).... :)

15:46:28 <GregElin> Yeah! I'm thinking that would be the date where we get a bunch of things working together: fotonotes, rdf, etc.

15:46:50 <libby> e.g. <Vevent><foaf:homepage="http://www2004.org"/><foaf:depiction rdf:resource="http://somephoto"/></Vevent>

15:46:56 <libby> then just harvest them all up

15:47:07 <libby> that's be neat

15:49:47 <GregElin> Excellent. Really glad you like the idea. I'm working on a little written statement to share with the local hosts regarding coordinating the digital photo gallery.

15:50:02 <GregElin> Wanted to know if you'd be interested in helping out.

15:50:10 <GregElin> Getting involved, etc.

15:51:20 <libby> that'd be cool, yes indeed

15:51:49 <GregElin> Great!

15:51:56 <_joshua> hey greg

15:51:59 <GregElin> I should have a draft written shortly.

15:52:12 <GregElin> Hey Josh! Have you been reading.

15:52:23 <_joshua> not really?

15:53:16 <GregElin> Discussing assembling a semantic photo archive for www2004. I'll catch you with messages...

15:53:44 <dajobe> I've been updating the wiki page with what you've said so far

15:55:45 <GregElin> dajobe, Libby - I started the wiki page [http://esw.w3.org/topic/WWW2004|WWW2004 thoughts]

15:56:16 <GregElin> brb...

15:56:19 <dajobe> uhoh

15:56:30 <dajobe> I just made one with WWW2004

15:57:15 * dajobe clearly failing with the wiki-Fu

16:05:49 * dajobe tries merging them

16:38:16 <GregElin> nice job of merging, dajobe.

16:49:17 <dajobe> ta

17:04:52 <md-work> md-work is now known as mdupont

20:01:49 <ericP> could someone check my SeRQL example --> http://www.w3.org/2001/11/13-RDF-Query-Rules/#SeRQL ?

20:39:17 * ronwalf munges foaf into OwlDL

23:31:33 * DanC answers a random DAML question... whee! just like HTML and XML...