Semantic Web Interest Group IRC Chat Logs for 2003-10-20

00:08:30 <JibberJim> Um, isn't that time correct for CEST?

00:09:25 <deltab> doh, CEST, right

00:09:30 <deltab> sorry

00:10:05 <deltab> I got confused by the use of CET

00:31:29 * mmealling shows his web services ignorance by asking if anyone has seen a web service that does a form of capabilities negotiation by returning snippets of WSDL when asked what methods it supports.

00:51:32 * sandro_ shares mmealling's ignorance, but thinks it sounds cool.

00:51:36 <sandro_> sandro_ is now known as sandro

00:56:57 <mmealling> I'm writing a standard for a group that needs a standard way of expressinga whole bunch of different kinds of queries. So I'd like to be able to connect to a service, ask it which queries/methods it supports in the form of a unique URI for each one, then it can also give me the WSDL for each one as well.....

01:50:06 <eaon> eaon is now known as eaon|zZz

02:07:10 * MarkB wonders what methods you use to ask a web service what methods to use ...

02:07:49 <JibberJim> You could use a web service perhaps?

02:07:56 <MarkB> brilliant!

02:08:05 <MarkB> 8-)

02:08:17 <JibberJim> That's one problem solved tonight, I can go to sleep now.

02:08:49 <MarkB> crisis averted. night night.

02:10:45 <MarkB> mmealing ... just to say that I don't care much for Web services. Prefer the Web instead.

02:11:17 <MarkB> oops, ^mealing^mealling

02:11:26 <MarkB> MarkB is now known as MarkB_Zzz

04:22:17 <chaalsNCE_> sh1mmer?

04:23:53 <chaalsNCE_> have you considered the idea of whitelists based on traceable trust delegation?

04:25:09 <chaalsNCE_> Danbri and I discussed it a while ago, and I was unconvinced that a trust-beased network couldn't be broken by spammers, by setting in place a number of legitimate paths, then letting some of the links go dead

04:25:31 <chaalsNCE_> (this is for peer-to-peer type certification)

04:26:29 <chaalsNCE_> after using them to pass on reliable trust information and set up lots of nodes that can later be used to authenticate spammers.

04:28:35 <chaalsNCE_> If you pass on lots of certifications, claiming that you personally certify them, having harvested them from moderately reliable sources, you should be able to add 10% of that number as spammers.

04:30:03 <chaalsNCE_> 1 million real addresses would give you enough dodgy ones to me a pain for a long time, and there are lots of verifiable real addresses if you trawl mailing lists.

04:30:25 <chaalsNCE_> I am actually hoping to find a flaw in my argument, but not succeeding so far.

04:31:46 <sh1mmer> moin chaalsNCE_

04:32:14 <sh1mmer> sorry i was busy working on something for a client

04:32:34 <sh1mmer> which is why i am up at this ridiculous hour (considering i have meetings starting at 9)

04:33:43 <sh1mmer> chaalsNCE_ this is old skool pgpnet and foaf type stuff yes?

04:35:10 <chaalsNCE_> yep.

04:35:23 <sh1mmer> the problem is the point of entry

04:36:04 <sh1mmer> also i have reservations, googles page ranking is something very similar and the spammers got into that pretty well

04:36:22 <sh1mmer> set up fake pages praising their websites

04:36:32 <sh1mmer> thoasands of them to boost rankings

04:37:21 <chaalsNCE_> true. A network of spammers would do better than a single spammer, and you can assume that they all have real lives and send legitimate email to people. So you can establish your own credentials as bona fide and not put anyhting too close to your real self.

04:38:05 <sh1mmer> charles the problem is still the point of entry. what if you know noone have just joined the internet and want to send mail?

04:38:30 <chaalsNCE_> then you don't have anyone to send mail to anyway.

04:38:33 <sh1mmer> the spammer want to set up lots of free yahoo accounts to spam with for 5 mins before the account gets shut then open another

04:39:27 <sh1mmer> you might have people you know but havent set up relationships with not having your own email addresss until you sign up to yahoo, only you cant because you have no trust

04:40:08 <chaalsNCE_> the point is you can sign up, but almost nobody will accept your email.

04:40:13 <sh1mmer> foaf has benefits but it suffers from legitimate low yield users differentiating themselves

04:40:37 <sh1mmer> thats not good enough. that wastes yahoo's server load

04:40:42 <chaalsNCE_> you and your friend sign each other's keys.

04:40:58 <chaalsNCE_> No, if people can't usefully send email from yahoo they stop signing up.

04:41:06 <chaalsNCE_> (car locator vs club lock)

04:41:34 <sh1mmer> being a cipherpunk i can tell you locator will never work on the net until ipv6

04:42:11 <sh1mmer> in terms of computer security point of entry is what yahoo, aol etc's biggest weakness is right now

04:42:27 * chaalsNCE_ wonders what makes ipv6 better...

04:42:59 <sh1mmer> there is better security built in at the packet level

04:43:13 <sh1mmer> although frankly a lot more could be done atm

04:43:19 <sh1mmer> thats another debate

04:43:39 <Mutiny> i bet yahoo's whack-a-mole scripts are majestic

04:43:40 <chaalsNCE_> the problem with a system like this for yahoo etc is that they need to get enough critical mass before they rely on it.

04:44:07 <chaalsNCE_> (and my argumment is that even then it is unreliable, although I would love to find out that I am wrong)

04:44:11 <sh1mmer> if the spammers can sign up new accounts in the 5 mins they use the account before it gets locked out for spamming they use 0.00x % of a server, if they make 100k new accounts...

04:44:33 <sh1mmer> the spammers have low hits anyway, unless everyone uses whitelisting it wont work

04:44:47 <chaalsNCE_> everyone: right.

04:45:11 <sh1mmer> as it happens i wrote a paper on white, grey, black and various other advanced email filtering facilities. not specifically aimed at spam but it was included

04:45:36 * chaalsNCE_ waits for a reference... ;-)

04:45:41 <sh1mmer> lol ill dig it ouy

04:45:42 <sh1mmer> out

04:45:56 <sh1mmer> it will be on a cd in my drawer somewhere

04:46:24 * chaalsNCE_ recently found a couple of those that were thought lost to history.

04:46:54 <sh1mmer> there is a great book

04:46:58 <sh1mmer> let me find the reference to it

04:49:45 <sh1mmer> Kaitlin Duck Sherwood's Overcome Email Overload book

04:49:47 <sh1mmer>http://www.overcomeemailoverload.com/eudora/html/index.html

04:49:47 <dc_rdfig> A: http://www.overcomeemailoverload.com/eudora/html/index.html from sh1mmer

04:49:54 <sh1mmer> damn

04:50:08 <sh1mmer> charles can you fix that compeltey non rdf related link?

04:52:38 <chaalsNCE_> ummm

04:53:03 <chaalsNCE_> A:|Overcome email overload - a book

04:53:03 <dc_rdfig> Titled item A.

04:53:26 <sh1mmer> er ok. not what i had in mind :)

04:53:35 <sh1mmer> i have no idea how to work that bot at all

04:53:42 <chaalsNCE_> A: stray reference from a discussion about whitelist-based filtering and shared trust.

04:53:42 <dc_rdfig> Added comment A1.

04:54:52 <sh1mmer> so anyway i know ducky through OSAF

04:56:02 <sh1mmer> mitch kapor's new company i keep an eye on the mailing list for chandler

04:56:03 <chaalsNCE_> A: unfortunately the chumpbot doesn't nkow where the logs are :( so you have to search manually. (How un-semantic is that ;-)

04:56:03 <dc_rdfig> Added comment A2.

04:56:26 * chaalsNCE_ offers tom the chumpbot manual: http://usefulinc.com/chump/MANUAL.txt

04:57:34 <chaalsNCE_> actually the chumpbot is hella cool, but I ahve a couple of tweaks to really make it rock....... as soon as I learn PERL (in my non-existent spare time)

04:59:15 <sh1mmer> maybe i will lend a hand at some point

04:59:40 <sh1mmer> perl is groovie but its ultra sparse lingo takes a while to get usable at

05:02:20 <chaalsNCE_> heh, the book is quite cool. But has a grammar mistake...

05:03:45 <chaalsNCE_> " before your correspondents--or whomever they forward the message to-- see the message."

05:04:37 <sh1mmer> chaalsNCE_ oh yes. thanks for the spelling correction on the wcag list. just so you know my current mua of choice is m2 and it doesnt spell check. i am also dyslexic. i didnt take offence and i dont mind corrections but i thought you should know

05:05:04 <chaalsNCE_> s/--[^-]*--/ or whoever (it is) to whom they forward it/

05:05:32 <sh1mmer> a comma would have sufficied.

05:05:46 <chaalsNCE_> (there's my little bit of pedantry for the day. Actually the "it" is pretty dodgy, but whatever)

05:07:30 <chaalsNCE_> sh1mmer, yeah, I noted that you said you were dyslexic. I don't use a spell checker much (certainly less than I should, since I type messily). The spelling risked distracting from the point, so I am glad it isn't at iss

05:07:39 <chaalsNCE_> s/at iss/the issue/

05:11:40 <sh1mmer> My speeling (haha) isn't that bad really. It's moslty not caring to take the trouble, since m2 haven't add the spell checker to 7.2 (I know it's in 8.0) I don't have much choice.

05:12:36 <chaalsNCE_> there is an apache module called mod-speling - at least we have one by that name at w3c

05:13:14 <sh1mmer> non interactive spell checkers are pointless for dyslexics.

05:13:22 <sh1mmer> and nothing plugs in, i have looked

07:25:41 <eaon|zZz> eaon|zZz is now known as eaon

07:53:18 <arnarl> hi

08:12:46 <mdupont> mdupont is now known as md-bk-2200CET

08:36:11 <eaon> eaon is now known as eaon|out

09:44:26 <chaals> chaals is now known as chaalsNCE

10:21:00 <JibberJim> chaalsNCE, the logger knows where the logs are, and you can ask him for a bookmark to tell the chump

11:02:48 <d2m_> d2m_ is now known as d2m

11:07:45 <chaalsNCE> ah, it's true.

11:08:19 <chaalsNCE> so if the chump-bot asked for a log bookmark life woulb be easier

11:09:00 <danbri_> hey charles

11:10:05 * mattb vaguely remembers discussing such things with chaals at www

11:10:15 <mattb> maybe by next year's www something will be done about it...

11:13:52 * darobin points out that NYC is more amenable to getting very drunk in a weird club than to coding IRC bots

11:14:18 <mattb> ok, you convinced me

11:14:33 <JibberJim> Hmm, but some weird club's in NYC probably have wireless, so you could do both...

11:15:17 <darobin> good idea

11:15:30 <darobin> and I guess drunk is the right state to code bots in

11:18:06 * monkeyiq wonders what the "test data" for the bots would be, something a little informal :)

11:20:36 <danbri_> logger, pointer?

11:20:36 <danbri_> See http://ilrt.org/discovery/chatlogs/rdfig/2003-10-20#T11-20-36

11:20:52 * danbri_ imagines patching logger would be better than patching chump

11:21:38 <danbri_> ie... logger, chump pointer A:? -> A:See [http://ilrt.org/discovery/chatlogs/rdfig/2003-10-20#T11-20-36|IRC discussion].

11:21:42 <mattb> yup

11:22:27 <danbri_> use the src., chaals... http://cvs.ilrt.org/cvsweb/redland/logger/logger

11:22:34 <danbri_> (or buy dajobe beer?:)

11:22:42 <dajobe> mmm, beer

11:23:35 <dajobe> it's time for V2 of the logger bot; I had a quick look at a jabber base but it's IRC support seems poor

11:23:57 <danbri_> could move it to POE?

11:24:05 <dajobe> would that help?

11:24:15 * dajobe doesn't know POE

11:24:36 <mattb> it's a bit of an effort to learn, but it does give you a multi-protocol reactor/messagebus

11:24:42 <monkeyiq> .google POE

11:24:44 <datum> POE: http://www.eapoe.org/

11:24:44 <dc_rdfig> Label POE not found.

11:25:26 * chaalsNCE starts a dutch auction for beers in return for the logger patch...

11:25:34 * chaalsNCE waits for bids.

11:25:57 <danbri_> don't you need a starting figure?

11:25:58 <Wack> 'dutch' auction?

11:26:11 <darobin> you need a start if it's dutch

11:26:22 <danbri_> probably called an "English auction"in Holland

11:26:28 <darobin> heh

11:27:08 * monkeyiq assumes POE == Perl Object Env

11:27:14 <mattb> poe.perl.org

11:27:27 * dajobe shocked danbri hasn't mentioned a ruby solution yet :)

11:27:47 <danbri_> yeah, you could rewrite it in Perl 8

11:28:25 <danbri_> I met the guy who did Ruby's Jabber library, and YAML support... he's into RDF... involved in DAML somehow too

11:28:28 <danbri_> rich kilmer

11:29:22 <darobin> when does Ruby get CRAN?

11:29:58 <danbri_> RAA :)

11:30:25 <danbri_> there are RAA++ RAA-NG proposals and efforts floating around, I did some rdfish prototyping but lost touch...

11:48:55 <r> r is now known as eikeon

12:48:31 <JimJibber> JimJibber is now known as JibberJim

13:13:28 * chaalsNCE offers 5 pints for the first person who patches logger or chump, so that a pointer to the logs is automatically inserted into every chump.

13:15:07 * chaalsNCE offers 15 pints for a patch so that chump checks things that it chumps and if they have been chumped before it provides a pointer (either to log or to previous chump) in the channel and in the chump.

13:15:31 * chaalsNCE offers a bonus 5 if someone does the lot.

13:15:44 <monkeyiq> heh, the shipping is not included I assume ;)

13:16:25 <chaalsNCE> well, we have to find a mutually convenient place to drink them - if that proves impossible I'll figure out a way to make sure they can be had somewhere reaonable.

13:16:56 <chaalsNCE> reasonable.

13:17:14 * monkeyiq has many pints awaiting him around the world... if only he could fly :/

13:17:18 <chaalsNCE> but:

13:17:47 <chaalsNCE> one pint subtracted from each offer every two days...

13:18:14 * chaalsNCE wonders where monkeyiq is.

13:19:11 <danbri_> hand delivered...

13:19:22 * monkeyiq only notices generic w3c for chaalsNCE, assumes he is in states/eu.

13:20:03 <chaalsNCE> the last three letters of my IRC nick are the nearest airport IATA code, as a rule.

13:20:15 <chaalsNCE> (ZZZ / Yum / Oot break the rule)

13:21:58 <chaalsNCE> typical places: MEL, BRS, NCE. Reasonably common: BOS, MAD, SYD, SGP, BKK, LAX, AKL, SFO, NYC, FRA, LON, (mostly major transit hubs...)

13:22:17 <monkeyiq> oh well, france is a long way away :/

13:22:24 <chaalsNCE> In principle I am in Footscray 3011

13:22:52 * monkeyiq pats his airport data that JibberJim gave him

13:25:58 <chaalsNCE> google knows where I live, let alone the semantic web

13:26:52 <monkeyiq> heh, I try to avoid google knowing where I live. its knows that I am reified which is enough :)

13:27:01 <chaalsNCE> (But where I am at a given time is not so obviously available. As it happens I'm in the EU for a few more weeks)

13:29:26 * chaalsNCE is guessing that if I head up the Hume I'd get close...

13:30:03 <monkeyiq> heh, well, the IP address I am on at current does leak data :)

13:35:50 * mmealling wonders if RFC2016 actually described RESTful web services back in 1996....

13:47:40 * chaalsNCE figures that fair conditions are that the first chumped success scores.

13:51:22 <libby> if chumped before, and other people chumped it, might make a little 'community of interest' on the topic :)

13:51:45 <libby> i.e. the fact that people tried to chump it uis interesting in itself

13:51:55 <JibberJim> Is that chumped before ever?

13:52:32 <danbri_> within same day vs 'at some point' is important distinction in the chump universe

13:52:54 * DanC_RSW waves

13:53:04 <chaalsNCE> in principle yes, although I would accept something that started recording from the date of its own existence. But 'within day' definitely doesn't qualify.

13:53:16 * chaalsNCE waves to dan.

13:53:26 * chaalsNCE is hoping to lose a lot of beer in a hurry.

13:54:19 <D[a]vey> D[a]vey is now known as Davey

13:54:26 * DanC_RSW breathes IP packets for the first time in almost a day

13:54:33 * DanC_RSW needs 110v somewhat urgently...

13:55:10 * libby should be able to show ever chumped before easily

13:55:19 <DanC_RSW> "... you can't have an attribute hanging around... you have to have the relevant element..." ugh. this workshop is getting into XML arcana?

13:55:20 <libby> ...but not as part of chump

13:58:27 <DanC_RSW> "... implemented in code." 'Complete Script in Rondo' slide

13:58:52 <DanC_RSW> "there's a paper in SigMod. code developed by sergey Melnick"

13:59:23 <dajobe> which workshop, the pssc?

13:59:35 <DanC_RSW> semantic integration

13:59:47 * DanC_RSW is trying to find the event where he is in http-space

13:59:52 <dajobe>http://smi.stanford.edu/si2003/

13:59:53 <dc_rdfig> B: http://smi.stanford.edu/si2003/ from dajobe

14:00:12 <dajobe> B:|SEMANTIC INTEGRATION WORKSHOP at ISWC 2003

14:00:12 <dc_rdfig> Titled item B.

14:00:25 <chaalsNCE> lib, if you can bolt that onto chump (for example as a bot that notes something being chumped and adds a comment) you'd be in for the 15 pints.

14:00:41 <chaalsNCE> (just don't drink them all on your birthday, please ;-)

14:00:51 <DanC_RSW> B:attendee [http://www.w3.org/People/Connolly/|Dan Connolly]

14:00:51 <dc_rdfig> Added comment B1.

14:00:59 <DanC_RSW> logger, pointer?

14:00:59 <DanC_RSW> See http://ilrt.org/discovery/chatlogs/rdfig/2003-10-20#T14-00-59

14:01:17 <DanC_RSW> B:cf [http://ilrt.org/discovery/chatlogs/rdfig/2003-10-20#T14-00-59|irc notes]

14:01:17 <dc_rdfig> Added comment B2.

14:01:21 <libby> hm, I think I'd fall over

14:01:39 <libby> do I have to drink them all in one session?

14:01:47 <DanC_RSW> B: is cyc:subEvents of [ foaf:homePage <http://iswc2003.semanticweb.org/> ].

14:01:47 <dc_rdfig> Added comment B3.

14:02:11 <DanC_RSW> "melnick developed similarity flooding; that's what he used here"

14:02:58 <DanC_RSW> B:packed room. firemarshall would go nutso

14:02:58 <dc_rdfig> Added comment B4.

14:03:06 <chaalsNCE> of course not.

14:03:12 <libby> heh

14:03:30 * chaalsNCE would prefer that you didn't, in fact.

14:03:50 * chaalsNCE notes that chaalsBRS would seriously prefer you didn't...

14:04:13 <JibberJim> You don't want to see libbyFallingOver ?

14:05:04 <DanC_RSW> PatH asks the speaker whether he's using "algebra" in the typical mathematical sense. [ Connolly has trouble finding the relevance of any of this to life as we know it ]

14:05:13 <dajobe> lol

14:05:23 <dajobe> maybe you should have gone to 'Practical' SSC

14:05:46 <DanC_RSW> This seems to be Invited talk: Phillip Bernstein "Generic Model Management: A Database Infrastructure for Schema Manipulation"

14:06:46 * DanC_RSW really wants to help PatH get connected to IRC as he asked the other day

14:07:18 <DanC_RSW> Welty: It's not a real algebra. ;-)

14:07:30 <DanC_RSW> B:attendee: Welty, Hayes

14:07:30 <dc_rdfig> Added comment B5.

14:07:55 <DanC_RSW> B:Invited talk: Phillip Bernstein "Generic Model Management: A Database Infrastructure for Schema Manipulation"

14:07:55 <dc_rdfig> Added comment B6.

14:10:03 * chaalsNCE doesn't want to get draawn into the conversation about libbyFallingOver (and (sort of;-) hopes that isn't a real WikiName anywhere ;-)

14:18:29 * chaalsNCE waves to reagle

14:19:55 <DanC_RSW> Hayes: I'm trying to relate this to Ontologies... how do you match [ontologies]?

14:20:05 <DanC_RSW> spkr: dunno; never tried...

14:20:28 <DanC_RSW> ?: I think the answer was in the 1st slides... he assumes models are structures... if you apply these techniques to the sentence structure, this stuff applies.

14:20:52 <DanC_RSW> [... something about negation...]

14:27:02 <Davey> Davey is now known as D[a]vey

14:27:18 <sh1m> chaalsNCE you around?

14:27:31 <chaalsNCE> yes...

14:28:08 <DanC_RSW> B:[Rondo|http://www-db.stanford.edu/~melnik/mm/rondo/] is the software Bernstein discusses

14:28:08 <dc_rdfig> Added comment B7.

14:30:05 <DanC_RSW> Bernstien: gotta have a good UI. 1st thing you'll hear from users is how klunky your system is.

14:30:25 <DanC_RSW> B:break...

14:30:25 <dc_rdfig> Added comment B8.

14:30:48 <mmealling> that needs to be said?

14:30:53 <DanC_RSW> B:attendee ter Horst

14:30:53 <dc_rdfig> Added comment B9.

14:31:02 <DanC_RSW> quite. "duh"

14:31:45 * reagleBRKLN waves back to chaalsNCE

14:31:59 * reagleBRKLN wonders why his irc client didn't highlight/catch that...

14:32:06 <mmealling> although I remember a time when I thought all I needed to do was show the underlying protocol doing its thing and people would just oooh and aaahhhh, and just get it. But that silliness disappeard quickly.

14:33:04 <sandro> It probably depends what your audience is, mm. If they are fellow protocol engineers, then sure.....

14:33:06 <reagleBRKLN> hrmm... highlighting is on...

14:33:13 * sandro waves to reagleBRKLN.

14:33:34 <reagleBRKLN> nope, kvirc isn't catching them for some reason

14:34:50 <mmealling> sandro, the sad thing I've learned over the past few years is that, while all of the protocol engineers could just love what you did, they really don't have the power to deploy anymore. You have to sell it to their bosses first.... gotta put lipstick on that pig since no matter how good the bacon, no one will ever take it to the dance.

14:35:25 <DanC_RSW> (oops.. I missed the 1st couple slides of this talk too; maybe there was some motivation in those slides, but it's now death-by-powerpoint)

14:35:52 <DanC_RSW> ah... getting medical info to the right place at the right time. I can grok that.

14:36:09 <sandro> ouch, DanC..... (On the other hand, +1 Laptops! )

14:36:38 <DanC_RSW> (mic is clipping. speaker blasts away...)

14:36:58 <sandro> Yeah, I understand, mm. I like to be able to explain my work to my friends and family, too, and that requires simple UI. (which let me tell you, doesnt happen very often. :-( )

14:37:49 <mmealling> no one in my family actually knows what I do for a living... I could be a drug dealer for all they know. ;-)

14:37:51 <DanC_RSW> "UMLS" on this slide... that's a buzzword I've heard EricM and Guus use.

14:38:34 <DanC_RSW> ask slashdot: how do you answer "what do you do for a living?" questions?

14:38:54 <dajobe> UMLS's a medical classification scheme

14:39:34 <mmealling> Cowboy Neals house boy....

14:39:53 * chaalsNCE got a nice comment from a psychiatric nurse who can barely operate a web browser, only knows "send" and "reply" in mail, and read my presentation from yesterday: "reminds me of semiotics stuff I did"

14:40:08 <sandro> "computers". "Oh? What kind?" "web. networking stuff." "Oh, .... my son does PHP4 servelets on Solaris....." or whatever.

14:41:15 <sandro> Yeah -- I'm always surprised how many people know that "ontology" actually means. Lots of people took a semester of philosophy or linguistics or something.

14:41:18 <mmealling> Mom: "What do you do again?" Me: "I design network protocols" Mom: "Oh, you make web pages?"

14:41:20 * chaalsNCE recalls the little johnny joke - in show and tell he says his mum's a prostitute and his dad's a pimp. The horrified teacher checks with them, and they admit they're really lawyers, but it's too hard to explain that to a kid.

14:41:31 * reagleBRKLN was thinking of writing a paper on the TAG URI issue and the semiotic issue (of Saussure)

14:41:44 <reagleBRKLN> of sign and signified

14:42:07 <reagleBRKLN>http://www.criticism.com/md/the_sign.html

14:42:07 <dc_rdfig> C: http://www.criticism.com/md/the_sign.html from reagleBRKLN

14:42:12 * chaalsNCE worked on accessibility for his mum before working for W3C. And was told that I should figure out what this metadata / semantic web thing was really about. At the time I didn't think much of it...

14:42:19 <DanC_RSW> in a cab ride last night, after baseball smalltalk, driver asked if I was in town for the "software conference". 'close enough,' I thought, and answered 'yes'. So I was quite surprised when I said "I work for this thing called the World Wide Web Consortium" and he said he knew what that was from investor's business daily.

14:42:26 <JibberJim> and now chaalsNCE?

14:42:27 <reagleBRKLN> c:The Sign, The Signifier, and The Signified

14:42:41 <reagleBRKLN> C:The Sign, The Signifier, and The Signified

14:42:42 <dc_rdfig> Added comment C1.

14:43:05 * chaalsNCE listens to what his mum says now.

14:43:33 <reagleBRKLN> So the semiotic stuff seemed relevant to the question of URI as a thing, or an identifier of a thing, etc.

14:43:33 <JibberJim> Glad to hear you've learnt your lesson.

14:43:39 <danbri_> so do I

14:44:16 <DanC_RSW> C:I've read stuff like this; maybe I don't get it, but it doesn't appeal to me at all.

14:44:17 <dc_rdfig> Added comment C2.

14:44:38 <DanC_RSW> Q/A time... Hayes obliges...

14:44:51 * chaalsNCE spends a lot of time dealing with "you know, this stuff is all very well in principle, but doesn't quite work in practice" mixed with ideas that I hope to understand soon as a result of learning that lesson.

14:45:01 <reagleBRKLN> yea, french theoreticians don't get me super psyched

14:46:10 <danbri_> Saussure's OK I think, but his terminology now is used largely by pomo types, who I really have a hard time reading

14:46:55 <reagleBRKLN> yes, that's what i'm buried in ;)

14:47:36 <reagleBRKLN> Saussure is a structuralist though :) (all the gibberish often gets cast as po-mo)

14:47:40 * DanC_RSW tries to remember the name of the guy who wrote "the name of the rose"

14:47:45 <reagleBRKLN> umberto eco

14:47:51 <DanC_RSW> B:SMAIL talk...

14:47:52 <dc_rdfig> Added comment B10.

14:47:52 <nmg> ack, beat me to it

14:47:55 * danbri_ was very amused by http://www.physics.nyu.edu/faculty/sokal/

14:48:04 <reagleBRKLN> his "postscript to the name of the rose" is very worthwhile

14:48:14 <DanC_RSW> I thought this sign/signifier/signified stuff was due to Eco. was he french?

14:48:26 <reagleBRKLN> Also, mentioned him on http://reagle.org/joseph/blog/culture/neo-blogs

14:48:31 <reagleBRKLN> no, eco is a nose picking italian

14:48:32 <nmg> italian, I believe

14:49:00 <reagleBRKLN> (i loved the book and move, and was a fan, but then was amused when he gave a talk at college park and he was on the cover of their paper picking his nose)

14:49:01 <DanC_RSW> B:folks here use "annotation" for what I'd call formalization. i.e. if you take a paper you wrote and formalize the dc:title, that's not an annotation, to me.

14:49:02 <dc_rdfig> Added comment B11.

14:50:19 * DanC_RSW wonders if nmg is travelling, and if so, what's his nearest airport

14:50:35 <nmg> not travelling - I'm in the room next door to you

14:50:54 <danbri_> lol

14:51:17 <DanC_RSW> er... what I meant by "travelling" is that your current foaf:nearestAirport isn't your usual foaf:nearestAirport

14:51:48 <danbri_> it's still contact:nearestAirport, we didn't steal that one yet :)

14:51:54 <DanC_RSW> ah

14:52:24 <nmg> given that my nearestAirport varies between SOU and Bristol on a weekly basis, I assume that I neve stop travelling ;)

14:52:24 * DanC_RSW feels just overwhelmed by all the knowledge being poured into and out of MS Powerpoint

14:53:13 <DanC_RSW> sbot, [ apt:iataCode "SUO" ] cyc:inRegion [ a cyc:Country; cyc:agentName ?WHERE ] ?

14:53:37 <DanC_RSW> .google SUO airport

14:53:39 <datum> SUO airport: http://cinema.multiplayer.it/bloopers/film/testo.php3?id_film=94&Lettera=A

14:53:45 <DanC_RSW> .google SUO airport code

14:53:46 <datum> SUO airport code: http://www.seniorsomething.com/apcS.html

14:53:47 <nmg> SOU - Southampton

14:53:59 <JibberJim> , England

14:54:03 <DanC_RSW> how far from SUO to BRS?

14:54:16 <nmg> about 50 miles

14:55:05 <DanC_RSW> ah; that's like LGW/LHR or the new york airports... sorta interchangeable for contact:nearestAirport purposes... hmm...

14:55:45 <nmg> and that's before you consider what you contact:mostConvenientAirport is

14:56:20 <sandro> nearestAirport is a perfect example of a hack. so elegant and so ugly at the same time. :-)

14:56:46 <DanC_RSW> What I really mean by [ ?WHO contact:nearestAirport ?APT] is that starting from an arbitrary place in the world, to visit ?WHO, flying to ?APT is likely to be nearly optimal.

14:57:03 <JibberJim> Very good nmg, it's 57 as crow flies!

14:57:21 <nmg> hmmm. there speaks a man who's never had to try and plan a journey through SOU airport...

14:57:53 <DanC_RSW> or: if two people have a common contact:nearestAirport and they want to get together, they shouldn't look at buying plane tickets.

14:57:57 * JibberJim really wants to do that, I've got route data for a variety of airlines in RDF, and people and their nearestAirport, I just need to get it all sorted out.

15:00:17 <sandro> Ah, that's more precise and focussed than I would have expected. I thought it was being used like I used to ask my Dad when were on long drives: "What's the nearest biggest city?"

15:00:17 * timbl-fl knows someone who flies a helicopter based between SOU and BRI - coul dbe just what you need

15:00:17 <nmg> danc: that's a more convincing example, but as sandro says, it's still a nasty hack

15:00:20 <sandro> I didn't call it "nasty". :-)

15:00:20 <sbp> it's good from a privacy point of view

15:00:20 <sandro> I think ontology design is all about making these kinds of tradeoffs, and I know I still have a lot to learn.

15:00:20 <timbl-fl> timbl-fl is now known as timbl-ma

15:00:20 <sbp> well, better than exact coordinates

15:00:20 <DanC_RSW> so perhaps contact:nearbyAirport would be better.

15:00:20 <DanC_RSW> or perhaps :airportWithinAnHoursDrive

15:00:20 <danbri_> yup, also contact:nearestAirport(X,Y) statements are prone to becoming false over time occasionally, due to new airport construction

15:00:20 <sandro> (although I think I'm very good at it.)

15:00:59 <DanC_RSW> .google TravelTools

15:01:07 <sandro> What's wrong with using spacial coordinates? Basically that they're hard for humans to work with, since they have so many bits?

15:01:23 * chaalsNCE notes that a friend who happens to travel a useful route is a good argument for proper privacy/auth management in this area.

15:01:29 <nmg> ideally we'de just be able to refer to a geographical gazetteer

15:01:29 * nmg must dig out his NIMA GNS dataset

15:01:31 <datum> TravelTools: http://www.traveltools.com/

15:01:37 <danbri_> we also have http://xmlns.com/foaf/0.1/#term_based_near

15:01:40 <sandro> If geocoding and geodecoding services were more handly it would help.

15:02:00 * chaalsNCE should read the spec - I just use nearestAirport as if it meant "the airport that strikes me as a good indicator of how to fly to roughly the same place from a long way away". Which is bound not to match what the social meaning spec might say I would mean, as far as I can tell...

15:02:09 <DanC_RSW> nothing's wrong with spacial coordinates. not sure I grok the question. most people can't answer "what's your lat/long to 3 significant digits" and even most aren't willing to in a public forum

15:03:13 <sandro> afraid someone might lob an icbm at them? :-) I suspect in 10 years your computer will nearly always know your spacials....

15:03:35 * JibberJim could do it to 10m accuracy, but I don't think that would be that useful, unless it could be decoded into nearestAirport/nearestTrainStation/nearestPlaceYouMightBe etc.

15:03:39 <chaalsNCE> because geocoding/geodecoding tools aren't handy, spatial coordinates are similar to directions in ancient icelandic for most people.

15:04:05 <DanC_RSW> hmm.. I think the common parlance is "airports that service this area include Fort Meyers". Usually followed by distance in driving-minutes or miles.

15:04:15 <chaalsNCE> If they are accurate they are bound to be useful, just as soon as you figure out what they mean.

15:04:20 <JibberJim> What's useful is how you can get to meet me, or knowing where I am likely to be see if it's worthwhile inviting me for a beer/to a meeting/etc.

15:04:34 * timbl-ma fears its not the icbm, its the radio controlled model plane with anthrax and gprs nowadays one fears

15:05:08 <DanC_RSW> yes, in 10 years I expect to have things inside my skin that know my lat/long and record everything I hear and see.

15:05:14 * chaalsNCE thinks timbl should stop promoting new techniques for terrorism in public...

15:05:58 * timbl-ma used to think that terrorists had been very unimaginative ... but on 9/11 took it all back

15:06:36 <DanC_RSW> I think 9/11 exceeded the perpetrator's expectations. I don't think anybody knew the buildings would collapse completely.

15:07:18 <DanC_RSW> talk starts "we have to deal with multiple ontologies in Semantic Web"

15:07:23 <nmg> nmg is now known as nmg_at_coffee

15:07:28 <DanC_RSW> hmm... do we?

15:07:43 <sandro> I think your timescale is off for having things really under your skin. I think that's more like 30 years out. Shall we take bets? :-)

15:07:48 <DanC_RSW> i.e. is that a motivating factor?

15:08:28 * DanC_RSW just read slashdot article about chip implants replacing daily pills, notes birth control patches and pacemakers, but won't take a bet on 10 vs 30 years

15:08:35 * sandro shudders at "in Semantic Web" with no definite article.

15:08:49 * JibberJim wouldn't even bet 30 due to the powering these things problem.

15:08:52 <DanC_RSW> .google wikipedia engrish

15:08:53 <datum> wikipedia engrish: http://www.wikipedia.org/wiki/Engrish

15:09:39 <sandro> Ah, not a native speaker? Okay. I thought it was a judgement call, reflecting a different conceptualization of what "semantic web" might mean.

15:09:52 <monkeyqi> and what if the implants had AI in them? you could be convicted for removing them

15:09:55 <DanC_RSW> power is a good point... battery technology is moving slower than cpu/net/disk. but something hearing-aid like seems not far away.

15:10:32 <mmealling> inductive power using a through skin transformer works fine for lots of embedded medical systems....

15:10:48 <mmealling> the patient wears it for a few hours each week....

15:10:51 * chaalsNCE didn't think 9/11 took masses of imagination, just careful planning and a lot of suicidal terrorists who could keep the plan together.

15:11:36 <DanC_RSW> I'm just struggling to get on this wavelength where pretty abstract claims are considered motivating. I need to eat and breathe. I like to have my computer do stuff for me. I don't feel "we need to deal with multiple ontologies" is beyond dispute at this point.

15:11:46 <monkeyqi> yeah, they were clearly not tom clancy fans

15:11:52 <mmealling> chaals, for most people the ability to think of using a plane as anything but a plane is hard... it requires abstract though and a large portion of the population isn't capable of that.

15:12:46 <mmealling> monkeyqi, more probably clearly not fans of reading in general.

15:12:53 <DanC_RSW> what's amazing is that the buildings were ensured for the case of *one* plane crashing into them, but not two. Insurance actuaries are evidently in that part of the population, mmealling.

15:13:43 * chaalsNCE thinks mmealling underestimates the population. They just use their said capacity for things like figuring out what the !@$%!#$!@#!! the UI engineer was thinking that day, or whether the odds of finding a landmine on taht road justify the likelihood of better money at the end of it.

15:13:52 <DanC_RSW> oops... flaw in my argument: insurance actuaries might have been covering accidental cases.

15:14:59 * chaalsNCE thinks insurance actuaries would be reasonably good at imagining terrorist attacks. Military actuaries are probably a lot better at it, of course.

15:16:13 <DanC_RSW>http://lists.w3.org/Archives/Public/www-tag/2003Oct/0027.html

15:16:13 <dc_rdfig> D: http://lists.w3.org/Archives/Public/www-tag/2003Oct/0027.html from DanC_RSW

15:16:24 <DanC_RSW> D:|Draft Finding on AbstractComponentRefs-37 issue

15:16:25 <dc_rdfig> Titled item D.

15:16:25 <chaalsNCE> And whole schools full of people get little badges to put on their shoulders or hide in their wallet if they were clever enough at it to impress the teachers, and appeared committed enough to be able to make it work.

15:16:31 <DanC_RSW> D:on today's TAG agenda

15:16:31 <dc_rdfig> Added comment D1.

15:17:39 <DanC_RSW> speaker explains some abstract "least upper bound (LUB)" concept using a restaurant example! Darn! I should have been paying more attention. They related this stuff to real life and I missed it!

15:18:50 <DanC_RSW> Hayes: seems dangerous to use LUB reasoning with negation. [cajun/afghani restaurant counter-example]

15:19:04 <DanC_RSW> ... non-mon

15:19:25 <DanC_RSW> ... do you take the LUB before or after you take the complement?

15:19:29 <DanC_RSW> A: before

15:19:29 <dc_rdfig> Added comment A3.

15:19:36 <DanC_RSW> Hayes: that's what's dangerous

15:19:37 <DanC_RSW> A3:""

15:19:37 <dc_rdfig> Deleted comment A3.

15:19:56 <DanC_RSW> Q[who?]: do you compute a normal form for your concepts?

15:20:22 <DanC_RSW> Aspkr: yes... we compute a normal form for the query. [.. herberand..?]

15:20:44 <DanC_RSW> J. Akahani, K. Hiramatsu, and T. Satoh "Approximate Query Reformulation for Integration" concludes

15:20:52 <DanC_RSW> [break]

15:33:04 <nmg_at_coffee> nmg_at_coffee is now known as nmg

15:37:44 * mmealling gives himself a headache by jumping into the whole REST vs SOAP issue

15:41:31 <DanC_RSW> [resume]

15:41:35 <DanC_RSW> B:M. Schorlemmer Y.Kalfoglou "On Semantic Interoperability and the Flow of Information "

15:41:35 <dc_rdfig> Added comment B12.

15:42:20 * mmealling slowly realizes that unless someone can quickly apply things like WS-Security, WS-Federation and Liberty he won't be able to make anything in this system RESTful....

15:46:45 <DanC_RSW> spkr refers to "logic of distributed systems" by Barwise/Seligman, which Ann W. has recommended to me in topicmap contexts.

15:46:58 <DanC_RSW> spkr calls it "Channel Theory"

15:47:12 * DanC_RSW looks up spkr...

15:47:34 <DanC_RSW> spkr = Schorlemmer

15:48:04 <DanC_RSW> (Frank vH distinguishes Schorlemmer among M. Schorlemmer Y.Kalfoglou)

15:49:06 <DanC_RSW> slide "channel theory": tokens vs. types

15:49:49 * nmg hadn't realised that Yannis and Marco had a paper in the other room

15:52:16 * DanC_RSW encourages ngm to chump the other workshop

15:52:42 <DanC_RSW> B:[Jan 2002 msg|http://lists.w3.org/Archives/Public/www-rdf-interest/2002Jan/0131.html] suggests IEEE SUO uses Barwise/Seligman logic

15:52:43 <dc_rdfig> Added comment B13.

15:53:20 <DanC_RSW> C:|THE SIGN, THE SIGNIFIER, AND THE SIGNIFIED

15:53:20 <dc_rdfig> Titled item C.

15:53:43 <DanC_RSW> C:"The sign, the signifier, and the signified are concepts of the school of thought known as structuralism, founded by Ferdinand de Saussure, a Swiss linguist, during lectures he gave between 1907 and 1911 at the University of Geneva."

15:53:44 <dc_rdfig> Added comment C3.

15:55:05 <DanC_RSW> B:slide notes SUO-IFF meta-level framework

15:55:05 <dc_rdfig> Added comment B14.

15:55:46 <DanC_RSW> gonna have to do the larch thing on barwise/seligman information flow stuff

15:59:20 <DanC_RSW> battery at 40%. I moved from my door-blocking spot at the break, but now I have no 110v.

16:00:06 <D[a]vey> D[a]vey is now known as Davey

16:00:29 <DanC_RSW> Menzel fidgets with mac/projector... "I've got a PC if you need it" -- Welty. [many laughs]

16:00:58 <DanC_RSW> Hayes encourages Menzel to press on...

16:04:02 <mmealling> ok, I give up for now. The lack of interface specification language for RESTful interactions makes it unusable for my app needs....

16:04:22 <mmealling> maybe SOAP 2.1 will fix that....

16:04:30 <MarkB_Zzz> MarkB_Zzz is now known as MarkB

16:04:46 <mmealling> or maybe Mark can correct me on all that....

16:05:09 <DanC_RSW> Menzel reviews motivation and history of KIF...

16:05:23 <MarkB> gotta run... but have you looked at RDF Forms? http://www.markbaker.ca/2003/05/RDF-Forms/

16:05:38 <DanC_RSW> you mean maybe WSDL v.next, I think/hope, mmealling.

16:05:59 <MarkB> RDF Forms = WSDL v.next

16:06:07 <MarkB> IMO

16:06:17 <MarkB> MarkB is now known as MarkB_lun

16:06:23 <MarkB_lun> MarkB_lun is now known as MB_lunch

16:07:48 * DanC_RSW wonders about making more time to noodle about WSDL and such

16:08:15 <deltab> is anything happening with WRDL?

16:08:16 <DanC_RSW> MarkB, would you be interested in a ScheduledTopicChat to explore "RDF Forms = WSDL v.next"?

16:08:21 <DanC_RSW> .google WRDL

16:08:22 <datum> WRDL: http://www.ashland.edu/~wrdl/

16:08:22 <dc_rdfig> Label WRDL not found.

16:08:29 <mmealling> Dan, Mark: yea, here's my problem: I'm involved with the Auto-ID standardization efforts. The goal is to build a service that can persist metadata about objects within the supply chain (this is a real world appliation that will mean big money to people very soon so its important to get it right).

16:08:33 <deltab> this: http://www.prescod.net/rest/wrdl/wrdl.html

16:09:37 <DanC_RSW> ah... so prescod's WRDL hasn't achieved googlemark yet

16:09:47 <mmealling> i.e. I need to be able to do things like: "for this EPC here is its PO#", "What are the hazmat characteristics of the product with this EPC", "Was this EPC recalled?", "When does this EPC expire", etc....

16:10:40 <mmealling> "what are all of the EPCs at this location x,y", "Tell subscribers that EPC 'x' has been seen and its temperature was above freezing for x minuts"

16:10:46 <DanC_RSW> it's clear to me how to map that into REST, mmealling, but I stipulate that it could be more automated, ala WSDL

16:11:48 <mmealling> sure, I could map all of those into REST. But we're talking thousands of potential methods that need to be concretely specified in a way that allows for the automatic generation of client side object classes.

16:13:08 <DanC_RSW> my intuition says 80% of the methods are straightforward RDF queries, 80% of the rest are SQL insert/update that map nicely to RDF, and the remaining .20*.20 need custom code.

16:13:48 <chaalsNCE> .google AA

16:13:48 <datum> AA: http://www.aa.com/

16:13:49 <dc_rdfig> Label AA not found.

16:13:55 <mmealling> the service will have about 10 basic methods, the most important being: what other methods do you support? I would prefer that method to have two response types: a unique URI that identifies it globally and a stream of WSDL-like interface specifications.

16:14:34 <r> r is now known as eikeon

16:14:36 * DanC_RSW down to 27% battery

16:15:25 <mmealling> i.e. if I connect to a service I can ask it if it supports the "uri:bla" method. That might be a namespace but somehow that doesn't feel right. This is more than an XML Namespace, its a method specificaiton which includes semantics.

16:17:10 <mmealling> and it all needs to work with the various ws-security, ws-federation, liberty, etc.. stuff since the security and trust models around this stuff are very complex.

16:17:13 <DanC_RSW> er... all the XML namespaces I know and use include semantics

16:18:03 <DanC_RSW> I can't promise that straightforward REST treatments play nicely with ws-security, ws-federation, liberty.

16:19:12 <mmealling> that's my problem... and why I don't think its possible for this version to be RESTful... the time frame is to short and the capabilities requirements are to significant.

16:19:32 <mmealling> I need the 1.0 spec for this to be done by the end of the year.

16:19:56 <DanC_RSW> how long do you want it to last?

16:20:22 <mmealling> a very long time.... this stuff is going to be the basis for the entire supply chain....

16:20:39 <DanC_RSW> I'd have to know a lot more about your application (and, well, about ws-security) to evaluate cost/benefits of REST vs. security needs.

16:20:57 <mmealling> security trumps any other feature....

16:21:12 <DanC_RSW> is 3 years a very long time?

16:21:18 <mmealling> more like 10

16:21:34 <mmealling> this industry moves very very slowly...

16:21:36 <DanC_RSW> are you familiar with kerberos?

16:22:14 <mmealling> yea, but these guys aren't.... (plus a lot of the ws-federation stuff inherited a lot from kerberos)

16:22:17 <DanC_RSW> does this industry have a trusted central party that can play the role of the keeper-of-the-keys ala kerberos?

16:22:26 <mmealling> not really....

16:23:03 <DanC_RSW> I agree that ws-fed looks a lot like Kerberos. If there isn't a social structure ala kerberos that can last 10 years, I don't see how ws-fed can meet your requirements.

16:23:33 <mmealling> the industry is to large... this app is for anyone that manufacturs, ships, or sells anything physical.

16:23:55 <DanC_RSW> B:Panel discussion "What are they smoking? Controversial issues in semantic integration" Panelists: Alon Halevy (chair), Pat Hayes, Len Seligman, Chris Welty

16:23:55 <dc_rdfig> Added comment B15.

16:24:14 <mmealling> the fact that its ws-fed (i.e. the protocol interactions) means the protocol level interactions are the same. Who the identify providers are can change.

16:24:42 <DanC_RSW> ah... so the keeper-of-the-keys situation isn't necessarily stable, but you think it's manageable.

16:24:54 <mmealling> in many cases the federations will be between retailers and suppliers....

16:25:17 <mmealling> yes.... it will be a competitive business so there will be several providers and they will migrate over time....

16:25:24 <DanC_RSW> if the keeper-of-the-keys situation is feasible, REST should work. HTTP digest authorization can work ala kerberos.

16:25:58 <MB_lunch> MB_lunch is now known as MarkB

16:26:23 * MarkB catches up

16:27:43 <mmealling> my problem is one of adoption rate. Must work still needs to be done to make REST consumable by the likes of Kraft Foods' IT department....

16:28:17 <MarkB> DanC, sure, a ScheduledTopicChat on RDF Forms/WSDL would be fun

16:30:47 <mmealling> how 'bout a more political/incremental approach: do you and Mark thing that the attempts to make SOAP (is it 1.2?) RESTful (or at least REST compatible) mean that I can safely specify SOAP now but make the RESTful features of future versions of SOAP mandatory to implement?

16:32:07 <mmealling> or is that the $10 million question that everyone wants to know the answer to?

16:32:13 <GabeW> mmealling: my guess would be that you'd create problems of expectation if you don't specify REST up front ...

16:32:30 * MarkB nods

16:32:54 <mmealling> but I can't make REST mandatory to implement now. The tools and specs just aren't there....

16:33:08 <GabeW> mmealling: I feel your pain

16:33:33 <mmealling> it has to be easy for these guys or else it falls flat on its face....

16:34:07 <MarkB> REST is more constraining than Web services. It's harder to do stuff, just like rubbing your tummy and chewing bubble gum is harder than just chewing bubble gum

16:34:08 <mmealling> but I would like to create a path that leads them to goodness and light....

16:35:15 <mmealling> mark, and if you're going to want to make REST meaningful you're going to have to figure out how to hide that fact or else developers will simply do what's easiest (i.e. what they're IDE tells them is easiest)

16:35:21 <GabeW> MarkB: so THATS what REST is about

16:35:26 <GabeW> (rubbing tummy and chewing)

16:35:58 <MarkB> mmealling. currently, the easy stuff is easy to most people, e.g. giving things URIs and using GET on those URIs

16:36:22 <MarkB> PUT naturally follows from that

16:36:26 <mmealling> mark, that means focusing effort on things that make it easy for developers to autogenerate code for talking to REST interfaces (both on the server and client side)

16:36:41 <MarkB> Knowing how to use POST is the really hard part, at least in my experience

16:37:02 <mmealling> mark, yea, but this stuff isn't easy when you're having to integrate into large ERP and business app systems...

16:37:04 <GabeW> MarkB: especially when SOAP specs suggest one way of using POST

16:37:47 <MarkB> I've never needed code generation for any RESTful work I've done. I could see some value to it though, but in a very different way than Web services use it

16:38:18 <MarkB> treat libwww as a generic stub

16:38:25 <GabeW> most "code generation" is to to tightly bound RPC

16:38:50 <mmealling> that's because you do this stuff, you know what HTTP is. This is an industry that still isn't even connected to the network. They use EDI via OSI streams sent over point to point VPNs.

16:38:55 <GabeW> right

16:39:01 <MarkB> yup

16:39:03 <GabeW> "file transport" - copying files around

16:39:14 <deltab> why are they changing?

16:40:17 <GabeW> deltab: one reason is that point-to-point integration costs aren't scalable (at least thats been what I've seen)

16:40:21 <MarkB> I think showing folks the value of GETtable URIs is the easiest way to enlightenment

16:40:27 <mmealling> detlab, we're tryign to bring them into the web model: adhoc, almost peer to peer nature of the web so that disintermediation can occur. Currently they're in the 'islands of information' model that had more in common with the pre-Internet days of AOL, Compuserve, Prodigy, etc.

16:41:10 <GabeW> MarkB: yes, but in the minds of many people, there's still a conceptual leap between GETtable URIs and application integration using the concept of resources

16:41:51 <mmealling> Mark, yes, but that is entirely hidden from them. They don't want to know about URIs or GETs, or even XML. They have an business system using SAP and Oracle and they want to write a SQL Stored Procedure that can simply "go get the Purchase Order # for this item and stick it here"....

16:42:03 <GabeW> mmealling says it better

16:43:29 <MarkB> understood ... I'm saying that if you or somebody could show them URIs/GET, they might "get it"

16:44:22 <GabeW> someone just needs to build and deploy something big and complicated (ie integration wise) using REST and write it up for "enteprise developers"...

16:44:52 <mmealling> why? the result is the same regardless of whether I use GET or SOAP or Gopher: some chunk of information on server x maintained by entity Alice is transported over the network to server y which is maintained by Bob.

16:45:05 * MarkB is doing that; KnowNow + Semantic Web for seismic data collection/distribution/integration

16:45:47 <MarkB> No, it's not the same from an integration POV, because you've got a persistent handle to the data, something you don't have with SOAP/getStockQuote

16:45:48 <GabeW> cool

16:46:16 * MarkB will writes about it when he can - probably not for many more months though 8-(

16:46:28 <GabeW> REST requires rearchitecting a layer beyond the "remote object invocation" layer

16:46:40 <GabeW> and thats a layer thats pretty baked in many enterprise architectures

16:47:02 <mmealling> The problem is that for now all of the implementations of this will be in very large data centers that already have very large ERP systems running the likes of SAP. I want to be RESTful so that the barrier to entry for the smaller guys is lower. But right now the barrier to entry for SOAP is lower than for REST.

16:47:14 <GabeW> yah, thats right

16:48:20 <MarkB> I don't know what to say. If that's the case, and you want to do REST, then you've got to sell it. All I can offer is the GET/URI suggestion; show 'em the money

16:48:39 <GabeW> rest needs a sales posse

16:48:50 <mmealling> definitely....

16:49:18 <mmealling> if I "show 'em the money" right now then SOAP/web services would win....

16:49:31 <MarkB> Our VP Sales is named "inevitability" 8-)

16:49:37 <GabeW> rotfl

16:49:57 <MarkB> he always gets the final say. bastard. 8-)

16:50:10 <GabeW> my attitude is to stick RESTful stuff in where its easy to do and doesn't raise any obvious political issues

16:50:58 <mmealling> the integration cost associated with not having features like auto code generation for standard languages just about trumps any lower level savings GET/URI might have....

16:51:12 <GabeW> XRI resolution is sorta RESTful.. emphasis on the GET/URI

16:51:47 <MarkB> ok, great, work with that mmealling.

16:52:17 <MarkB> I havent' done code-gen in about 10 years, but recall it being horrid

16:52:27 <GabeW> codegen--

16:52:39 <mmealling> no time.... the genesis of this discussion was whether or not the REST work was heading in that direction in a time frame that was useable for these customers.

16:53:16 <mmealling> I'll simply have to stick with SOAP/WSDL/etc and hope that I can retrofit REST into it over time....

16:53:36 <mmealling> or that the community does that for me.

16:53:40 <GabeW> mmealling: I would think you could specify the SOAP interactions in a way which adheres to some of the REST constraints

16:53:53 <MarkB> Hard to know for sure, but I'd say that *today*, the REST toolset is far more capable and mature than Web services

16:54:10 <mmealling> oh sure... even in SOAP 1.1 there are ways and bindings that allow you at least some RESTful features....

16:54:34 <GabeW> mmealling: beyond the GET/SOAP-Response MEP

16:55:40 <GabeW> i mean, you can "hide" REST in soap (if you have folks who insist on SOAP) with REST-like SOAP messages.. as a transition to real REST

16:56:07 <mmealling> mark, I think we're talking at two different layers. I'm talking about a toolset that allows near instantaneous integration into legacy apps.... Teaching SAP about REST is impossible.

16:56:56 <MarkB> I think we're talking at the same layer. I would surprised if developers at SAP didn't know how to configure a Web server and write a simple CGI/Servlet to answer "GET"

16:57:35 <GabeW> i think it woudl be an interesting excercise to refactor some of these enterprise integration protocols/soap messages as series of REST messages

16:58:12 * MarkB curses at state of cross-compilation tools on Linux/open-source-in-general

16:58:13 <GabeW> not sure it would result in the best REST architecture, but it *would* show the "completeness" of REST w/r/t to expressiveness of REST verbs (or disprove)

16:58:13 <mmealling> but has that integration work been done so that the users who already have SAP installations (and won't be ugprading for another 3 years) can access and use that remoe CGI?

16:58:55 * GabeW goes off to phone call

16:59:34 <mmealling> right now the most compelling thing about web services for these guys is that they can run a script that creates C++/Java classes that hide the fact that a given piece of information is on a remote server.

17:00:24 <mmealling> (add Cobol and other legacy languages to that list)

17:01:02 <MarkB> a URI hides that too

17:01:15 <mmealling> nope....

17:01:18 <MarkB> but sure, I see your point ... integration is two-way

17:01:33 <MarkB> no?

17:01:47 <mmealling> a URI requires your app to know about URIs, and the web, and networks, and things like asynchronous I/O, etc...

17:02:46 <MarkB> I wouldn't say that. Consider java. Have a URI, use java.net.HttpUrlConnection.

17:02:48 <mmealling> these apps need to be network enabled without knowing they've been connected to a network... they will treat this system as simply another database.

17:03:43 <GabeW> it seems to me that you could model any exchange of SOAP messages as a series of REST interactions (even with the same API presented to the application developer) - the tools would change.

17:04:17 <mmealling> Gabe, sure, but those tools don't exist yet....

17:04:23 <GabeW> yes

17:04:26 <GabeW> correct

17:04:29 <GabeW> i agree tools do need to exist

17:04:38 <mmealling> I need them to exist before I can require their use...

17:04:39 <GabeW> as silly as we may think it is

17:04:52 <GabeW> because its a conceptual mapping

17:05:02 <MarkB> the proper libraries can hide much of the network from a developer. I don't think what java.net hides is too different than what a stub hides; a little sure, but not too much

17:05:05 <mmealling> and these guys don't do things conceptually....

17:05:28 <GabeW> they do things in the concept of objects (these days)

17:05:41 <mmealling> I'm having enough trouble getting them to even understand the concept of a network.....

17:06:04 <GabeW> "its all objects"

17:06:21 <mmealling> and don't get me started on the idea of a protocol not having to specify every single query it might ever be capable of handling....

17:06:21 <MarkB> "a resource is any thing ..."

17:07:36 <mmealling> these guys would have required that HTML's IMG tag specify whether or not it was valid for the image to contain pictures of people or not.

17:08:04 <MarkB> ah, like rdf:datatype

17:08:07 * MarkB ducks

17:10:26 <GabeW> "its all objects and you access objects via methods specific to the objects' class".. people with this mentality think of REST as giving you only four methods..

17:11:23 <MarkB> btw, mmealling... ws-security isn't so bad from a REST POV

17:12:10 <MarkB> in general, the higher up the WS stack you go, the more reasonable the specs since they tend to deal in data

17:13:05 <GabeW> MarkB: What sort of layering do you see in WS-*?

17:13:53 <MarkB> what do you mean?

17:14:11 <GabeW> you said "higher up the WS stack"...

17:14:21 <GabeW> maybe I misread

17:14:23 <deltab> WS-Federation builds on WS-Security, -Policy and -Trust

17:14:27 <deltab> like that?

17:14:44 <GabeW> well, thats my question - if thats what MarkB was referring to

17:14:53 <deltab> -- http://www-106.ibm.com/developerworks/webservices/library/ws-fed/#twomodel

17:17:02 <MarkB> I just mean that the higher up you go, the greater the chance of avoiding the "specific interface" assumption since it's all just data

17:17:12 <MarkB> that increases the probability of it being REST friendly

17:17:28 <GabeW> got it

17:18:41 <mmealling> mark, and that's my question: is the direction of web services within the W3C such that I can assume that the process will ensure that SOAP v.next will be RESTful?

17:22:03 <MarkB> SOAP 1.2 is RESTful, as is 1.1 too (but not 1.0). But as with anything, it can be used non-RESTfully

17:22:53 <MarkB> I'm not sure what the W3C is going to have to say about this. I'm hoping the TAG does via webarch, but it may very well not. See the "interaction" section in webarch, for example

17:23:16 <MarkB> very "lame duck", IMO

17:30:49 <deltab> ^not be stuck

17:31:34 <deltab> oops

17:34:51 <danbri_> crap. pasted a URI into mutt/vi when it wasn't in 'insert' mode, and lost the reply to a msg of DanC's I was working on.

17:35:30 <sbp> hence DontModeMe

17:35:52 * danbri_ enters GrabbingSomeFood mode

17:36:01 <sbp> :-)

17:36:50 <Schuyler> danbri_: your trip back all right?

17:38:50 <danbri_> yes, just lengthy!

17:39:45 <danbri_> was good to meet you finally. in fact generally it was a great trip for putting names and faces together :)

17:58:07 <mdupont> mdupont is now known as md-afk

17:58:51 <mmealling> sorry, had to stop the group I've been talking about from defining data types as use case specific (i.e. <kidimage ref="foo://bar">, <dogimage ref="foo://baz"> instead of <img src="foo://bif"> and then something available via "foo://bif" that says whether the image contains a kid or a dog)

18:14:43 <Davey> Davey is now known as D[a]vey

18:26:41 <D[a]vey> D[a]vey is now known as Davey

18:33:21 <timbl-ma> DanC, latest check-in of cwm has @keywords function added and a couple of simple tests for it.

18:46:05 <eikeon_> eikeon_ is now known as eikeon

18:47:25 <danbri_> is there an OWL 2 UML convertor? or other diagramming tool that makes use of OWL?

18:50:56 <dajobe> see http://esw.w3.org/topic/UML_20and_20RDF

18:53:40 <danbri_> thanks

18:53:47 * dajobe tried the chat search

18:53:56 <dajobe> seems I should default to match words & be case independent

18:54:48 <DanC_RSW> ok, found 110v

18:55:15 <DanC_RSW> ah... much better. Whacked the volleyball around a bit. I had exceeded my threshold for being this close to a beach without playing volleyball.

18:55:17 <timbl-ma> eikeon?

18:55:41 * eikeon waves

18:55:47 <DanC_RSW> note to self: bring extra t-shirt tomorrow

18:55:54 * timbl-ma frost last night in MA

18:56:15 * eikeon reminded never picked squash yet

18:56:31 <timbl-ma> Eikeon, did you do anything special to prevent python expatreader crashing when it found an XML namespace?

18:57:25 <eikeon> I put a workaround in somewhere... let me find where.

18:57:39 <timbl-ma> There seems to be a clear bug in expaeader.py wwith pytrhon2.2 and the latest expat - and rdflib doesn't crasgh though

19:02:11 <eikeon> cvs history does not go back far enough: http://rdflib.net/cvs/rdflib/rdflib/syntax/parser.py?rev=1.34&content-type=text/vnd.viewcvs-markup

19:02:33 <eikeon> ... but the workaround is in the above

19:02:45 <eikeon> [[[

19:02:48 <eikeon> def parse(self, source):

19:02:48 <eikeon> parser = make_parser()

19:02:48 <eikeon> # Workaround for bug in expatreader.py. Needed when

19:02:48 <eikeon> # expatreader is trying to guess a prefix.

19:02:48 <eikeon> parser.start_namespace_decl("xml", "http://www.w3.org/XML/1998/namespace")

19:02:53 <eikeon> parser.setFeature(handler.feature_namespaces, 1)

19:02:54 <timbl-ma> Ok, I will send a bug report to the fink people. Maybe it gets cl;eared up in python 23

19:02:55 <eikeon> parser.setContentHandler(XMLRDFHandler(self.add))

19:02:57 <eikeon> parser.setErrorHandler(ErrorHandler())

19:02:59 <eikeon> parser.parse(source)

19:03:01 <eikeon> ]]]

19:03:35 <timbl-ma> Ah.

19:03:38 <timbl-ma> Thank you.

19:04:05 <timbl-ma> BTW, is there a way I can get namespace declarations though rdflib?

19:04:18 <eikeon> I fired off a bug report to PyXML... but no joy.

19:04:38 <DanC_RSW> no joy in PyXML bug system? ouch!

19:04:42 <eikeon> The prefix bindings?

19:04:49 * DanC_RSW wonders about escalation paths

19:05:45 <timbl-ma> yes

19:05:52 <timbl-ma> cwm accumulates them to resue on output

19:05:58 <timbl-ma> reuse

19:06:13 <Davey> Davey is now known as D[a]vey

19:06:33 <eikeon> Can set them in rdflib... but rdflib does not automatically try and reuse them.

19:06:55 <timbl-ma> All cwm needs it to be notified when the parser finds one

19:07:00 <timbl-ma> (a new one)

19:07:21 <timbl-ma> (although replication i snot a problem excep[t a watse of time)

19:08:35 <eikeon> Hum... looks like the map is not all that exposed. There is a method for setting:

19:08:38 <eikeon> def prefix_mapping(self, prefix, namespace):

19:08:38 <eikeon> self.__namespace_prefix_map[namespace] = prefix

19:10:37 * eikeon wonders if question is a feature request?

19:18:58 <ndw_> ndw_ is now known as ndw

19:19:14 <DanC_RSW> hallway conversation: "... RDF, with the triples, subject object..."

19:52:34 <ndw> newline in field name `1.21'

19:52:34 <ndw> idea how to fix this:

19:52:36 <ndw> lol

19:52:42 <ndw> part of that was /msg'd, I swear

19:56:52 <dajobe> hi ndw, reading your laptop woes

19:57:17 <ndw> i think things are fixed mostly. but horribly painful.

19:57:27 <dajobe> it read that way

20:00:54 <DanC_RSW> danbri, I continued the discussion of domain names, currencies, and armies with jjc...

20:01:06 <DanC_RSW> ... he pointed out that "a language is a dialect with an army and a navy:"

20:01:24 <DanC_RSW> he says that's the best definition in the literature

20:02:30 <DanC_RSW> hi bijan... have you ever heard "a language is a dialect with an army and a navy"

20:11:59 * chaalsNCE notes that some languages just have a terrorist group - apparently those aren't armies if they're not fighting for the right kind of self-determination

20:13:48 <D[a]vey> D[a]vey is now known as Davey

20:16:14 * mmealling reads the WSDL 1.1 spec and tries to understand why the HTTP GET & POST bindings aren't sufficiently RESTful (or are they?)

20:18:21 <GabeW> mmealling: "WSDL includes a binding for HTTP 1.1's GET and POST verbs in order to describe the interaction between a Web Browser and a web site. "

20:18:27 <GabeW> is that what you are talking about?

20:23:41 <MarkB> they're not very RESTful, because it forces a distinction between a wsdl operation and an http verb

20:24:43 <GabeW> iow, it creates verbs on top of HTTP verbs

20:25:16 <MarkB> pretty much, but *occasionally* you might want to do that. My objection is that it forces this distinction.

20:25:45 <MarkB> (*very* occasionally)

20:28:08 <Davey> Davey is now known as D[a]vey

20:32:30 <GabeW> so WRDL instead of WSDL..

20:40:54 <MarkB> or WIDL

20:41:01 <MarkB> or RDF Forms

20:41:11 <MarkB> (others?)

20:44:18 * aharth is at iswc, anybody else?

20:47:17 <ndw> are you alone? :-)

20:47:36 <_joshua> _joshua is now known as VanessaCarlton

20:47:48 <VanessaCarlton> VanessaCarlton is now known as _joshua

20:48:02 * timbl-ma has noticed a few peole on the channel are at iswc

20:48:55 * ndw was making a bad joke

20:49:34 <aharth> ndw: no, actually a lot more people than i expected :)

20:51:21 <aharth> starting to get interesting only slowly though, mainly tutorials today

21:15:26 <GabeW> MarkB: does rdf-forms describe how to construct URIs or how to construct the (content of a POST request | query part of a GET request)?

21:20:46 <reto> I've made a draft of a schema for rendering RDF resource to some text format, there is also a demo implementation in java which you can try out at: http://test.wymiwyg.org/rdfstyles/

21:20:51 <jeremiah_> jeremiah_ is now known as jeremiah

21:21:07 <mortenf> looks interesting reto, will have a close look later

21:22:15 <sandro> wymiwyg is a cute name. :)

21:22:22 <reto> thanks!

21:23:44 <reto> not starting from xslt approach I could avoid RDF-path problems (at least till now)

21:57:16 <Schuyler> reto: rdfpath problems? (curious to hear more)

22:01:36 <sbp> the problem being: what RDFPath language do you use?

22:01:38 <sbp> I assume

22:01:54 <sbp> you can't use XPath very well unless you have a canonical subset of RDF/XML that you're dealing with

22:02:02 <sbp> otherwise you end up writing an RDF/XML parser in XSLT

22:02:22 <sbp> I should do that

22:02:28 <sbp> but I think DanC's already done it...

22:02:54 <dajobe> it's been done at least 3 times

22:04:22 <sbp> yeah. just found Jason Diamond's one: http://lists.w3.org/Archives/Public/www-rdf-interest/2000Sep/0097 (doesn't look anywhere near complete, though)

22:04:51 <dajobe> like his repat; old and hasn't been completed or updated

22:05:20 <sbp> perhaps I'll write one for the REC

22:05:35 <dajobe> I wouldn't encourage yet another one; there are both more complete and more modern ones. It's just a make-work job, IMHO.

22:05:48 <sbp> okay. ta!

22:06:14 <dajobe> did you notice rdf:RDF was made optional?

22:06:16 <sbp> dajobe: saving me from my own stupid ideas since 2003

22:06:20 <sbp> yeah. already implemented it

22:06:24 <dajobe> ah, great!

22:06:27 <sbp> only if there's a single nodeElt, no?

22:06:31 <dajobe> made (more) optional that is

22:06:32 <dajobe> yes

22:06:38 <sbp> good, good

22:06:39 <dajobe> I wanted to make it pretty precise

22:06:45 <sbp> yeah. it's a good call

22:26:41 * mmealling missed the end of the conversation with gabe and Mark WRT to WSDL and WRDL.

22:26:47 <mmealling> gabe, you still there?

22:52:53 <GabeW> i am here

22:52:55 <GabeW> now

22:53:26 <GabeW> it didn't really go anywhere besides mentioning WIDL and rdf forms

22:57:02 <GabeW> mmealling: does that answer your question?

22:58:00 <mmealling> yea, saw that, but I"m still trying to understand Marks assertion that WSDL 1.1's HTTP GET binding "forces a distinction betweena wsdl operation and an http verb"a wsdl operation and an http verb

23:00:37 <GabeW> we didn't go much further there

23:01:08 <GabeW> i think the idea is that wsdl forces you to declare an operation beyond "GET", "POST", etc

23:01:11 <mmealling> yep... so I'm bouncing around google looking up various articles from folks concerning SOAP, WSDL and REST... brain breaking stuff

23:02:00 <mmealling> 'beyond'? I was thinking of it as more of a subclassing....

23:02:10 <GabeW> i guess it depends on your POV

23:02:15 <GabeW> i see it as "on top of" maybe

23:02:21 <mmealling> i.e. there is GET(); then there is GETPurchaseOrder()

23:02:40 <GabeW> well, could be DOJumpingJacks

23:02:43 <GabeW> ()

23:03:12 <mmealling> you wouldn't use GET though since that has side effects (i.e. makin someone do jumping jacks)

23:03:27 <GabeW> heh

23:03:35 <GabeW> fair enough

23:03:38 <GabeW> well

23:03:47 * GabeW is on phone now

23:08:40 <libby>http://swordfish.rdfweb.org/discovery/2001/04/rdfig/

23:08:41 <dc_rdfig> E: http://swordfish.rdfweb.org/discovery/2001/04/rdfig/ from libby

23:08:55 <libby> E:|chump search, now with url search too

23:08:55 <dc_rdfig> Titled item E.

23:09:12 <libby> E:danbri suggests make an annotea server out of it

23:09:12 <dc_rdfig> Added comment E1.

23:09:46 <libby> E:currently just html output though

23:09:46 <dc_rdfig> Added comment E2.

23:09:49 <libby>http://swordfish.rdfweb.org/discovery/2002/02/paths/

23:09:50 <dc_rdfig> F: http://swordfish.rdfweb.org/discovery/2002/02/paths/ from libby

23:09:50 <danbri> E:At some point, ericm was tapping off XML data from chump and making into annotea data, not sure what became of that.

23:09:50 <dc_rdfig> Added comment E3.

23:10:13 <libby> F:|foaf codepiction paths demo resurrected

23:10:13 <dc_rdfig> Titled item F.

23:10:48 <libby> F:now with name-based search as well as by email/mbox_sha1sum

23:10:48 <dc_rdfig> Added comment F1.

23:11:27 <libby> F:updated once a day from the [http://swordfish.rdfweb.org/discovery/2001/08/codepict/|codepiction database]

23:11:27 <dc_rdfig> Added comment F2.

23:12:31 <libby> F:code by Damian Steer (shellac): [http://rdfweb.org/people/damian/2002/02/foafnation/|explanation of how he did it]

23:12:32 <dc_rdfig> Added comment F3.

23:13:23 <JibberJim> How much do you think it would take to make it so it searched on any IFP? as simplisitic looking at my Smushing process suggests that a significant number of Agents in the foaf world only have a homepage/weblog and not an email

23:13:56 <JibberJim> I've been looking at changing all the foafnaut queries to work like that.

23:14:03 <libby> F:I also added in a few famous people from [http://rdfweb.org/2002/01/photo/|danbri's codepiction writeup] to make it more interesting

23:14:03 <dc_rdfig> Added comment F4.

23:14:22 <libby> eah, makes sense jim

23:14:41 <libby> I'd have to have a bit of a think, but seems feasible

23:14:53 <JibberJim> I don't know if that's so relevant to codepiction though, as there are quite a few groups and things now.

23:16:00 <libby> it is, e.g. email addresses for sinatra,chet baker etc

23:16:06 <libby> is bad

23:16:50 <sbp> sbp is now known as esoteric

23:16:57 <esoteric> esoteric is now known as sbp

23:23:04 <danbri> http://ilrt.org/discovery/chatlogs/foaf/latest/ [[

23:23:04 <dc_rdfig> G: http://ilrt.org/discovery/chatlogs/foaf/latest/ from danbri

23:23:05 <danbri> Your software has been discovered sending too many requests to the chatlogs area ignoring the robots.txt entry for /discovery/chatlogs that forbids this. This page will be returned by all such requests from now on.]]

23:23:07 <danbri> boo!

23:23:21 <danbri> oops meant for #foaf, but still, boo! :)

23:24:04 <libby> uhoh

23:24:15 <libby> is that all of us here?

23:24:22 <danbri> same ip yeah

23:24:33 <libby> dajobe will remove us

23:24:39 <JibberJim> You only got one IP?

23:24:48 <sbp> mask?

23:24:54 <danbri> shared cable modem

23:25:15 * JibberJim shares ADSL with himself, but still has 8...

23:28:47 * chaalsNCE wonders about moving into Jim's house instead of Zetland

23:29:09 <libby> yes, but we have heating :)

23:29:11 <JibberJim> I have no heating at all chaalsNCE, and it's freezing outside.

23:29:24 <libby> poor jibberjim :(

23:29:49 <JibberJim> The laptop keeps me warm

23:29:56 * chaalsNCE has a Mac - what more heating can you ask for?

23:31:09 <chaalsNCE> E:

23:31:09 <dc_rdfig>http://swordfish.rdfweb.org/discovery/2001/04/rdfig/

23:31:10 <dc_rdfig> chump search, now with url search too

23:31:11 <dc_rdfig> (1:libby) danbri suggests make an annotea server out of it

23:31:12 <dc_rdfig> (2:libby) currently just html output though

23:31:13 <dc_rdfig> (3:danbri) At some point, ericm was tapping off XML data from chump and making into annotea data, not sure what became of that.

23:31:38 <libby> yeah, not what you want chaals, but sorta getting there, a bit

23:32:05 <JibberJim> sshhh libby, wait until he says it's not what he wants, you might just have talked yourself out of the beers

23:32:24 <chaalsNCE> E: [mailto:charles@w3.org|I] have a half-baked bot that can pump URI's and comments to an annotea server, and fetch them back, that was done as a student project (written in C I think)

23:32:24 <dc_rdfig> Added comment E4.

23:32:41 <libby> heh

23:32:54 <JibberJim> I've got one of those written in javascript too, but it's broke at the moment.

23:33:02 * libby will demand birthday beer from chaals anyway

23:33:14 <chaalsNCE> Lib, all you need is a bot that looks for something being chumped, queries your server, and chumps back the last reference...

23:33:32 * chaalsNCE already said he would make you a birthday pavlova. You don't get much better than that...

23:33:54 <libby> indeed, and you are most kind

23:39:03 <JibberJim> hmm the semantic web doesn't seem to know when Libby's b'day is

23:41:48 <libby> it's thursday, when I have an all-day swad-europe meeting, excitingly.

23:42:00 <libby> shellac's is tomorrow. well today really

23:42:10 <JibberJim> What a day to turn 40

23:42:17 * JibberJim ducks :-)

23:42:18 <libby> heheh

23:42:22 <libby> git

23:59:10 <Davey> Davey is now known as D[a]vey