Semantic Web Interest Group IRC Chat Logs for 2003-10-09

00:06:19 <sandro> I wonder when to close the world.... My owl test-results generator does a pass where it adds a lot of closed-world properties on things, for a test, how many systems have passed it. Right now it just adds that property knowing it's not going to leak out again, but..... hrm.

00:23:27 * bijan notes that it's often clearer in deductive databases, for some reason

00:23:48 <bijan> I guess because they so regularly do it, the have the tricks and tradeoffs worked out

00:24:41 * bijan would love to see a mad panel with deductive database, semi-structure data, and kr semantic webists going at each other

00:26:04 <bijan> Right now, via the ascendence of OWL and RDF semantics, the KR folks are in the lead

00:26:06 <bijan> In a big way

00:26:21 <bijan> Even given the sharp battle between the DL and the RDF camps

00:26:59 <bijan> If the KRists win on rules (by, perhaps, crushing RuleML) then I think the die will have been cast

00:27:19 <bijan> Hmm. Points to a weakness, perhpas, in the current efforts

00:27:42 <bijan> Presumably, the webby thing to do would be to encapsulate or unify all these things in some way

00:27:56 <bijan> Behind hypertext (or some relevant analogue)

00:28:19 <bijan> Be a hub and integrative, interstial membrane, rather than the meat

00:30:28 <sandro> "If the KRists win on rules (by, perhaps, crushing RuleML)" the JC was very very strange yesterday.

00:30:40 <bijan> heh

00:30:50 <bijan> I forgot that you're all way more plugged into all that

00:30:57 <bijan> I'm just going from DAML PI rumors

00:31:08 <bijan> And the IEEE article

00:31:36 <sandro> We have two proposals on the table, on from the RuleML side one from Peter/Ian. But there doesn't seem to be any process for merging, deciding between the two, or anything. It's surreal. (or maybe going on behind doors I dont know about.)

00:31:58 <sandro> IEEE article?

00:32:19 <bijan> No, my impression is that they're largely disjoint :)

00:32:21 <bijan> current issue

00:32:23 <Davey> Davey is now known as D[a]vey

00:32:31 <bijan> stefan stabb's "slot"

00:32:37 <bijan> 6 views on rules

00:32:57 <bijan> Ian, michael kifer, ben grosof, stefan decker, and two other people

00:33:12 <sandro> well, we've managed to get the two proposals to look the same at the right distance -- they're both just positive horn rules with URIs for symbols,.... or maybe datalog -- a little fuzzy there.

00:33:24 <sandro> it's not online is it?

00:33:34 <bijan> "DL compatible FOL", "F-Logic", "RuleML/Situatied courteous logic", "Triple"...

00:33:41 <bijan> No idea

00:34:01 <bijan> I have hard copy becaue 1) jim has a slot thus copies, and 2) i've coauthored a task puting article in it

00:34:07 <sandro> yeah, those sound like the sides. :)

00:34:35 <bijan> most of the bits are funny

00:34:39 <bijan> They tend to start off with:

00:34:48 <bijan> "What can we derived about rules from first principles?"

00:34:57 <bijan> "1) Web rules must be declarative"

00:35:07 <bijan> 2) My particular view is obviously correct"

00:35:22 <bijan> s/view/view, software, current research/

00:35:22 <sandro> Decker seems to have quietly dropped out of the discussion, but maybe he's just busy elsewhere for a bit. But I couldnt blame him for taking his marbles and going home.

00:35:36 <bijan> I thnk it's the galloway move

00:35:44 <bijan> With his other stuff

00:35:46 <bijan> Esp. ISWC

00:35:47 <sandro> yeah, sounds about right.

00:35:53 <bijan> he's become quiet everywhere

00:36:03 <bijan> And stefan is a bit of a trouper, so I doubt he's marble hording :()

00:36:04 <bijan> :)

00:36:07 <sandro> He's going to Ireland? Good for him.

00:36:13 <bijan> I hope it is :)

00:36:32 <bijan> He is, last we spoke, rueing the weather trade

00:36:34 <sandro> I know he was none-too-fond of California.

00:36:42 <bijan> Other way round, actually

00:36:45 <bijan> AFAICT

00:37:06 <sandro> Okay, maybe this will make him appreciate CA more when he goes back. :-)

00:37:20 <bijan> Or maybe it's "All the charm of CA, all the weather of, well, ireland, and less cosmopolitan than idaho"

00:37:33 <sandro> Although I think his main complaints were politcal/social, and after yesterday, ... who knows.

00:37:42 <bijan> Well, yes

00:37:45 <bijan> let's not go there

00:37:51 <sandro> Have you ever been to Idaho?

00:37:52 <bijan> e.g., CA :)

00:37:55 <bijan> Nope

00:37:59 <bijan> It was a slander

00:38:24 <bijan> I totally reserve the right to completely slander places i've never beena nd have no intention of going to

00:38:33 <bijan> Outrageously

00:38:47 <sandro> Yeah. My father was born there, but I've never been. My wife's from Iowa, which is pretty close, right?

00:38:55 <bijan> No!

00:39:00 <bijan> iowa is totally cool and hip

00:39:04 <bijan> Worldly even.

00:39:11 <bijan> HEy, you're wife's from Iowa

00:39:12 <sandro> Sioux city?

00:39:20 <sandro> Siouxland!

00:39:27 <bijan> There's an *excellent* Dar williams song called Iowa

00:39:33 <bijan> One of my favorite songs, actually

00:39:35 <sandro> lol

00:39:44 <bijan> I had to change all my iowa jokes and scorn after hearing this song

00:39:56 <bijan> (to idaho)

00:40:02 <bijan> (or indiana)

00:40:15 <sandro> I haven't heard any new music outside of MusicTogether ans Suzuki since Gregorian was born. (practically.)

00:40:38 * bijan will see the new sondhiem musical in Nov

00:41:02 <sandro> but man can I play a mean Twinkle Twinkle on 12 different instruments. :-)

00:41:09 <bijan> heh

00:41:30 <bijan> when I was a babe, that was, I'm told, the sure way to put me to sleep

00:41:36 <bijan> Hmm.

00:41:39 <sandro> and now?

00:41:49 <bijan> When I was a bab*y*, lest anyone doubt my attractiveness

00:42:04 <sandro> lol

00:42:04 <bijan> Doesn't work

00:42:16 <bijan> Would that it did during all my years of extensive insomnia

00:42:23 <sandro> I have a babe (er.... baby) sitting in my lap right now.....

00:42:29 <bijan> Would that it had...

00:42:31 <bijan> heh

00:42:46 <bijan> Your baby was a babe, as I recall :)

00:43:07 <sandro> she is, she is. I'd put pics on the web if I could figure out which URIs to use. :-]

00:43:27 <bijan> That's an interesting result of the "cool uris" dogma

00:43:33 <bijan> People get afraid to put stuff on the web

00:43:36 <bijan> I know i do

00:43:39 <bijan> or to update stuff

00:43:50 <sandro> Why afraid to update stuff?

00:44:09 <bijan> Well, if the uri's weird it can be hard to "fix" information structure

00:44:20 <bijan> The site gets more tedious to deal with

00:44:28 <sandro> Have I mentioned my introhash idea? That's my best bet for pictures, but it's like protocol5, which is way too complex for sw-meaning right now.

00:45:18 <sandro> but maybe with the right working code it'll be obvious how it works.

00:45:35 <bijan> uh...

00:45:38 <bijan> prolly not :)

00:45:45 <sandro> yeah.

00:45:54 <bijan> Hmm. If we get our http headers right, with the semistatic test page update itself?

00:46:56 <sandro> nah. for now, I have to type "make", look at the page, and the commit it. the http headers only affect how fast the generation runs.

00:47:02 <bijan> Aha

00:47:08 <bijan> Oh, I see you updated

00:47:10 <bijan> Grand

00:47:12 <bijan> Thanks :)

00:47:43 <sandro> Yeah, I changed the format a bit today, in response to lots of requests from DL folk who want to get 100% without doing Full stuff.

00:47:50 <bijan> yeah

00:47:53 <bijan> Ooo!

00:47:55 <bijan> i didn't notice that yet

00:47:57 <bijan> I see

00:48:27 <sandro> I'm not thrilled with having 40 different tables, but... nothing better has come to mind yet.

00:48:38 <bijan> yep

00:48:50 <bijan> Well, I'm pleased with the pellet results

00:48:57 <bijan> Oooh, there's Cerebra now!

00:49:05 * bijan note sthat pellet is kicking *serious* butt

00:50:01 <sandro> Yeah -- that was the real kicker for me -- Network Inference said "look, we can pass the two that no one else can, but ... for the others, well, change your format around and we'll think about it." And it was a reasonable request anyway.

00:50:12 <bijan> Yep

00:50:14 <bijan> Excellent

00:50:23 <bijan> Scotched the whole HP thing

00:50:30 <sandro> hm?

00:50:42 <bijan> Reraising the objection

00:51:04 <bijan> i mean, i kinda saw it being beat down, but having some passes sorta clinches it in a nice way

00:51:35 <sandro> yeah -- there are a ton of tests passed by 5+ systems now!

00:52:31 <bijan> Though, i'll say, the tests are in, one sense, somewhat silly

00:52:44 <bijan> I'm unclear that they show much *either* way

00:52:49 <sandro> Surnia's looking pretty pathetic. :-) But that's only like my 8th priority; mostly I needed to bootstrap this process.

00:53:16 <bijan> So, i'm very skeptical about some of the fowl and Euler results

00:53:34 <bijan> Inconsistency especially

00:54:01 <bijan> And even if they passed them all, I'm unclear what that would exactly show. it's tricky to interpret

00:54:13 <bijan> OTOH, they do provide a useful initial set to get started with

00:54:28 <bijan> There's a lot worse you can do than start by trying to pass (and udnerstand) most of the tests

00:54:29 <sandro> Yeah -- I'd like to see some other kinds of demos/tests. This is kind of like benchmarks-- not even close to a substitute for trying the systems yourself.

00:55:18 <bijan> yeah, espeically hard to predict how you'll do with realtively arbitrary kbs

00:55:18 <sandro> Right. Without the tests, given how complex OWL is, darn near anything could be advertized as an OWL implementation. Now it's a lot more clear what that might mean.

00:55:59 <sandro> And relatively arbitrary kbs can be SOOOOO arbitrary. The range of kinds of data people want to put in RDF is insane.

00:56:09 <bijan> yep

00:56:31 <bijan> Personally, the next "goal" for me is Galen

00:56:42 <bijan> We can't yet even check galen

00:56:53 <sandro> I still feel like we're on the web before the first browser. It's all so hypothetical, even if some stuff works.

00:56:54 <bijan> Though we finally managed to absorb the entire tbox

00:57:11 <sandro> What is Galen? (aside from a horse I know)

00:57:31 <bijan> The medical ontology that fact (and the optimizations) was developed for

00:57:35 <bijan> 2000 some classes

00:57:39 <bijan> A bajillion gcis

00:57:44 <bijan> SHIF, i think

00:57:58 <sandro> and Pellet is supposed to handle this?

00:58:01 <bijan> Fairly tricky nut to crack

00:58:08 <bijan> Well, it's a good test of a reasoner

00:58:22 <bijan> Forces you to engineer things well

00:58:33 <bijan> And get a certain core set of optimizations together

00:58:59 <bijan> i think the record for classifying it is like <10 seconds

00:59:06 <bijan> which is pretty amazing, when you think about it

00:59:11 <sandro> Forgive my naive question but what are you going to ask Pellet to do with it?

00:59:22 <bijan> Classify it?

00:59:54 <sandro> what does a user do with it?

01:00:09 <bijan> The med folks use it for something

01:00:14 <bijan> I mean, it's a live ontology

01:00:19 <bijan> anotomical stuff, i believe

01:01:29 <bijan> See: http://lists.w3.org/Archives/Public/www-webont-wg/2003Jul/0303.html

01:02:28 <bijan> We can do ka.owl

01:04:09 <bijan> Sandro: http://www.opengalen.org/

01:04:11 <bijan>http://www.opengalen.org/

01:04:11 <dc_rdfig> A: http://www.opengalen.org/ from bijan

01:04:20 <bijan> A:|Open Galen

01:04:20 <dc_rdfig> Titled item A.

01:04:41 <bijan> A:Excellent FAQ about Galen.

01:04:41 <dc_rdfig> Added comment A1.

01:04:58 <sandro> "OpenGALEN: Making the impossible very difficult" I haven't laughed so hard in a long long time.

01:05:15 <bijan> A:Galen was the medical ontology that drove the development of FaCT and optimized implementations of very expressive DLs (starting with SHF)

01:05:15 <dc_rdfig> Added comment A2.

01:05:39 <bijan> A: And, as Sandro just pointed out, they have the *best* motto: "OpenGALEN: Making the impossible very difficult"

01:05:39 <dc_rdfig> Added comment A3.

01:06:11 <bijan> I'm giong to point lots and lots of people to this site

01:06:14 <bijan> It' sreally interesting

01:09:06 * sandro keeps reading

01:10:13 <bijan> Be interesting if OpenGALEN coverts to OWL

01:11:06 <sandro> " it has lots of features similar to KR languages (sometimes also known as ‘description logics’) such as KL-ONE, CLASSIC and BACK. However, it has been developed specifically to support the requirements of medical terminology, and has specific features to provide that support; for example, its handling of partonomies and the integration of multiple classifications for different purposes is unique, as far as we have been able to determine."

01:11:26 <bijan> Wow, query language, authoring tools, intermediate langauge

01:12:44 <bijan> software tools and engines exist. Reengineering the underlying formalism and inference engine, for example to FACT [20], is also under consideration.

01:13:39 * bijan happy to have confirmation that all those "GRAIL concepts" comments scattered through the FaCT code refered to this

01:13:46 <bijan> I thought it meant te "holy grail" concept

01:13:52 <bijan> That had some algorithmic meaning :)

01:14:00 <sandro> heh.

01:14:10 <bijan> It was confusing!

01:14:16 <sandro> at least it's not in python. (ba da bum.)

01:15:03 <bijan> I have wondered if Lisp FaCT is dead

01:15:08 <bijan> The released version is quite old

01:15:14 <bijan> And a bit, well, icky :)

01:16:18 <bijan> Hmm. the IR seems good

01:17:06 <sandro> What I was trying to ask without prejudicing the answer is do you expect GALEN to be used (at least by you guys) in checking for consistency (in ontologies and or ontologies+instance data) or for checking entailments (ie answering questions about ontologies, or about ontologies+instance data). That's four possibilities. There may be others.

01:17:32 <bijan> Er..I wasn't referring to galen the system

01:17:36 <bijan> But Galen the ontology

01:17:41 <bijan> Transliterated by ian into owl

01:17:57 <bijan> And the current use is to stress test our implementation :)

01:18:38 <bijan> And the real stress test (and one main use) is classification

01:18:42 <bijan> Which is sorta inbetween

01:18:57 <bijan> Finding hiddent subsumptions is sort of a query

01:21:27 * bijan should start using the badges zoe made

01:21:36 <bijan> Though it needs a "consistent" version too

01:23:02 <sandro> yeah, I don't really understand the DL mindset. I have a tenuous hold on "classification".

01:23:15 <bijan> Really?

01:23:24 <bijan> you shoudl play with the pellet web demo

01:23:30 <sandro> yeah>

01:23:32 <sandro> yeah?

01:23:43 <bijan> Classificaiton is just getting all the implicit subsumption relation made manifest

01:23:54 <sandro> url?

01:23:58 <bijan> I really need to lean on my guys to improve the display of the tree

01:24:06 <bijan> So you can see the hidden ones more easily

01:24:16 <bijan> sure: http://www.mindswap.org/2003/pellet/

01:26:25 <sandro> "This isn't entirely accurate or helpful, but I'm feeling drawing impaired at the moment, and I think it is somewhat helpful." That sounds like my friend Bijan.

01:26:35 <bijan> Heh

01:26:49 <bijan> That's so out of date

01:26:52 <bijan> i rell must fix

01:26:55 <bijan> really must

01:28:18 <sandro> so half my questions were about whether there was an A-Box. (ie ontology + instance data)

01:28:24 <bijan> Right.

01:28:26 <bijan> We do

01:28:31 <sandro> I know thats a false generalization, but stuill

01:28:36 <bijan> That was a super top priority

01:28:59 <bijan> We will shortly, i hope, have conjunctive Abox query

01:29:04 <bijan> Based on ian's iswc paper

01:29:17 <bijan> We're going to use that in an analysis of rdf path query languages

01:29:37 <bijan> (I.e, try to figure out some tradeoffs twixt speed and completeness)

01:30:20 <sandro> But I'm not seeing the demo you wanted me to see.

01:30:32 <bijan> See the links at the top of the page?

01:30:40 <bijan> and see in the scroll back /demo for the demo?

01:30:42 <bijan> :)

01:31:00 <sandro> ah

01:32:29 <sandro> Yay. It's great. I love it.

01:33:06 <bijan> :)

01:33:09 <bijan> Me too, actually.

01:33:28 <bijan> I'm going to have somone integrated with the Emacs owl mode

01:33:29 <sandro> oops. failed on http://www.w3.org/2002/03owlt/testOntology.

01:34:02 <bijan> N3?

01:34:12 <bijan> I could get an n3 switch

01:35:09 <karl_> karl_ is now known as karlcow

01:35:17 <sandro> n3? No its correctly getting the rdf. If I add ".rdf" i get the same error. It looks like the error is Jeremy's, not yours.

01:35:25 <bijan> Wait...what did it try

01:35:33 <bijan> Hmm When I look in a browser I get n3 :)

01:35:42 <sandro> Yeah, content negotation.

01:35:53 <sandro> It's aweful. :-) / 2

01:35:57 <bijan> Yes :)

01:36:08 <bijan> I so want a "mime/type" button in my browser

01:36:10 <bijan> Two sorts now

01:36:17 <bijan> one to force content negotionation to what i want

01:36:23 <bijan> One to force mimetype to what i want

01:36:30 <bijan> Forget stinkng "View source"

01:36:34 <bijan> Jsut do text/plain!

01:37:04 <sandro> is there a difference between forcing coneg and forcing mimetype? Oh, you want to override what the server told you the mime type was. Yes.

01:37:12 <sandro> Screw mime type.

01:37:15 <bijan> :)

01:37:27 <sandro> not text/plain -- whatever -- just ignore it.

01:37:47 * sandro hears his co-workers gathering kindling (for the heresy bonfire)

01:38:25 <bijan> :)

01:38:58 <sandro> Meanwhile, the pellet page should be using GET not POST. I can't paste the URL of the page I want to show you, because you guys used POST!

01:39:09 <bijan> It does both, actually

01:39:20 <bijan> It might require javascript to do so

01:39:26 <bijan> That was my first complaint about hte demo :)

01:39:50 <bijan> See: http://www.mindswap.org/cgi-bin/2003/pellet/pelletGet.cgi?inputString=&inputFile=http%3A%2F%2Fwww.mindswap.org%2F%7Ebparsia%2Fontologies%2Fsws%2Fiope-consistent-example.owl&conclusionsFile=&classifyFormat=NONE

01:39:59 <bijan> So you could compose the get yourself

01:40:00 <sandro> But I dont understand, does "The following error occured:

01:40:00 <sandro> UnsupportedFeatureException: Unsupported feature: Subject and object of owl:equivalentProperty should be a class [http://www.w3.org/2002/03owlt/testOntology#importedPremiseDocument, http://www.w3.org/2000/01/rdf-schema#subPropertyOf, http://www.w3.org/2000/10/rdf-tests/rdfcore/testSchema#inputDocument]

01:40:00 <sandro> Please contact the system administrator." mean the tested RDF had an error, or Pellet has an error?

01:40:24 <bijan> I'm betting a file error

01:40:36 <bijan> Someting weird has been happening today

01:40:44 <Cardinal> bijan: It sounds like you want a feature D[a]vey and I were talking about a little while back. The ability for a browser to tell you "This document is available in the following additional mime types: <foo>"

01:40:48 <bijan> I don't like these eror message

01:40:56 <bijan> Cardinal: maybe

01:41:05 * D[a]vey remembers that conversation

01:41:10 <sandro> To go along with "and the following styles...."

01:41:17 <Cardinal> Right.

01:41:31 <bijan> That file fails in the w3c validator too

01:41:33 <D[a]vey> I can't remember what we concluded. heh

01:41:39 <bijan> Parser Loading Error

01:41:45 <Cardinal> We concluded it would require an additional HTTP header, I think. :)

01:41:50 <bijan> Exception parsing: The scheme is not conformant.

01:41:51 <D[a]vey> yes, thats right :)

01:41:57 <bijan> Not for me

01:42:14 <bijan> I just want to be able to say, "Trye text/plain. Try application/xml.

01:42:18 <sandro> HTTP doesnt really give you enought for your browser to give you the choices well, alas.

01:42:27 <bijan> "If it breaks bad on me."

01:42:33 <D[a]vey> thats what we were saying bijan

01:42:37 <D[a]vey> uhm, sandro

01:43:00 * bijan does that now by saving a doc, changing the extension to something else, and reloading

01:43:06 <sandro> prove me wrong, Davey. :-)

01:43:14 <bijan> Or if I'm debuging wikis, I screw with the mimetype ;)

01:43:22 <D[a]vey> sandro: I'm not disagreeing.

01:43:57 <Cardinal> sandro: We were thinking of something like Also-Available-As: text/html, image/jpeg, ..

01:44:11 <sandro> Ah, makes sense.

01:44:36 <Cardinal> Since having the UA poll the server for "other stuff it may understand" would be prohibitively spammy.

01:44:40 <bijan> Weird

01:44:53 <bijan> That ontology parses if i paste it in, but doesn't if I retrieve the url

01:44:55 <bijan> In the rd fvalidator

01:45:36 <sandro> But I think I'll stick with arguing that Con-Neg is really only there to offer a simple migration path. It's okay for image formats, but that's about it.

01:46:13 <bijan> Oh, eww. that thing uses rdf:Class

01:46:16 <sandro> conneg between text/plain and text/html, or text/html and application/xml, or whatever.... that's just bad.

01:46:17 <bijan> It's not even an owl document

01:46:19 <bijan> Fssst

01:47:13 <bijan> Actually, seems quite borken in many ways.

01:47:18 <bijan> And the n3 seems unlike the rdf/xml

01:47:40 <bijan> The n3 is a lot shorter, for example

01:48:09 <bijan> And messed up the cardinality datatype

01:50:05 <bijan> Argh, NONE of my borwsers understand application/xml+rdf

01:50:08 <bijan> Everyone shoudl stop using it

01:50:40 <sandro> tell RDF core immediately if you have a good argument there.

01:50:55 <danbri> everyone should help patch mozilla

01:51:01 <bijan> "It doesn't work with any of my browsers" seems good enough to me

01:51:13 <bijan> "Including mozilla 1.5 which freakin ghas rdf builtin"

01:51:22 * Cardinal ponders a Moz extension to render application/rdf+xml as a XUL list.

01:51:31 <bijan> No no!

01:51:33 <bijan> Just as XML

01:51:37 * sandro types "cvs delete testOntology.n3" and contemplates hitting return.

01:51:53 <bijan> Why do these stupid browsers not go down to xml mode?!

01:52:01 <bijan> Give me a choice!

01:52:09 <bijan> Better

01:52:12 <bijan> Just do something usefuL!

01:52:12 <sandro> Maybe XML mode should be a switch on everything XML, include xhtml ?

01:53:32 * sandro presses return. no more testOntology.n3

01:53:48 <bijan> Woohoo!

01:53:54 <bijan> yes that woudl be groovy

01:54:25 <sandro> (some early techno song riff about having the power....)

01:54:28 <sandro> it's done. do you still get n3?

02:01:00 * bijan no longer getting the n3

02:18:53 <bijan> it's a little annoying that the owl test ontology isn'tin...owl

02:20:45 <danbri> what is it 'in'?

02:21:09 <bijan> RDFS

02:21:13 <bijan> Ish

02:35:50 <tim_away> Ah! iCal 1.5 allows URLs for events.

02:36:51 * tim_away noted concern about browsers not handling application/rdf+xml in TAG meeting

02:37:06 <bijan> it sucks

02:37:31 <bijan> but, I'll say, my ire focuses on the people using appication/rdf+xml too :)

02:37:35 <bijan> Maybe even primarily

02:37:42 <tim_away> It is part of a general longstanding problkemn - browsesr let you choise to save to disk or open with a external app, but don't give tyou the coice of handling ait as any ofth e many format sthey suport internally (plain text, XML, SVG, etc)

02:38:00 <tim_away> You would prefer?

02:38:08 <bijan> A way of handling it

02:38:10 <tim_away> for the mime type fo an RDF file?

02:38:20 <bijan> Oh, inprinciple I like the mime type

02:38:31 <bijan> And I made our guys change to it

02:38:36 <bijan> but then I culdn't browse any of our rdf

02:38:43 <bijan> So i started whining for a standard xml one :)

02:39:19 <tim_away> I guess we need a rev of CUAP, the coommon user agent problems note.

02:39:54 <Cardinal> bijan: You could employ a mod_rewrite hack not unlike the one some folks use for XHTML.

02:39:58 <tim_away> And while we are at it, please never have a "type URL here", jhave an icon well, like NeXTStep.

02:40:16 * tim_away really wanted NexTStep's tear-off menus today in OS x

02:40:58 <tim_away> If you use application/XML then the fragment ID has no meaning.

02:41:19 <bijan> Eh.

02:41:20 <danbri> maybe someone is packaging GNUStep for Fink... (but that's not really what you asked for...)

02:41:29 <danbri> application/xml you mean?

02:41:33 <danbri> (case mattering?)

02:41:39 <bijan> I so don't care about that. I'd rather be able to look at it in a browser

02:41:49 <bijan> And when I load it in an rdf library, well, tis' rdf and taht's that

02:43:45 <tim_away> application/xml yes

02:56:08 <_joshua> hey danbri, see you in two days

03:03:17 <danbri> :)

05:04:21 <Davey> Davey is now known as D[a]vey

09:44:41 * dajobe looks at logger stuff

09:55:40 <eaon> eaon is now known as eaon|lunch

10:38:44 <eaon|lunch> eaon|lunch is now known as eaon|busy

11:35:31 <dajobe> ok, logs from yesterday repaired

11:36:36 <dajobe>http://ilrt.org/discovery/chatlogs/rdfig/2003-10-08.html#T17-22-28

11:36:37 <dc_rdfig> B: http://ilrt.org/discovery/chatlogs/rdfig/2003-10-08.html#T17-22-28 from dajobe

11:36:50 <dajobe> B:|Yesterday's #rdfig logs repaired, starting from this date

11:36:50 <dc_rdfig> Titled item B.

11:37:11 <kao> what happened?

11:37:27 <dajobe> networking problems at our site

11:37:48 <kao> ic

11:37:59 <dajobe> B:up to [http://ilrt.org/discovery/chatlogs/rdfig/2003-10-08.html#T21-53-04-3|21:53:04]

11:37:59 <dc_rdfig> Added comment B1.

13:06:29 * sbp works on an OWL Syntax Checker

13:06:48 <sbp> currently passes 95/122 test cases--roughly 77%

13:06:57 <sbp> but the test cases seem rather... odd/broken

13:07:19 <sbp> a lot of them use out of date OWL terms, though, so perhaps I just need to find a later version

13:07:27 <bijan> Looks like the current download of racer has owl support!

13:08:03 <sbp> http://www.w3.org/TR/2003/CR-owl-test-20030818/ => http://www.w3.org/2002/03owlt/ => http://www.w3.org/2002/03owlt/approved.zip

13:09:05 <sbp> ah... approved.zip 28-Mar-2003, and yet the Tests CR is from August (and OWL Semantics and Reference)

13:09:53 <sbp> are the tests really five months behind the rest of the OWL specs suite, or am I missing an obvious link?

13:10:18 <D[a]vey> hmm, I'm still not sure how to decide what mime-type to set on my annotea messages

13:11:29 <D[a]vey> I guess its upto the the client... I'll make it user configurable.

13:11:50 <D[a]vey> although, I think I'll build some detection in there.

13:11:56 <sbp> it seems that many of the directories in /2002/03owlt/ have been modified after the zip, too

13:13:27 * sbp finds http://www.w3.org/2002/03owlt/editors-draft/draft/approved.zip - "Last-Modified: Wed, 01 Oct 2003 20:00:15 GMT" - seems more promising...

13:14:11 <dajobe> there was no owl test cases pubed since March, wonder why?

13:14:35 <dajobe> ah, owl home page is wrong

13:15:01 <sbp> no test case files since March, no test case spec since August

13:15:19 <sbp> AFAICT...

13:15:24 <dajobe> mail the comments list

13:15:37 <sbp> okay

13:15:52 <dajobe> it if's not obvious to an outsider, I don't see why they can't explain

13:16:04 <dajobe> AFAIK most of the OWL test cases work is done by WG members

13:30:33 <DanC> " It has been generated automatically from a file containg DAML annotations" yeah! " See here" boo. http://www.daml.org/services/members.html

13:31:09 <DanC> uses foaf... nifty!

13:34:53 <JibberJim> Hmm, Rosco seems to have trouble with the file... but I think foaf:Phone has rdf:resource not literal as its range - although not quite, it's a testing property and doesn't specify it...

13:36:25 * DanC notes that rdfs:Literal is a subclass of rdfs:Resource

13:36:58 <JibberJim> It is nifty though - could use FOAF:Group too to define their group.

14:17:04 Topic now Semantic Web hack-n-chat http://rdfig.xmlhack.com/ latest logs http://ilrt.org/discovery/chatlogs/rdfig/latest wiki http://esw.w3.org/topic/FrontPage

14:17:04 Users on #rdfig: logger mdean_ dajobe monkeyiq eikeon deusx tim_away bijan ericm larsbot JibberJim iwaim d2m nmg merriam chrisc burtonator xover jeen Esaj eaon|busy mea_culpa D[a]vey dngor mmealling ear1grey datum Ol bitsko karlcow darobin josek MarkB shellac kao sbp DanC kham Oblomov Wack dhmspector GabeW grumbles sandro grault _joshua ericP deltab tav eikco golbeck xower jordan grove dc_rdfig

14:17:04 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

14:48:43 <D[a]vey> D[a]vey is now known as Davey

14:52:14 <bijan> Ugh, right into wsdl telecon

14:52:24 <Davey> Hey libby_

14:52:47 <libby_> hi Davey

14:53:07 <Davey> how are you today?

14:53:31 <libby_> having a laptop crisis ;)

14:53:40 <libby_> stupid computers

14:53:40 <Davey> hehe, sounds like fun

14:53:56 <libby_> fortunately shella has lent me his

14:54:04 <Davey> :)

14:54:08 * Davey waits for his apple lappy

15:01:28 <libby_> libby_ is now known as libby

15:08:17 Topic now Semantic Web hack-n-chat http://rdfig.xmlhack.com/ latest logs http://ilrt.org/discovery/chatlogs/rdfig/latest wiki http://esw.w3.org/topic/FrontPage

15:08:17 Users on #rdfig: logger dajobe iwaim libby mdean_ monkeyiq eikeon deusx tim_away bijan ericm larsbot JibberJim d2m merriam chrisc burtonator xover jeen Esaj eaon|busy mea_culpa Davey dngor mmealling datum Ol bitsko karlcow darobin josek MarkB shellac kao sbp dc_rdfig grove jordan xower golbeck eikco tav deltab ericP _joshua grault sandro grumbles GabeW dhmspector DanC kham Oblomov Wack

15:08:17 <NickServ> Password accepted - you are now recognized

15:08:17 <MemoServ> You have no new memos

15:08:17 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

15:27:05 Topic now Semantic Web hack-n-chat http://rdfig.xmlhack.com/ latest logs http://ilrt.org/discovery/chatlogs/rdfig/latest wiki http://esw.w3.org/topic/FrontPage

15:27:05 Users on #rdfig: logger dajobe_ iwai libby mdean_ monkeyiq eikeon deusx tim_away bijan ericm larsbot JibberJim d2m merriam chrisc burtonator xover Esaj eaon|busy mea_culpa Davey dngor mmealling datum Ol bitsko karlcow darobin josek MarkB shellac kao sbp DanC kham Oblomov Wack dhmspector GabeW grumbles sandro grault _joshua ericP deltab tav eikco golbeck xower jordan grove dc_rdfig

15:27:05 <MemoServ> You have no new memos

15:27:05 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

16:06:45 <sbp> up to 201/212 passes; 94.8%

16:16:42 <bijan> Cool

16:45:54 <Davey> Davey is now known as D[a]vey

17:07:28 <D[a]vey> D[a]vey is now known as Davey

17:14:56 <shellac> ls

17:15:11 <libby> heh

17:15:11 <shellac> doh - two monitors, no brain

17:15:15 <sbp> hehe

17:15:50 <libby> least it wasn't your password :)

17:16:05 <libby> how you doing sbp? take it you got back alright

17:16:09 <shellac> msg nickserv identify froody1

17:16:25 <libby> hehhe

17:16:41 <sbp> whoo: 207/212; 97.6%

17:16:52 <sbp> oh, hey libby. sorry, didn't realise we hadn't spoken since then

17:17:04 <sbp> I did indeed! Jim streaked across the UK in no time

17:17:19 <libby> excellent

17:17:35 <libby> well, it was great fun having you guys staying

17:17:35 <sbp> now working on a crazy, crazy OWL Syntax Checker

17:17:44 <libby> crazy fool!

17:17:45 <sbp> yes. and thanks again for having us

17:17:55 <sbp> sorry I wasn't up in the morning either to say bye and thanks etc.

17:18:06 <libby> s'ok :)

17:18:43 <sbp> how's your laptop? hope it wasn't my influence that borked it

17:19:20 <libby> heh, nah, it was being a bit flakey for a few days

17:19:33 <Davey> what brand was it libby?

17:19:40 <libby> it's pretty knackered, but apple-care to the rescue!

17:19:44 <sbp> phew

17:19:44 <libby> a mac

17:19:48 <libby> ibook.

17:19:53 <sbp> heh. all being taken care of?

17:19:54 <libby> more than 2 years old now

17:20:02 <libby> heh, yep

17:20:04 <Davey> ah, cool :)

17:20:09 <sbp> good, good

17:20:17 <Davey> I want a 15" PowerBook :/

17:20:37 <libby> they're a bit big to lug round, but very cool

17:20:41 <shellac> try it on #santa

17:20:44 <Davey> I'm asking apple for 1*12", 2*15" and 1*17" with all the trimmings.

17:21:08 <Davey> I would love the 17" but I want something I can actually *use* on a plane.

17:21:39 <libby> yeah. not in cattle class

17:21:41 <shellac> if you can afford an 17" you can afford business class

17:21:54 <Davey> who said I'm paying for the lappies?

17:22:10 <Davey> hell, I'm not even planning to pay for the flights :)

17:22:17 * Davey hugs the PHP Conference Circuit

17:22:31 <libby> heh

17:46:52 <Davey> Davey is now known as D[a]vey

18:08:36 <sbp> 210/212, but the remaining two are just timing out rather than actively giving the wrong answer

18:13:54 <libby>http://www.kingtiny.net/weblog/RSStoOutlook.html

18:13:54 <dc_rdfig> C: http://www.kingtiny.net/weblog/RSStoOutlook.html from libby

18:14:34 <libby> C:|RSS to Outlook by Rahul Singh

18:14:35 <dc_rdfig> Titled item C.

18:15:23 <libby> C:in response to [http://www.ozzie.net/blog/2003/09/20.html|Ray Ozzie's questions about RSS and iCal]

18:15:23 <dc_rdfig> Added comment C1.

18:18:07 <libby> C:relatedly, [http://igargoyle.com/rss/1.0/modules/event/|an updated version of RSS events module] (by nym, I think), and [http://planb.nicecupoftea.org/archives/000075.html|me suggesting some possible improvements] or just using [http://www.w3.org/2002/12/cal/|RDFical] instead :)

18:18:07 <dc_rdfig> Added comment C2.

19:31:14 <GabeW> quick question: if I'm appending arbitrary text to the end of a HTTP URI, do I need to follow any escaping rules specific to HTTP (I don't see any in RFC 2616, but that may not be everything)..

19:31:31 <GabeW> in other words, any escaping rules beyond RFC2396

19:33:56 <larsbot> what do you mean by "to the end of a URI"?

19:34:06 <GabeW> hi larsbot

19:34:07 <GabeW> i mean

19:34:23 <GabeW> http://www.example.com/something and I want to add some arbitrary string

19:34:24 <dc_rdfig> D: http://www.example.com/something from GabeW

19:34:26 <GabeW> erg

19:34:29 <GabeW> D:|Ignore me

19:34:30 <dc_rdfig> Titled item D.

19:35:05 <GabeW> I need to run that arbitrary string through some escaping

19:35:19 <GabeW> and I want to know exactly the requirements for that escaping and whether there is anything specific to the HTTP URI scheme

19:35:29 <GabeW> over and above whats required in RFC 2396

19:35:59 <larsbot> there isn't anything specific to HTTP URIs, no

19:37:03 <GabeW> ok, i don't see anything in rfc 2616, but I was wondering if there were any practical issues with just strictly applying only the escaping rules in 2396 and it sounds from your response that there is not problem.

19:38:09 <sandro> So what's the right way to give an RDF URI (eg http://www.w3.org/1999/02/22-rdf-syntax-ns#type) in a nice HTML document? In the OWL Test Results page the name of the test itself is an a/href to the RDF URI, but that strikes me as a for-experts-only approach. You don't want random folks having their browser showing them RDF/XML.

19:38:23 <sandro> (at least not unless they click on a link like [data] )

19:43:33 <sbp> whoo, 100%!

19:45:18 <bijan> Congrtats!

19:45:22 <bijan> Is this in python?

19:45:36 <sbp> thanks. it is indeed; based on the RDF API thing that I've been hacking on

19:45:37 <larsbot> GabeW: there is one issue with non-ASCII characters

19:45:39 * bijan will steal that, then; plannin gon porting pellet to python

19:45:45 <bijan> Ugh. Port it ot rdflib :)

19:45:53 <sbp> ugh right back at you

19:45:57 <sbp> :-)

19:46:09 <GabeW> larsbot: well, non-ASCII characters are a whole nuther ball of wax

19:46:32 * sbp wonders if you can do stuff like "G += Graph(uri=uri)" in rdflib

19:46:59 <bijan> Ask daniel :)

19:47:06 <sbp> heh

19:47:12 <sbp> it'll be interesting to see what he thinks of it

19:47:19 <sbp> he'll probably just kick me

19:47:40 <sbp> but I'm hoping that the two implementations will be so dissimilar as to merit this new one

19:52:04 <sbp> I've called my OWL Syntax Checker "hoot", btw

19:52:21 <sbp> it reminds me of both OWL and DanC

19:52:57 <JibberJim> The joys of implementing early - you get to pick the bad puns for your product names...

19:53:06 <sbp> heh, heh. indeed

19:53:30 <sbp> actually, there's a "hoolet", but I think that sufficiently distinct enough...

19:53:54 <sbp> hoot was a rather automatic choice

19:55:12 <sbp> if I do a Consistency Checker, I'm thinking of "hooter"

19:58:24 <D[a]vey> D[a]vey is now known as Davey

19:59:17 <GabeW> larsbot: what was that issue, btw, just to make sure I'm already aware..

20:03:40 <larsbot> just the encoding of the characters

20:04:21 <GabeW> right

20:04:22 <GabeW> ok

20:04:42 <JibberJim> And if you do a pair of consistency checkers?

20:05:22 <sbp> heh, heh. maybe hooterus/hooterii then...

20:20:09 * DanC wonders if anybody here has read the draft TAG finding on extensibility and versioning http://www.w3.org/2001/tag/doc/versioning-20031003.html

20:27:26 <GabeW> wow

20:28:35 <DanC> wow what... it's so long?

20:28:52 <GabeW> yes, that, and it seems ambitious and meaty

20:29:00 <GabeW> and a good thing to produce

20:29:06 <GabeW> this comes up a lot

20:29:21 <DanC> too ambitious, I think, maybe.

20:29:36 <GabeW> that was my first thought

20:29:48 <GabeW> obviously, I haven't had time to parse it in detail

20:30:08 <GabeW> there are documents out there on extensibility and xml schema

20:30:28 <GabeW> i wonder how much this overlaps

20:31:06 <DanC> "Definition: a language is an identifiable set of vocabulary terms that has defined constraints."

20:31:15 <DanC> that's a very non-traditional definition of language

20:38:27 <sandro> Yay. Very nice discussion of forwards/backwards compatible. Although Google argues strongly they should be called "forward" and "backward" (no "s").

20:38:56 <sandro> (more strongly on forward)

20:40:10 <sandro> :-) " * Big bang. Applications are expected to abort if they see an unexpected version."

20:41:08 <GabeW> yah, I like the attempt to classify the strategies..

20:42:27 <DanC> hm... I don't like the definitions of fwd/back compat

20:42:44 <DanC> they say nothing about meaning of documents, only what software does

20:48:38 <sandro> you have higher standards. I'm happy that the defns given aren't totally wrong. :)

23:29:42 <Davey> Davey is now known as D[a]vey

23:29:45 <D[a]vey> D[a]vey is now known as Davey