Semantic Web Interest Group IRC Chat Logs for 2002-07-11

01:49:26 <connolly> connolly is now known as DanC_

02:40:48 <MysticOne> [GlobalNotice] Hello everyone, we hope you're having a pleasant evening/morning. If you're in the mood for some music, fun, all that good stuff, check out the Open Projects Network's streaming radio station, WOPN! Check #WOPN for URLs to the streams and to take part in the show... online *now*! Thanks!

04:21:21 * Seth waves for eikeon

04:40:51 * Seth wonders what % of rdf out there does not conform to new WG standards, and if that % is high, what use would a conforming parser be in today's world

04:42:08 <Seth> for example ... does http://www.w3.org/2000/01/rdf-schema conform ?

04:46:22 <Seth> does anyone know if the document at url http://www.w3.org/2000/01/rdf-schema conforms to new WG standards ?

06:02:11 <xovver> xovver is now known as xover

07:25:12 <connolly> connolly is now known as DanC_

07:25:33 <DanC_> cool... converting dia files to RDF works, at least for the first case in http://www.swi.psy.uva.nl/usr/Schreiber/docs/owl-uml/owl-uml.html

07:27:20 <DanC_>http://www.w3.org/2002/03owlt/umlp/dia2owl.xsl

07:27:20 <dc_rdfig> A: http://www.w3.org/2002/03owlt/umlp/dia2owl.xsl from DanC_

07:27:31 <DanC_> A:|dia2owl -- convert UML diagram from dia to OWL

07:27:31 <dc_rdfig> titled item A

07:27:42 <DanC_> A:works for just one simple case, so far.

07:27:42 <dc_rdfig> added comment A1

07:28:11 <DanC_> A:details in the [Makefile|http://www.w3.org/2002/03owlt/umlp/Makefile]

07:28:11 <dc_rdfig> added comment A2

07:32:07 <DanC_> hmm... dia doesn't seem to support multiplicity/cardinality on attributes.

07:36:41 <DanC_> B:skipping 2nd example... dia doesn't seem to support multiplicity/cardinality on attributes, and the enumeration seems goofy.

07:36:42 <dc_rdfig> Label B not found.

07:36:50 <DanC_> A:skipping 2nd example... dia doesn't seem to support multiplicity/cardinality on attributes, and the enumeration seems goofy.

07:36:50 <dc_rdfig> added comment A3

08:00:54 <DanC_> ok... got another case working...

08:13:03 <DanC_> "A: The native Dia file format (extension .dia) is XML compressed with GZip. You can select not to compress the files in the Preferences."

08:13:04 <DanC_> duh

09:25:29 <rreck> anyone know how to convert RDF to GML ?

09:28:20 <larsbot> what's GML here? Generalized Markup Language?

09:28:31 <larsbot> (hi rreck :)

09:35:34 <kao> rreck: if you mean GML=Geography Markup Language, they got a Profile for expressing GML in RDF

10:03:01 <rreck> GML=Geography Markup Language, yes but i want to go from rdf to GML

10:14:03 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

11:01:21 <lilo`> [GlobalNotice] Hi all. Some rehubbing will be necessary in a moment; we're still adapting to the placement of new hubs. Please bear with us.

11:09:57 <lilo`> [GlobalNotice] Okay folks. Here we go. This will be a fairly major rehub, but we'll try to make it short.

11:13:05 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

11:14:56 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

11:18:27 Topic now RDF/Semantic Web 24x7 chat - http://rdfig.xmlhack.com/ blog

11:18:27 Users on #rdfig: logger quux Hixie larsbot dajobe JibberJim deltab kao DanC_ Skyline AndyS shellac dc_rdfig sandro MarkB dom thelsdj xower eikeon RM_39_ eXpat rreck AaronSw` sbp danbri dmiles oierw datum grove timmmit danb_lap jang lasDesk1

11:18:27 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

11:21:03 <lilo`> [GlobalNotice] Hi all. Rehubbing is complete. Thanks for your patience.

11:35:40 <em_lap> em_lap is now known as emmit

13:42:02 * Seth bemoans that rdf parsers don't seem to want to tell you what the name space bindings are

13:43:06 <Seth> i mean if they ~always~ gave us triples of the following sort, then we could ~always~ convert the rdf back to some kind of human readable data:

13:43:25 <Seth> <http://www.w3.org/1999/02/22-rdf-syntax-ns#> namespaceName "rdf".

13:43:47 <Seth> for some meaning of ~always~

13:46:07 <kao> i believe youre not the only one who bemoans this fact

13:47:13 <Seth> well its nice to know that i have fellow bemoaners ... have we met before?

13:47:32 * JibberJim has bemoaned it regularly too...

13:47:34 <kao> not in person, but probably on irc

13:47:52 <danbri> still leaves you with problem of re-serialising bnodes

13:48:32 <Seth> danbri, what is that problem ? .. havent run into it yet

13:48:42 <danbri> would be nice if you coudl get this from parsers, but probably can make do with getting it from descriptions of namespaces, eg. that 'dc' is favoured short name for the http://purl.org/dc/elements/1.1/ namespace

13:48:58 <danbri> problem: if you have lots of bNodes all interconnected, you can't rdf/xml-serialize without inventing uris

13:49:35 <Seth> well yes one can compile their own dictionary, but what heppens when you encounter a new namespace .. where do hou get the data?

13:49:43 <kao> but if your source was rdf/xml, you dont have that problem

13:50:55 <Seth> kao, huh? ... if my source is rdf/xml and all i see is the triples i get from the parser there is no way i can get the namesapace binings in a guaranteed manner

13:51:15 <kao> sorry, i was referring to the interconnected bNode problem

13:51:47 <kao> in that, if you have no interconnected bNodes in your source, you wont have them when you serialize it again

13:52:26 * danbri nods

13:52:56 <Seth> danbri, yes that is a problem with researilizing to rdf/xml, but not to quads ... with quads you can just user the bnode id _:000000000123

13:53:47 <kao> quads? subject, predicate, object, something?

13:53:53 <JibberJim> I still want to know how I refer to a bnode I've named in a seperate rdf/xml document in a way that an agent can understand.

13:54:35 <Seth>http://robustai.net/sailor/grammar/quads.html

13:54:36 <dc_rdfig> B: http://robustai.net/sailor/grammar/quads.html from Seth

13:54:43 <Seth> B:| Quads

13:54:45 <dc_rdfig> titled item B

13:55:10 <Seth> jibber, you can't thats why we call it a bnode

13:55:54 <JibberJim> Yes, but if I label it in the XML with rdf:about="http:something" then it's now no longer a bnode

13:56:07 <Seth> kao, yes the short answer is that quads are <subject> <property> <object> <context>

13:57:00 <Morbus> danbri: you editing the FOAF wiki? wanna fix the broken "old list" link too?

13:57:00 <JibberJim> So now in a different XML/RDF file I want to say something is <http:something> I want to be able to point to a uri where that something is defined.#

13:57:09 <Morbus> was just about to, but got yelled out for conflicting edits ;)

13:58:17 <kao> Seth: sounds interesting, for doing formulas etc. if i say s p o c, can i say c x y? ie. is a context a regular resource?

13:58:17 <Seth> jibber, can't do it if its a bnode

13:58:28 <JibberJim> It's no longer a bnode, I've labeled it.

13:58:49 <Seth> kao, yes a context is a regular resource

13:59:14 <Seth> jibber, ok then you gave it a uri

14:01:16 <JibberJim> so yeah, I give it a URI http://jibbering.com/moomin/snork#123443432432423 which I then use in another file, how do I tell an agent that that there's more information about that URI at such and such a URI.

14:01:41 <kao> rdfs:seeAlso ?

14:02:25 <JibberJim> I think that's a woolier relationship than I'm wanting to say and rdfs:isDefinedBy is a stronger one.

14:02:37 <DanC_> the swap RDF parser (sax2rdf.py) passes on namespace bindings. (not as triples, though. you can't make up triples that were'nt in the document, or stuff like log:notIncludes gets screwed up)

14:02:48 <Seth> jibber, <rdf description about=http://jibbering.com/moomin/snork#123443432432423> should work

14:02:48 <danbri> JibberJim> I still want to know how I refer to a bnode I've named in a seperate rdf/xml document in a way that an agent can understand.

14:03:22 <danbri> ...you usually don't want to refere to the bNode, but to the thing the bNode denotes. That's easier (though there are no guarantees) sicne you can use identifying properties of that thing to write further descriptoins of it in a 2nd doc.

14:03:31 <danbri> morbus: which 'old list' link? happy to fix.

14:03:51 <danbri> yes, rdfs:seeAlso

14:04:34 <Morbus> danbri: ldodds is editing it now for his matic... but the "old list of classes and properties" is borked.

14:05:07 <danbri> ah, that one. ok. thought you meant mailing list.

14:05:30 <Seth> danbri, how does rdfs:seeAlso help? the object of rdfs:seeAlso cannot be a bnode in another rdf/xml doc ... im confused

14:05:37 <kao> Seth, what do the 'terminals' (. ? !) in quads mean?

14:05:51 <Seth> . means assert

14:05:57 <Seth> ? means question

14:06:06 <Seth> ! means imparative

14:06:11 <kao> you can say rdfs:seeAls <the-document-whery-you-defined-that-node>

14:06:14 <Seth> same as english

14:06:41 <Seth> kao, but how do you call out 'that node' ?

14:07:38 <kao> i thought the node had an uri by then? http://jibbering.com/...#1234...

14:08:10 <kao> i was referring to "tell an agent that that there's more information about that URI at such and such a URI"

14:08:48 <kao> otherwise, you'd need some path language

14:09:32 <Seth> kao, well yes if it has a uri there is no prob .. if i have a document with a lot of foaf people in it, there is no way to say that <http:.....#seth> seealso <foaf:document> and have it mean anything interesting

14:09:39 <kao> i experimented with that, but found it is difficult as you can have multiple occurences of a property in no fixed order

14:12:17 <danbri> morbus, fixed it.

14:13:30 <Seth> kao, im curious what you really meant when you said " i was referring to "tell an agent that that there's more information about that URI at such and such a URI"

14:14:21 <kao> it was a question by JibberJim above

14:15:03 <danbri> oops, fixed the old list, not the link to it!

14:15:19 <JibberJim> Okay, Seth, look at http://jibbering.com/foaf.rdf - there I say that some URI's are img:depictions of me, but those URI's are meaningless unless you've also got the RDF from the rdfs:seeAlso.

14:17:20 <danbri> doesn't quite parse as rdf

14:17:33 <JibberJim> Sorry, I must've just broken it...

14:19:43 <danbri> I'm not quite sure what you're trying to say in that markup...

14:20:42 <Morbus> danbri: i'm not sure where to go with FOAF, or where best documentation needs to occur. another tutorialish doc like edd's? a clearer spec?

14:20:58 <JibberJim> Fixed now, I'm trying to say that the resource "such and such" is a img:depiction of me, and that is defined at the uri ".../rdfweb/combined.rdf"

14:21:03 <ambient> by asserting <a> <rdfs:seeAlso> <b>, am I also asserting whatever the document b says about a?

14:21:56 <danbri> the bit that's least documented, I think, is the 'and then what...?' aspect. People can write these files, but isn't clear how to get them linked up, how they find their friends ones, what conventions there might be for naming them, how to link from their homepages, etc.

14:22:13 <Morbus> alright.

14:22:16 <danbri> ...which services can consume the data (libby's, jim's, edd's, mine...).

14:22:27 <DanC_> no, ambient, rdfs:seeAlso isn't as strong as that. (that's sorta what daml:imports means)

14:22:31 <danbri> So there's a gap between understanding theres some abstract rdf vocabulary, annd getting it used

14:22:40 * danbri doesn't really undestand daml:imports

14:22:51 * DanC_ doesn't think anybody really does

14:22:54 <danbri> ...is it recursive? can it do versioning?

14:23:13 <danbri> ah ok, I don't feel so bad then

14:23:17 <DanC_> yes, it seems to be recursive. Dunno what you mean by 'do versioning'?

14:23:50 * DanC_ tries converting http://cgi.omg.org/docs/ad/01-02-03.txt to RDF using http://www.interdataworking.com/converter/

14:24:04 <danbri> I mean, if I daml:import http://example.com/connolly-useful-info.rdf, I'd much prefer to qualify the import with a version stamp (hash etc)

14:24:38 <danbri> ....in case you later change that doc to daml:import http://www.urbanlegends.com/stuff-a-friend-told-us.rdf

14:24:56 <danbri> ie its too risky without some such mechanism

14:25:23 <kao> danbri, what about the recently mentioned contexts?

14:26:10 <danbri> yeah, in practice, most rdf systems have some such notion. Doesn't help clarify (or make deployable) daml:imports, as far as I can see...

14:26:42 <DanC_> the only way I can see to do qualified import, at this point is to say: { <> daml:imports <useful.rdf>. <useful.rdf> crypto:sha-1 "23lk23". ...etc... }. i.e. state an assumption about the document, which, if false, makes the document sorta nil

14:27:02 * danbri nods

14:27:35 * DanC_ notes my use of "the document" was ambiguous, but gathers danbri got the gist

14:27:41 <danbri> I was chatting with Edd about this stuff. We've been signing RDF docs w/ GPG, and want a way of associating big load of RDF docs without having to sign each of them.

14:27:58 <DanC_> did anybody look at dia2owl? I didn't get very far, but I wonder if anybody else thinks it's promising

14:28:15 <danbri> ...eg I "stand by" what I said in lots of photo metadata .rdf files, but can't be bothered to GPG sign them individually.

14:28:22 * Seth sigh redfoot doent give me valid looking triples for jibbers foaf

14:28:36 <danbri> But I might GPG sign a table-of-contents / manifest that listed their sha1 hashes

14:28:53 <danbri> the rdf is malformed, I think

14:28:59 <danbri> dia2owl: not yet

14:29:04 <JibberJim> I think I've fixed it, the validator is happy with it.

14:29:25 <larsbot> Danc_, I've got a bound version of the XTM spec here that says "please return to Dan Connolly"... :)

14:29:26 <danbri> linking WebOnt to the modelling work in the UML tradition seems very worthy

14:29:36 <larsbot> you still want it? :)

14:29:54 <DanC_> holy cow, really larsbot? yes, I still want it.

14:30:28 <DanC_> that's from the March 2001 conference in Austin.

14:30:32 <larsbot> yeah, Steve Pepper found it in the office here

14:30:40 <larsbot> that might be. we have no idea where it comes from :)

14:31:02 <larsbot> if you'll be in Montréal or Baltimore you can get it there

14:31:06 <larsbot> otherwise I can snail mail it

14:31:35 <DanC_> I get nervous about committing my knowledge to paper, cuz I might lose it. This is the first time the "please return..." thing has actually worked. No, I take that back; it worked one other time, but I had only misplaced the thing for a few minutes, and it was only returned from a distance of a few hundred feet.

14:31:54 <DanC_> I do plan to be in Montreal, but I might ask you to snail mail it sooner.

14:32:10 <larsbot> that's up to you

14:32:17 * danbri is briefly reminded of http://wwww.bookcrossing.com/

14:32:18 <Seth> jibber, are you fairly sure your foaf is valid rdf ?

14:32:22 <DanC_> ok, I'll send you mail, larsbot

14:32:26 <larsbot> fine :)

14:33:16 <larsbot> danbri, that's an interesting site

14:33:25 <JibberJim> Well the RDF validator passes it fine Seth, check you've got the most up-to-date one.

14:33:54 <Seth> jibber, ok must be my redfoot parser

14:34:07 <danbri> yes, I thought so too. If only I could bear to part with books...

14:34:29 * DanC_ checks out bookcrossing... cool! harks to http://dm93.org/z2001/ShareLifeThruArt

14:34:54 <larsbot> heh. a friend of mine has a giving-away-books party every time he moves

14:35:01 <larsbot> I've been considering whether I should do the same

14:35:07 <larsbot> *some* books I can part with easily

14:35:51 * kao is afk

14:35:53 <DanC_> logger, pointer?

14:35:53 <DanC_> See http://ilrt.org/discovery/chatlogs/rdfig/2002-07-11#T14-35-53

14:38:19 <JibberJim> The problem with bookcrossing is, you need to register your book before releasing it, I often leave books places (when travelling) but wouldn't have the opportunity to register them first.

14:38:50 <danbri> good point

14:39:39 <JibberJim> Now if I could give the book id, just as a URI on a domain I control we'd not have the problem...

14:42:31 <danbri> <s:BookRelease s:date="2002-07-11" s:comment="Left in Bristol Temple-Meads station" s:geoLoc="123-321-321"><s:bookDetail><s:Book s:title="The RDF Specifications" s:publisher="cheesebag reprints in" .../</..etc

14:42:45 <danbri> that's all you need...

14:42:48 * DanC_ does some family time...

14:42:58 <danbri> and an ID inside the cover would just be another property

14:43:20 <danbri> of the release not the Book, since the latter is the book-as-a-work not a particular copy of it (arguably...)

14:44:14 * libby talking to Eamonn and Tony from elsevier

14:44:49 * libby shows them the chump...

14:44:55 <libby>http://hdl.handle.net/1014/yads

14:44:55 <dc_rdfig> C: http://hdl.handle.net/1014/yads from libby

14:45:30 <libby> C:|YADS - a description service

14:45:30 <dc_rdfig> titled item C

14:46:35 <libby> C:an interesting model which uses RDF designed for manageing resource collections

14:46:36 <dc_rdfig> added comment C1

14:46:47 <Seth> jibber, your rdf does not pass the validator at http://www.w3.org/RDF/Validator/ARPServlet

14:47:16 <JibberJim> It does for me :-(

14:47:19 <libby> C:the method of using RDF - which we think fecili9tates the use of scemas to produce just in time instance data

14:47:19 <dc_rdfig> added comment C2

14:47:34 <libby> C:and stom ereally neat SVG pictures - lovely

14:47:34 <dc_rdfig> added comment C3

14:47:55 <Seth> hmmm are we talking abut the same rdf ? [[

14:47:56 <Seth> <?xml version="1.0" standalone="yes"?>

14:47:57 <Seth> <rdf:RDF xmlns:img="http://jibbering.com/2002/3/svg/#" xmlns:wn="http://xmlns.com/wordnet/1.6/" xmlns:foaf="http://xmlns.com/foaf/0.1/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:an="http://rdf.desire.org/vocab/recommend.rdf#" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#">

14:47:57 <Seth> <foaf:Person>

14:47:57 <Seth> <foaf:name>Jim Ley</foaf:name>

14:47:58 <Seth> <foaf:mbox rdf:resource="mailto:jim@jibbering.com"/>

14:48:00 <Seth> <foaf:depiction rdf:resource="http://jibbering.com/dinner/Jbeer.jpg"/>

14:48:02 <Seth> <foaf:homepage rdf:resource="http://jibbering.com"/>

14:48:04 <Seth> <img:depiction rdf:resource="http://jibbering.com/rdfweb/#1016648055840_0">

14:48:06 <Seth> <rdfs:seeAlso rdf:resource="http://jibbering.com/rdfweb/combined.rdf"/>

14:48:08 <Seth> </img:depiction>

14:48:10 <Seth> <img:depiction rdf:resource="http://jibbering.com/rdfsvg/#1025619832190_0">

14:48:10 <JibberJim> That's the old RDF!

14:48:13 <Seth> <rdfs:seeAlso rdf:resource="http://jibbering.com/rdfweb/combined.rdf"/>

14:48:14 <Seth> </img:depiction>

14:48:14 <JibberJim> get the new one.

14:48:16 <Seth> <img:depiction rdf:resource="http://jibbering.com/rdfsvg/#1025708928490_2">

14:48:18 <Seth> <rdfs:seeAlso rdf:resource="http://jibbering.com/rdfweb/combined.rdf"/>

14:48:20 <Seth> </img:depiction>

14:48:22 <Seth> </foaf:Person>

14:48:24 <Seth> </rdf:RDF>

14:48:26 <Seth> ]]

14:48:28 <Seth> oh you switched on me

14:48:56 <libby> C:Here's a coool [London Tube map example |http://www2.elsevier.co.uk/~tony/yads/lib/tube/svg.html]

14:48:56 <dc_rdfig> added comment C4

14:49:10 <danbri> C:Interesting, it has a handle. I had a RDF resolver thingy for handles somewhere. (rummages) [http://www.ilrt.bristol.ac.uk/discovery/rdf-dev/roads/cgi-bin/desire/handle2rdf?URI=hdl%3A%2F1014%2Fyads|resolution info for hdl:/1014/yads] (scraped,...).

14:49:10 <dc_rdfig> added comment C5

14:49:29 <Seth> jibber, what's the url again?

14:49:40 <JibberJim> - http://jibbering.com/foaf.rdf

14:50:41 <Seth> jibber, kewl .. that makes more sense ... now it looks like i though you meant

14:54:12 <Seth> jibber, do you have a python method to display a polypath ?

14:55:25 <Seth> or i guess a rdf method to make a svg of a polypath ?

14:55:43 <JibberJim> Erm, not in python, basically you put the various properties into an SVG document that looks like http://jibbering.com/2002/7/example.svg

14:58:17 <JibberJim> Or you can use the various undocumented client-side SVG stuff available on my site e.g. http://jibbering.com/rdf/highlight-3.svg?url=http://jibbering.com/rdfweb/1020892863024.rdf

14:58:55 <Seth> aaah .. jibber, ... in your rdf, the data that is the object of polypath ... how do i transform tha into a svg? .. or do i?

14:59:21 <JibberJim> You just place it in an SVG d attribute of a path element.

14:59:35 <JibberJim> and then clip an image to that path.

15:00:14 <Seth> .. hey im a newbe ... can you break that down for me?

15:02:53 <Seth> .. i mean are these supposed to be seen as pictures by humans .. how do i transform them to pictures in my sailor agent?

15:06:28 <JibberJim> Not easily, it's not simple unless you know SVG, basically you create an SVG document as explained in http://jibbering.com/2002/7/example-img-depiction.html

15:08:53 <Seth> what is the point of passing this data in rdf rather than putting it in a svg file and passing the uri of that file?

15:10:11 <JibberJim> because we can do things with the RDF, we can say find me a picture of a Jim Ley which takes up at least 50% of the photo and doesn't also have a beer in the photo.

15:10:48 <JibberJim> Why mark-up anything in RDF?

15:11:55 <Seth> i know how to use rdf for stuff ... i just dont know how to use the data that you put in there as the object of polypath

15:12:27 <JibberJim> Also you can combine the information from multiple RDF sources - for example http://jibbering.com/rdf/co-search.1?person=mailto%3Adanbri@w3.org is a combination of all the pictures of danbri which I've got.

15:13:43 <jang> quick question, for which I apologise in advance

15:14:28 <jang> is there any way of getting xslt to rewrite java webapp web.xml files?

15:16:49 <jang> ...which are distinguished by not using namespaces

15:17:53 * danbri tried to get namespaces/rdf into the Jakarta stuff a while back... http://w6.metronet.com/~wjm/tomcat/ToFeb11/msg00005.html

15:18:17 <danbri> is that the same format? tomcat config files?

15:18:21 <danbri> bummer

15:50:25 <Seth> where is the rdf schema that defines namespace? ... in other words [[

15:50:26 <Seth> <http://www.w3.org/1999/02/22-rdf-syntax-ns#> type namespace.

15:50:26 <Seth> <http://www.w3.org/1999/02/22-rdf-syntax-ns#> nsPrefix "rdf".

15:50:26 <Seth> namespace type class.

15:50:26 <Seth> namespace definedIn ?schema.

15:50:36 <Seth> ]] what is the value of ?schema

15:55:28 <Seth> am i to take silence her to mean that ?schema does not exist ?

15:57:40 <relaxedguy> lunchtime

16:00:05 * las waves as she's muted....

16:00:18 <jang> seth: that's correct

16:00:44 * Morbus waves to las.

16:01:28 <Seth> jang, wow i would have expected that the w3c would not have left such a important class go undescribed :(

16:02:25 <Seth> i wonder what the thinking was that they did not put that class into some standard schema

16:02:39 <Seth> isnt it even defined in daml ?

16:04:04 <jang> we had a long chat about it at the last rdf core f2f

16:04:21 <Seth> and ?

16:05:22 <sandro> I imagine the problem is there's no consensus on what a Namespace is.

16:05:57 <Seth> kind of wierd, considering we put them in every piece of rdf we write

16:06:16 <sandro> You may. I never touch namespaces -- just prefixes. :-)

16:06:41 <Seth> how do you assicoate the prefix with the namespace uri ?

16:06:42 <sandro> (except when I use XML, which I do from time to time)

16:07:28 <sandro> ns + "#" + something == identifier

16:07:49 <Seth> im after the bloody triple itself .. obviously i can make one up ... but since it is such an important tiple, i somehow dont feel adequate to the task

16:08:38 <Morbus>http://kairosnews.org/modules.php?op=modload&name=News&file=article&sid=549&mode=thread&order=0&thold=0

16:08:39 <dc_rdfig> D: http://kairosnews.org/modules.php?op=modload&name=News&file=article&sid=549&mode=thread&order=0&thold=0 from Morbus

16:08:43 <Morbus> ack. poop.

16:08:44 <Seth> sandro, no not how to make a uri ... what is the official triple for:

16:08:47 <Seth> <http://www.w3.org/1999/02/22-rdf-syntax-ns#> type namespace.

16:08:56 <Morbus> D:|A Brief Summary of Blosxom, Peerkat, and AmphetaDesk

16:08:56 <dc_rdfig> titled item D

16:09:08 <Morbus> D::'ve been playing with blog tools and desktop news aggregators lately, so I thought I'd post some short summaries of what I've seen and what I've run into as problems.

16:09:08 <dc_rdfig> added comment D1

16:09:16 <jang> officially, there isn't one.

16:09:22 <jang> danbri wants "schema" or something

16:09:30 <jang> to tie together with "isdefinedby"

16:09:31 <Seth> in other words, the term namespace is never defined in any rdf document that i can find

16:09:33 <sandro> That would depend what you mean by "namespace". Pick some meaning, and some URIRef for it.

16:09:38 <jang> but don't expect me to advocate the idea.

16:09:59 <sandro> I don't think namespaces belong in RDF documents. They're an XML thing.

16:10:06 <Seth> namespace is obviously that which is identified by a namesapce uri

16:10:19 <sandro> But what is that thing? No one knows!

16:10:47 <sandro> (because there is no such thing -- the question is fallacious -- the namespace URI is NOT an identifier. It's just a prefix string.)

16:11:23 <sandro> If you want to use it as an identifier (as I do, in XML), that's fine -- but you need to define that. W3C does not.

16:11:30 <Seth> <http://www.w3.org/1999/02/22-rdf-syntax-ns#> is not a uri ?

16:12:12 <bijan> sandro, you need to distinguish the use of uris for ids of XML namespaces and the use (misuse) in RDF (RDF/XML especially)

16:12:23 <sandro> "is a URI" is kind of silly. That's like "Sam the Green Alien from Canada" is an identifier, but it's a kind of nonsense one -- I certainly have no idea what/who it might identify.

16:12:57 <bijan> Quite clearly, I believe, an XML Namespace URI identifies a set of names and *not* a prefix abbrev of a bunch of uris.

16:13:03 <Seth> it identifies an instance of the class namespace

16:13:24 <bijan> Well, I don't know abou that. That doesn't tell me anything at all.

16:13:29 <sandro> Bijan, the set of names is simply the set of strings with that as a prefix, no?

16:13:33 <bijan> No.

16:13:44 <bijan> Actually.

16:13:49 <bijan> That's why a qname isn't a uri

16:14:09 <bijan> And qnames aren't rdf identifiers.

16:14:14 <sandro> Ah, okay, you're being more precise about XML.

16:14:16 <bijan> Yes.

16:14:17 <bijan> That's all.

16:14:34 <bijan> You're quite right that there are no namespaces in RDF.

16:14:42 <bijan> NO, XMl namespaces certainly.

16:15:03 <sandro> The qname is the pair of (namespace name, element name) -- the namespace name doesn't identify anything, though, for that formulation.

16:15:43 <bijan> IIRC the spec, the namespace name is suppose to identify a set of names.

16:15:54 <bijan> It doesn't make it retrivable.

16:16:04 <bijan> But it's conceived as naming, well, a set.

16:16:17 <sandro> It's neither retreivable nor limitted in any way -- it's hard for me to see that as a set.

16:16:32 <Seth> sandro, yes withing a uri like <http://xmlns.com/foaf/0.1/name> the "http://xmlns.com/foaf/0.1/" part is not anyting , it is just a part of the uri.

16:16:35 <bijan> Well, no sets are retrievable, so I don't see why that matters.

16:16:43 <Seth> me thinks that is another issue

16:17:24 <sandro> I think we're in about the same territory, Bijan -- any arguing would be within the territory of fuzziness in my brain, which I don't feel much need to defuzzify right now.

16:17:26 <bijan> And that the set is specified by "all the lnames that appear in some qname as the second element" is a definite limitation.

16:18:07 <sandro> On the other hand, I'd love to argue about my axiomatic semantics for "positive triples" since I can't get them working right.

16:18:15 <bijan> Sure, but even if they don't name, they aren't prefix abbreviations of uris :) Tis all.

16:18:25 <bijan> Oh dear.

16:18:29 <bijan> Uh...meeting time! :)

16:18:34 <sandro> ha ha ha

16:18:52 <Seth> but if i make a rdf node like <RDF:description ..... RDF:about="http://xmlns.com/foaf/0.1/" ... > it denotes a set of names, like bijan said.

16:19:13 <bijan> I didnt' say that.

16:19:16 <sandro> [ You can make (nsname, elementname) to "$nsname$sepchar$elementname". Of course "#" isn't the right sepchar, really.

16:19:45 <bijan> Sandro, of course you can make such a mapping.

16:19:50 <bijan> RDF/XML *does* make a mapping :)

16:20:24 <bijan> But the same uri can be generated by multiple, non-equivalent qnames.

16:20:37 <bijan> Which are considered distinct idenitifers in XML.

16:20:49 <bijan> Hence, a qname *isn't* an abbreviation of a URI.

16:20:54 <sandro> How, if $sepchar is not a urichar? (eg "}" or "'")

16:21:07 <bijan> Oh. But then it's not a uri.

16:21:12 <bijan> SO it's not an rdf id.

16:21:12 <sandro> ah, yes.

16:21:27 <sandro> So, "#" is a pretty good compromise.

16:21:46 <bijan> Well, actually I beleive the compromise is to concat without a separator.

16:22:00 <bijan> And not to worry, as RDF doens't care about qnames anyway :)

16:22:14 * Seth defines his own class of namespaces

16:24:09 <sandro> Does http://...foo mean the same thing as http://...foo# Hrm. I guess that depends on the media type.

16:25:01 <Seth> depends upon the rdf header

16:25:42 <Seth> in other words what you put in xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"

16:36:44 <JibberJim> Bookcrossing doesn't have the problem with needing to know the book before you're out, you can pre-get numbers...

16:37:27 <JibberJim> via http://bookcrossing.com/labels

16:38:00 <JibberJim> (only the part after the - is significant apparently, you can make up the first 3 numbers apparently.)

16:59:50 <JibberJim> Does anyone know if you're supposed to be able to use the Mozilla RDF parser in a webpage, and if not, why not?

17:00:53 <danbri> "in" a web page?

17:01:05 <danbri> I don't tihnk it can slurp RDF out of HTML

17:01:35 <JibberJim> Yeah, no, in some javascript that is in a webpage (like the foafmatic demo where I use my own rdf parser) I'd like to parse and query some RDF.

17:03:48 <danbri> ...and call their parser via XPCOM? I did that once, I think.

17:03:50 <danbri> Long while ago.

17:04:01 <danbri> see http://www.mozilla.org/rdf/doc/faq.html for some examples

17:04:29 <JibberJim> Yeah, I took one of them, and I just get permission denied errors even creating the RDF parser object.

17:09:58 <danbri> yeah, is a bit paranoid

17:11:48 <JibberJim> It seems to me the security policy of mozilla is "disallow everything" then we'll be safe...

17:23:16 <eikeon> logger, pointer?

17:23:16 <eikeon> See http://ilrt.org/discovery/chatlogs/rdfig/2002-07-11#T17-23-16

17:30:51 <eikeon> Seth, where did you end up re: trying to parse jibbers foaf?

17:31:13 <JibberJim> Bug in my foaf not redfoot

17:32:16 <eikeon> JibberJim, thanks.

17:34:11 <eikeon> Redfoot's parser... now actually in rdflib... has been updated to that latest working draft. So is a bit more strict about attributes being qualified.

18:14:00 <Seth> eikeon, his new one worked fine ... after he corrected it :)

18:22:38 <Seth> eikeon, the extra strictness really works against me here ... i mean the rdfs schema document wont even parse

18:28:57 <rreck> NSA's yearly electrical bill is more than $21 million.

18:29:05 <Seth> weee .. how do we become members of the W3R ... count me in !!

18:49:53 <eikeon> Seth, I agree that the extra strictness is a pain when there are not versions of rdf-schema/rdf-syntax that conform to the spec. The goal of rdflib's rdf/xml parser is to stick as close to the latest WD as possible to help identify such issues. I will better document this and can add an optional mode for the parser to be less strict if need be.

18:50:37 <danbri> the rdfs doc will be fixed. please don't use it as an excuse to encourage parsers to be gentle.

18:50:56 * JibberJim thinks "get what you can mode" is always useful in a parser.

18:51:11 <JibberJim> but is possibly unique...

18:52:15 <Seth> eikeon, would it be possible to have a strict\forgiving option in the parser .. otherwise i fear i will be forced to chop at your code

18:52:49 <eikeon> For now I have just made the mods to rdf-syntax rdf-schema that I needed and put them up at: http://redfoot.net/2002/07/08/

18:54:06 <eikeon> I have a feeling the parser will have to support such an option due to customer demand... but it will not be the default.

18:54:08 <Seth> but i fear there are zillions of rdf files all over the place that i want to read .. that doesnt solve ~my~ problem

18:55:18 <Seth> that's kewl ... as long as i can set the mode when i call the parser

18:58:38 <eikeon> Are RDF/XML files that conformed to the last recommendation going to be distinguishable from RDF/XML files that follow the upcoming Recommendation?

18:59:04 <dajobe> I wouldn't think so

18:59:07 <danbri> they'll glimmer slightly in the moonlight

18:59:18 <dajobe> but those are dark glimmers ;)

18:59:29 <Seth> and break when read into a strict parser

19:00:47 <eikeon> Almost seems as they should be... is there a short answer to why there will not be?

19:00:58 <dajobe> eikeon: do you use any extra tests for your parser, beyond the rdf test cases wd?

19:01:31 <eikeon> No... have been following the approved test cases that are in the manifest.rdf

19:01:39 <dajobe> that's what I wondered

19:01:47 <dajobe> those tests don't claim to be comprehensive

19:01:51 <dajobe> just address particular points

19:02:43 <eikeon> Yes, I do not have a comprehensive set of test cases yet... but that will not make the parser stricter than it should be ;)

19:02:44 <dajobe> I tried to check in some more tests, for eg. ntriples, but I can't do that just now. That exercises the parser a bit more

19:03:02 <dajobe> if you want to try my raptor rdf/xml -> ntriple tests, feel free

19:03:20 <dajobe> from http://cvs.ilrt.org/cvsweb/redland/raptor/tests/

19:03:20 <eikeon> Sure, would love to. Pointer?

19:03:24 <eikeon> Thanks

19:05:13 <eikeon> What does raptor do re: strictness issue? Is it forgiving by default? Not at all? or optional mode?

19:05:32 <dajobe> strict

19:05:42 <dajobe> in the sense it tends to abort(). I've been more gentle now

19:05:46 <dajobe> will add a lax mode

19:05:51 <dajobe> just like others - ARP already has this

19:12:26 <eikeon> dajobe, Wouldn't happen to have a .tgz with the test cases in it that I could grab? (or a manifest.rdf for them?)

19:13:47 <dajobe> hmm, they are in the raptor nightly snapshot

19:14:07 <dajobe> . http://www.redland.opensource.ac.uk/dist/snapshots/source/raptor-0.9.6.tar.gz

19:14:11 <dajobe> no manifest yet

19:14:13 <eikeon> That should work... thank you.

19:14:45 * Seth wishfully types in http://lists.w3.org/Archives/Public/www-archive/2002Jul/0021.html to his agent to see if it can read

19:15:41 * Seth it can't :(

19:16:51 <Seth> eikeon, is your parser supposed to extract rdf from html files ?

19:17:25 <eikeon> It will look into xml files... but have not yet added a layer for looking into non xml.

19:17:41 <Seth> how about xhtml ?

19:17:47 <eikeon> Looks like there will be another release soon to add these features :)

19:17:57 <eikeon> It should work with xhtml.

19:18:17 <Seth> so it can read in the foaf document? ... humm ill try it

19:20:11 <Seth> KEWWWEL ... it worked :))))

19:22:13 * eikeon :)

19:22:15 <dajobe> rdf parsers should/can search for rdf:RDF blocks in xml (such as xhtml, svg). I've added it to mine, I think ARP does it too in some mode

19:23:31 <danbri> cool

19:23:57 * DanCon isn't sure how cool this is.

19:24:10 <DanCon> e.g. be careful doing that in my XSLT documents!

19:24:17 <sbp> $ GET http://lists.w3.org/Archives/Public/www-archive/2002Jul/0021 | python -c "import sys, re; print re.compile('<rdf:RDF.*?/rdf:RDF>', re.S).findall(sys.stdin.read().replace('&lt;', '<').replace('&gt;', '>'))[0]"

19:24:24 * JibberJim 's parser just generates triples of XHTML...

19:24:28 <danbri> this japanese rss feed is a good test case for unicode etc -- http://www.kanzaki.com/info/rss.rdf -- and a nice example of XSLT + CSS + XHTML + RSS + RDF combination...

19:26:15 * JibberJim gets 110 triples of that in nice shiny unicode...

19:26:23 <dajobe> rendered: http://www.redland.opensource.ac.uk/rss/?parser=repat;box=yes;uri=http%3A%2F%2Fwww.kanzaki.com%2Finfo%2Frss.rdf

19:26:34 <sbp> argh, needs "replace('&quot;', '"')" too. nasty hack

19:26:38 <dajobe> or try the feed in your browser

19:26:50 <JibberJim> redland got 111 :-(

19:27:19 <dajobe> lol

19:27:39 <dajobe> right and ARP gets ...

19:27:45 <dajobe> 111

19:27:46 <dajobe> :)

19:27:46 <eikeon> DanCon, Do you have an example of your XSLT documents? Why is caution needed for them?

19:27:46 <JibberJim> Can I get an NTriple or something so I can compare dajobe?

19:28:05 <dajobe> the rdfdump program in raptor can do that

19:28:10 <JibberJim> Ah no I don't I get 111, I trusted the status, which is 1 less!

19:28:22 <DanCon> sure, eikeon...

19:28:42 <DanCon> take dia2owl, which I just chumped...

19:28:45 * JibberJim now wonders if it's worth announcing his parser on www-rdf-interest if it's getting random ones correct...

19:28:52 <DanCon> ... http://www.w3.org/2002/03owlt/umlp/dia2owl.xsl

19:29:01 <dajobe> how does it do on the test cases JibberJim?

19:29:04 <DanCon> <xsl:template match="/">

19:29:05 <DanCon> <r:RDF>

19:29:05 <DanCon> <xsl:apply-templates/>

19:29:05 <DanCon> </r:RDF>

19:29:05 <DanCon> </xsl:template>

19:29:12 <eikeon> Ah... I see :)

19:29:20 <DanCon> I'd expect an RDF parser to get confused about <xsl:apply-templates/>

19:30:02 * eikeon should have been able to figure out the issue without an example... but now it is abumdantly clear :)

19:30:04 <dajobe> it ought to be told to scan, since it can't tell that it is allowed to assume the optional <rdf:RDF> is missing

19:30:06 <dajobe> I think

19:30:15 <DanCon> it's likely to spit out { _:something rdf:type xsl:apply-template }, treating <xsl:apply-templates/> as a typednode, when I didn't mean to assert any such thing.

19:30:18 <dajobe> I meant, optional outer <rdf:RDF>

19:30:27 <JibberJim> I've never done them dajobe as I've only ever thought of it as an internal toy parser I can use in javascript.

19:31:00 <sbp> one could simply check for the document being XSLT, and leave well alone if so

19:31:10 <dajobe> that's not scalable

19:31:32 <DanCon> scalable solutions are hard to come by here...

19:32:02 <dajobe> I'd prefer not to add checks for lists of namespaces to an rdf parser unless really necessary

19:32:06 <DanCon> ... TimBL has suggested looking for a schema for the root element; the schema should tell you whether it's RDF transparent (ala XHTML) or not (ala XSLT)

19:32:23 <dajobe> some are useful to do, such as daml (maybe for daml=>owl)

19:32:37 <sbp> even XHTML may not be RDF transparent: what if I put a bit of XML RDF in a blockquote? am I asserting it?

19:33:11 <DanCon> yes, the XHTML schema should point out blockquote as an exception, I think timbl would say.

19:33:44 <sbp> good point. I wonder if there are any machine readable schemata that let you say things like that around, though?

19:34:00 <DanCon> "With that exception, it clearly mekes sense to allow any language which has the concept of a sentence -- maybe any language at all - to allow sentences from other languages to be included anywhere where a sentence of its own could go. This should be a generic feature of XML schemas." -- http://www.w3.org/DesignIssues/Meaning.html

19:34:42 * sbp marvels at how quickly DanC found that quote

19:34:58 <DanCon> yeah, but it's not the one I was looking for.

19:39:08 <sbp> DanC: 404 at http://www.w3.org/2000/elxml/ from http://www.w3.org/2000/08/jr456

19:39:42 <sbp> now I think of it, perhaps it should point to http://www.w3.org/2000/08/elxml/xml-mode.el

19:50:38 * Seth wonders what "to allow sentences from other languages to be included anywhere where a sentence of its own could go. " actually means

19:51:48 <Seth> I eat means "Ich essa" in Deutch. .... something like that?

19:53:38 <Seth> {i eat whatever} (means the same as) "Ich essa wassever". for those who like triples

19:59:49 <DanCon> sbp, try http://www.w3.org/2000/08/elxml/xml-mode.el

20:00:01 <DanCon> phpht. too slow.

20:22:41 * eikeon thinks it would be nice to have some bit of glimmer to distinguish RDF conforming to current Recommendation from the future recommendation... to help aid in existance of strict parsers. Else it seems people are almost forced to write and use parsers that do not conform to the latest WD re: strictness.

20:24:20 <Seth> like for example where would it glimmer?

20:24:43 <JibberJim> in the rdf namespace - how else do you do versioning?

20:25:11 <Seth> great then all our triples will suddenly not match ... yuckey poo

20:25:57 <JibberJim> Yes, I'm not saying it's a workable solution, just the only versioning capability there is isn't it?

20:25:58 <Seth> cours their quads would match :))

20:26:01 <eikeon> Triples could still match if only the rdf:RDF namespace glimmered?

20:26:35 <Seth> cept a triple that was based on that namespace ... like rdf:type

20:26:40 <JibberJim> They're kind of important part of the triples thought aren't they.

20:26:50 <JibberJim> s/the/lots of the/

20:27:33 <eikeon> a new rdf:RDF could reference rdf:type from the current rdf: namespace no?

20:28:34 <Seth> how?

20:29:54 <Seth> me thinks that by binding the naespace into the triple, the w3c has kind of painted themselves into a corner here

20:30:39 <eikeon> rdf:RDF does not end up in any triples tho?

20:30:58 <eikeon> r1 = current rdf: namespace

20:31:07 <eikeon> r2 = new rdf: namespace

20:31:26 <Seth> but the namespaces designated by the rdf header do end up in the triples, change the namespace and you change the triples

20:32:08 <eikeon> then <r2:RDF> <r1: > <r1: > etc </r2:RDF>

20:32:43 <JibberJim> That would work wouldn't it...

20:33:18 <Seth> yep, that might work ... doubt you'll get them to go for it

20:33:21 <eikeon> In short, all rdf:foo's would stay in the current namespace except for RDF... which would be the bit that glimmers.

20:34:23 <Seth> ... wait a minute ... paste some exact rdf here .. im not sure what you mean

20:36:28 <eikeon> [[[

20:36:30 <eikeon> <?xml version="1.0" encoding="UTF-8"?>

20:36:30 <eikeon> <r2:RDF

20:36:30 <eikeon> xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"

20:36:30 <eikeon> xmlns:r1="http://www.w3.org/1999/02/22-rdf-syntax-ns#"

20:36:30 <eikeon> xmlns:r2="http://www.w3.org/2002/07/11-rdf-syntax-ns#"

20:36:31 <eikeon> >

20:36:33 <eikeon> <r1:Description r1:about="http://rdflib.net/">

20:36:35 <eikeon> <rdfs:label>RDFLib.net</rdfs:label>

20:36:37 <eikeon> </r1:Description>

20:36:39 <eikeon> a</r2:RDF>

20:36:41 <eikeon> ]]]

20:39:41 <Seth> hey that works for me ... the rdf:RDF was a bit redundant anyway .. now we would get a real syntactic versioning that would preserve the triples themselves :)) .... you should propose this to the working group

20:41:00 <sbp> ooh: "The group is (2002/06) looking at schema annotation ways of declaring html:href to carry the significance of an xlink. These are known as "hlink"." - http://www.w3.org/DesignIssues/XLink

20:42:21 <dajobe> rdf:RDF redundant? it's the XML document element, required by XML

20:42:24 <dajobe> and SGML

20:48:41 <Seth> eikeon, thing is that this versioning switch is not the only reason to have a strict/forgiving option

20:48:52 <niq> SGML would permit rdf:RDF to be omitted

20:49:41 <Seth> rdf:RDF is redundant because it says the same thing on both sides of the semi colon

20:50:47 <JibberJim> but it could be MOO:RDF and would mean the same.

20:51:08 <Seth> but i guess the lower case rdf is the namespace and the upper case is the element name .. so eikon is just changeing the optional namespace identifier

20:51:26 * niq is away: please don't disturb, I'm having a nervous breakdown

20:52:21 <Seth> actually, eikeon, the uri specified for xmlns:r1 and xmlns:r2 should be the same, that way you preserve the integrety of the triple .. huh?

20:57:32 <JibberJim> No, they have to be different as that's how you know that the root element is different so you should use a different parser - the triples are all the same as rdf:RDF never appears in a triple.

20:58:07 <eikeon> The r2:RDF element above does not end up in any of the generated triples. And the point is to have r2!=r1 so that r2:RDF is a newly defined element corresponding to a new version of RDF... yep, what JibberJim said.

21:00:01 <Seth> huh, i must be missing something .. a triple that uses http://www.w3.org/2002/07/11-rdf-syntax-ns# will never match a triple that uses http://www.w3.org/1999/02/22-rdf-syntax-ns# .... ?

21:00:23 <JibberJim> The first namespace you've got there will never be seen in a triple.

21:01:28 <Seth> oh i see what your doin

21:05:17 <Seth> id make 'r1' just be 'rdf'

22:30:38 * DanCon works on Montreal travel stuff... trying to get cwm to compute my travel request based on records I already keep...

22:55:00 <DanCon> hey... is log:uri broken?

22:56:01 <DanCon> no...

22:56:44 * DanCon realizes I'd forgotten the --think option. duh.

23:05:54 * DanCon finds a bug with log:uri and str:endsWith

23:06:15 <DanCon> log:uri is making ascii strings, but N3 literals are unicode strings; endsWith fails as a result.

23:08:53 <DanCon> well, it's starting to work..

23:08:55 <DanCon> [ a mt:TravelRequest;

23:08:55 <DanCon> mt:for [

23:08:55 <DanCon> k:destination__RoundTrip <status-formal.rdf#_rdfxg34>;

23:08:55 <DanCon> k:passengers <status-formal.rdf#who_connolly>;

23:08:55 <DanCon> mt:forEvent <status-formal.rdf#L3846> ] ].