Semantic Web Interest Group IRC Chat Logs for 2003-11-19

01:33:18 <mdupont> mdupont is now known as zzz-md-zzzz

01:33:53 <zzz-md-zzzz> zzz-md-zzzz is now known as zZz-md-ZzZz

02:04:32 <Davey> Davey is now known as D[a]vey

03:09:29 <_deusx> _deusx is now known as deusx

05:05:02 <deusx> deusx is now known as _deusx

06:20:22 <CaptSolo> hi dajobe :>

06:20:36 <dajobe> hi

06:38:28 <CaptSolo> what's new?

06:39:45 <deltab> my discovery of how to turn a cube into a dodecahedron

06:42:14 <CaptSolo> hm? how?

06:43:48 <deltab> add a new edge to each face, breaking it into two faces and turning the faces adjacent to the new vertices into pentagons

06:44:20 <deltab> well, all of the faces

07:09:52 <zZz-md-ZzZz> zZz-md-ZzZz is now known as mdupont

08:32:24 <D[a]vey> mattmcc: you about?

08:32:45 * mattmcc nods.

08:33:31 <D[a]vey> I'm just curious if I need to do something like ?d <rdf:type> <d:dublincore> in the RDQL to get the elements in the d: namespace

08:34:07 <D[a]vey> or if I use the ?a (from ?a <rdf:type> <a:Annotation>) still

08:35:54 <mattmcc> I'm not sure I follow.

08:39:10 <D[a]vey> I think I'm still hung up on my RDF as XML stuff

08:40:43 <D[a]vey> wow...

08:40:47 <D[a]vey> RDQL *rocks* :D

08:40:55 <D[a]vey> check this out: http://davey.is-a-geek.com/php-cvs/pear/XML_Annotea/Annotea_Server.php :D

08:41:13 <D[a]vey> hmm, missing one thing

08:41:45 <D[a]vey> got it :)

08:42:24 <D[a]vey> btw mattmcc, if you missed this: RdqlEngine::writeQueryResultAsHtmlTable($result); - its awesome :)

08:42:40 <mattmcc> Yeah, it has some nice output routines.

08:45:03 <D[a]vey> mattmcc: I'm having some problems trying to get to the body content when the data is inline...

08:45:30 <D[a]vey> mattmcc: http://www.w3.org/2001/Annotea/User/Protocol.html#PostABody <--- see the example?

08:46:29 <D[a]vey> hmmm....

08:46:33 * D[a]vey ponders

08:53:02 <Wack> well, display the rdfxml as triples :]

08:53:20 <D[a]vey> (?a <h:Body> ?body) <--- any idea why that doesn't grab the content of <h:Body> ?

08:53:24 <Wack> sure

08:53:51 <D[a]vey> I mean, unless its a coincidence, (?a <ns:element> ?foo) has worked till now...

08:54:16 <Wack> but it's better if you understood rather than have me give you a working query no? ;)

08:54:42 <D[a]vey> Wack: if you could explain, thats best for sure :)

08:54:48 <Wack> ?body is now a blank node, whereas is was a resource with an uri before

08:55:06 <Wack> you'll have to look into properties about ?body now to get more information about ?body

08:55:23 <Wack> h:Body is a property of ?body, not of ?a

08:55:35 <Wack> well, display the rdfxml as triples

08:55:51 <Wack> I think you'll see what's different between the examples

08:56:35 <D[a]vey> hmm, you know whats worrying? the query from earlier: SELECT ?s, ?p, ?o WHERE (?s, ?p, ?o) doesn't show the content of <h:Body>, but anyways onto the triples.... how do I do that? :)

08:57:33 <Wack> well, SELECT ?s, ?p, ?o WHERE (?s, ?p, ?o) is a crude way to achieve the same

08:58:04 <Wack> I thought there were convenience methods to 'dump' a model as a html table right? you used it yesterday as well

08:58:28 <D[a]vey> ah, yes, thats the triples? :)

08:58:30 <Wack> (btw, the reason it doesnt show the contents is propably because it's a xml literal)

08:59:22 <D[a]vey> fyi, I worked it out (kinda) from what you said, but I don't understand why, so I'll look at the triples now :)

09:00:09 <Wack> see the Subject

09:00:09 <Wack>

09:00:10 <Wack> Predicate

09:00:10 <Wack>

09:00:10 <Wack> Object

09:00:16 <Wack> columns?

09:00:16 <D[a]vey> yes

09:00:40 <Wack> well, a triples is one row in that table

09:00:42 <Wack> -s

09:00:42 <D[a]vey> I see how Object changes to bNode2, and then the subject changes...

09:01:12 <D[a]vey> so every blank node you encounter, it becomes the subject?

09:01:16 <Wack> no, no

09:01:31 <Wack> uhmm

09:01:44 <D[a]vey> then why? I'm still stuck on this parent, child XML tree idea... trying to get my head past that.

09:01:53 <Wack> ah!

09:01:53 <mattmcc> RDF triples aren't ordered.

09:02:01 <Wack> rdf model is _not_ a tree

09:02:04 <mattmcc> Remember, each statement is independent.

09:02:05 <Wack> it's not like xml

09:02:27 <D[a]vey> Wack: right, this is where my head is getting screwed

09:02:36 <D[a]vey> D[a]vey is now known as Davey

09:02:52 <Wack> right, well, each triple consists of three nodes (the subject, predicate and object)

09:03:14 <Wack> those nodes are not unique to each triple, triples share nodes

09:03:27 <Wack> one triple's object, can be another triple's subject

09:03:34 <Wack> that's how you traverse the data

09:03:56 <Wack> <bnode1> <ns:prop1> <bnode2>

09:04:06 <Wack> <bnode2> <ns:prop1> <bnode3>

09:04:15 <Wack> 2 triples

09:04:29 <Wack> but you could kinda traverse it in your mind like:

09:04:35 <Wack> <bnode1> <ns:prop1> <bnode2> <ns:prop1> <bnode3>

09:04:56 <mdupont> mdupont is now known as Work-md-Work

09:04:58 <jql> looks suspicious

09:05:23 <Wack> but, there are better introductions written down

09:05:25 <mattmcc> Davey: Have you looked at an RDF graph visualization before?

09:05:41 <Davey> mattmcc: I've looked at something... not sure if its what you mean.

09:05:48 <Wack> better than with which I can come up with while not having had my coffee yet ;)

09:05:58 <mattmcc> Try punching your FOAF into the RDF validator and asking for a graph.

09:06:04 <Davey> I guess its really time to wade through the RDF Primer?

09:06:06 <mattmcc> (Graph and triples, rather, so you can compare them)

09:06:16 <Davey> I've started that primer so many times now... heh

09:06:32 <Wack> skim it, look at the pictures ;

09:06:33 <Wack> ;)

09:06:40 <Davey> sounds like playboy...

09:06:44 <jql> Davey: don't worry. it wasn't until I read all the documents -- not just the primer -- that I started to grasp it

09:07:04 <Wack> yeah, you need an aha-erblebnis of kinds

09:07:12 * mattmcc nods.

09:12:53 <Wack> it's also very essential to know that rdfxml and the rdf model are two seperate things

09:13:31 * jql uses rdf+xml exclusively, but only after understanding the model. that model is there for a reason

09:14:28 <Wack> rdfxml is 'just' a serialization format that uses xml to format rdf model data (triples) and has all kinds of quirks and shortcuts that can may confuse people that try to see it as 'regular' xml

09:14:49 <jql> the predicate attribute shortcut, for one

09:15:25 <Davey> ok, I think I understand this...

09:15:33 <jql> <foo ex:bar="baz"><ex:bar><Messes><withYourMind/></Messes></ex:bar></foo>

09:16:49 <Wack> the revised rdfxml syntax spec is also greatly improved by accompanying every example with what triples it results in

09:18:00 <Davey> if I understand these two triples (looking at them in order, though as mattmcc says, theres no order) 9 says that bNode1 contains bNode2, 10 says, bNode2 contains <h:Body>

09:18:24 <Davey> but couldn't it then deduce, that bNode1 contains <h:Body> ?

09:18:36 <mattmcc> It doesn't, though.

09:18:36 <Wack> well, 'contains' is a bad choice of words

09:18:58 <mattmcc> And, yeah, vocabulary can be tricky.

09:18:58 <Davey> Wack: what would be better choice?

09:19:16 <Wack> references/refers to, or targets

09:19:37 <Davey> cause how I'm seeing this now, is again a tree, is <bNode1><bNode2><h:Body /></bNode2></bNode1>

09:20:09 <mattmcc> For arguments sake, you could rearrange that so that h:Body isn't 'inside' bNode2, but the triples would be the same.

09:20:10 <Wack> well, that's wrong when it comes to rdf ;)

09:20:25 <jql> I see <node><arc><node/></arc></node>

09:20:37 <Wack> it even could be declared in another file ;)

09:21:29 <Davey> does <h:Body> have *any* relationship to bNode1?

09:21:38 <Wack> no

09:22:15 <Davey> OK. Then once again my fundamentals are screwy.

09:22:34 <Wack> btw, that's another issue; <h:Body> is not it's contents

09:23:10 <Wack> h:Body is the relationship (predicate) between two nodes (bnode2 and the xhtml)

09:23:37 <Davey> I was under the impression that for instance, if you have these triples (english, forgive my vocabulary) "Davey is 19" "Davey is the same age as his cat" you could then deduce that "19 is the age of Daveys Cat"

09:23:59 <mattmcc> Only with the additional help of some schema work.

09:24:07 <jql> hrm. you're hitting the inference level stuff. see cwm. :/

09:24:29 <Davey> mattmcc: hmm

09:24:31 <Wack> dont even go there as long as you're groping with the basics ;)

09:24:38 * mattmcc grins.

09:24:54 <Davey> ok, so I need to forget the inference stuff?

09:24:55 * jql uses alot of rdf, but doesn't go through inference like that

09:25:05 <mattmcc> Yeah, leave inference off to the side for now.

09:25:22 <Davey> and just remember that triples can share stuff, but are not necessarily related?

09:25:42 <Wack> try to find some tool that can visualize rdfxml

09:25:54 * mattmcc points at the validator.

09:26:14 <jql> <You><caretakerOf><Cat><named>Fluffy</></></></>

09:26:33 * jql confoozles the issue a bit

09:26:56 <jql> <You><caretakerOf rdf:resource="http://fluffy.the.cat"/></>

09:27:19 <jql> <Cat rdf:about="http://fluffy.the.cat"><named>Fluffy</></>

09:27:31 <Davey> can you bring up http://www.w3.org/RDF/Validator/ARPServlet.tmp/servlet_58683.png ?

09:27:42 <Wack> yes

09:28:44 <Wack> looks like the code behind it needs some whitespacestripping :]

09:28:48 <Davey> see, what I'm *still* seeing here is that parent -> child -> grandchild relationship (on the lowest branch), but I do see that the grandchild is only connected to the parent in that they share child...

09:29:24 <Wack> well, the rdfmodel is just the connections (the arrows and which way they point) between nodes

09:30:05 <Davey> so they're directional arrows, not connections?

09:30:09 <jql> xml is misleading, because parent -> child -> parent is both legal and common

09:30:12 <Wack> that 2 nodes are connected via another node is left up to other software (for instance rdql)

09:30:48 <Wack> yes, the connection is always one-way (a directional graph)

09:31:02 <Davey> *think* I'm getting this...

09:31:15 <mattmcc> Davey: http://www.w3.org/RDF/Validator/ARPServlet.tmp/servlet_58687.png is your FOAF, fwiw. :)

09:32:07 <Davey> so, with this in mind... how do I get to the CDATA of <h:Body>...

09:32:23 <jeen> sandro, you in?

09:32:34 <jql> implementation defined behavior there, Davey. :)

09:33:02 <Davey> gee... helpful :P

09:33:20 * Davey var_dumps

09:33:30 * jql was just trying out the W3C-approved wording of RTFM :P

09:33:44 <mattmcc> Well, something to note in the graph.. h:Body is only there as a predicate.

09:34:18 <Wack> is that a recommendation jql? :]

09:34:30 <jql> working draft

09:35:14 <jql> of course, http://www.w3.org/TR/RTFM/

09:35:30 <jql> current version is no doubt RTFM-20021018

09:36:32 <Davey> mattmcc: soemthing the FOAF graph illustrates, though I don't know the vocab, is how your FOAF is a node of mine... but is the same as mine. (talking about the relationship with the right-most oval)

09:37:14 * jql browses the http://www.w3.org/TR/rdf-syntax-grammar/ spec yet again... have it memorized soon

09:37:20 <mattmcc> Two statements in your FOAF have the same URI as their object.

09:37:27 <mattmcc> So the graph is illustrating that.

09:37:38 <Davey> yeah, what he said.

10:04:21 <swh> swh is now known as swh_afk

10:38:03 <Davey> Davey is now known as D[a]vey

10:48:44 <dajobe>http://www.schemaweb.info/

10:48:45 <dc_rdfig> A: http://www.schemaweb.info/ from dajobe

10:49:01 <dajobe> A:|SchemaWeb by Victor Lindsey

10:49:01 <dc_rdfig> Titled item A.

10:49:42 <dajobe> A:|SchemaWeb by Victor Lindesay

10:49:42 <dc_rdfig> Titled item A.

10:50:06 <dajobe> A:he seems to have invented an XML triples syntax

10:50:06 <dc_rdfig> Added comment A1.

10:50:55 * monkeyiq looks at site

10:51:01 <dajobe> nmg mentioned at http://www.schemaweb.info/forum/ShowPost.aspx?PostID=1

10:51:17 <nmg> so I noticed...

10:51:24 <dajobe> A:a slick looking & working site

10:51:24 <dc_rdfig> Added comment A2.

10:51:53 <nmg> "W3C guru and RDF expert"? well, gawrsh!

10:52:05 <dajobe> A:a [http://www.schemaweb.info/rss.aspx|RSS 1.0 feed] of schemas

10:52:06 <dc_rdfig> Added comment A3.

10:52:08 <dajobe> A:!

10:52:08 <dc_rdfig> Added comment A4.

10:52:49 <dajobe> A:a few mime types for the rdf content would be nice

10:52:50 <dc_rdfig> Added comment A5.

10:54:53 <swh_afk> swh_afk is now known as swh

10:55:08 <monkeyiq> hmm, no schemas for bibtex :(

10:56:13 <dajobe> that is thus a market opportunity

10:57:06 <monkeyiq> heh, it dawned on me that the bibtex thing would be a obvious extension for the FOAF stuff. Not having to hack in the names of authors and publishers each time.

10:57:13 <dajobe> A:a nice site

10:57:13 <dc_rdfig> Added comment A6.

10:57:39 <dajobe> well, there's always more you can capture but in this case maybe some way to convert to/from bibtex would be nice

10:57:52 <dajobe> should be stuff like that in prism already I think

10:58:00 <monkeyiq> though I suspect that I'd have to write latex and emacs support stuff for anyone to want to use such a schema

10:58:24 <monkeyiq> oh ok... I should check out prism.

10:59:12 <dajobe>http://www2.elsevier.co.uk/~tony/spec/rss/prism.html

10:59:12 <dc_rdfig> B: http://www2.elsevier.co.uk/~tony/spec/rss/prism.html from dajobe

10:59:18 <dajobe> B:|RSS1.0 prism module

10:59:18 <dc_rdfig> Titled item B.

10:59:29 <dajobe> maybe not

10:59:43 <monkeyiq> I was also considering if bibtex was in RDF then folks could start embedding their own bibtex in XMP/RDF in pdf/ps files for everyone. Nice to just drag a file into a new paper and have the citation generated from the embedded RDF.

11:03:11 <monkeyiq> hmm, also seems to have no reference to FOAF, which IMHO most coauthors are also fairly close FOAFs too.

11:03:38 <dajobe> or just friends

11:03:42 <dajobe> or supervisor/student

11:04:01 <monkeyiq> yeah, that last one might not imply as much friends :)

11:04:03 <dajobe> sometimes it just doesn't need modelling for citations

11:04:28 <dajobe> you'll rapidly get into the problems nmg & swh are familar with re peoples' names

11:05:15 <monkeyiq> I was looking at the FOAF linking so that I can consider it in reverse, so I can say "what papers did X write" given X is a URI for a friend.

11:06:04 <monkeyiq> seemed better to allow such linking explicitly in the RDF instead of dumping the char string for the name into the RDFbibtex file.

12:03:33 <arnarl> hi

12:54:01 <Wack>http://www.angelite.nl/aquamarine-v2/

12:54:01 <dc_rdfig> C: http://www.angelite.nl/aquamarine-v2/ from Wack

12:55:52 <Wack> C:|Aquamarine v2 API

12:55:52 <dc_rdfig> Titled item C.

12:57:01 <Wack> C: It's a revision of our older API, adding querying, remote models and model aggregation and renaming stuff here and there.

12:57:01 <dc_rdfig> Added comment C1.

12:57:09 <Wack> C: Comments are welcome!

12:57:09 <dc_rdfig> Added comment C2.

12:57:10 <shellac> does this mention which languages are supported?

12:57:14 * shellac dense

12:57:24 <Wack> no, it's language independant

12:58:02 <Wack> (the html is generated from CORBA idl file)

12:58:10 <D[a]vey> ok, I still don't quite see how to get to the uhm, object of the <h:Body> subject...

12:58:23 <shellac> aha - just saw the idl

12:58:55 <shellac> D[a]vey: it's deja vu all over again

12:59:23 <D[a]vey> shellac: hehe, you missed another debacle earlier

13:01:14 <shellac> aha - you can't via rdql

13:01:41 <shellac> did that come up earlier?

13:01:59 * shellac checking logs

13:02:21 <D[a]vey> I *can't*?

13:02:26 <D[a]vey> you mean: http://davey.is-a-geek.com/php-cvs/pear/XML_Annotea/Annotea_Server.php

13:02:33 <shellac> if you look you'll see:

13:02:42 <Wack> (annotation? <a:body> ?body), (?body <http:body> ?contents)

13:02:53 <shellac> oh - sorry, wrong body

13:02:55 <shellac> doh

13:03:20 <shellac> I thought you wanted the html body

13:03:33 <Wack> ah you already did that

13:03:37 <Wack> well, does RAP support xml-literals?

13:03:59 <shellac> have you looked at the triple dump?

13:04:13 <D[a]vey> If you refresh the url, I need to get at whats in <h:Body>

13:04:25 <D[a]vey> shellac: thats the thing... its not in there :/

13:05:55 <shellac> oh yes: you've go Literal: <zip>?

13:06:28 <D[a]vey> yup

13:06:31 <shellac> looks like a bug

13:06:36 <D[a]vey> ack :/

13:06:39 <D[a]vey> wait...

13:06:43 * D[a]vey commented something out...

13:07:22 * D[a]vey uncomments

13:09:13 <D[a]vey> odd, its already uncommented

13:11:07 <D[a]vey> oops, wrong copy.

13:11:22 <Wack> what part is commented?

13:11:25 <shellac> have you checked using something other than the html table result method?

13:12:36 <D[a]vey> shellac: yes, I've dumped the contents of the result

13:13:07 <Wack> my guess is that there is some problem with xml literals in the parser

13:13:11 <shellac> I see a commented // $this->_report_start_parse_type_literal();

13:13:23 <shellac> in the parser

13:14:09 <D[a]vey> it seems that the problem is at the case starting line 1448

13:14:17 <shellac> oh - but not corresponding method

13:14:48 <D[a]vey> theres an echo which throws an E_NOTICE and a commented out $this->_report_start_element()

13:15:14 <D[a]vey> you'll notice that IN_PROPERTY_PARSE_TYPE_LITERAL falls throw to the IN_XML case which is what I'm talking about.

13:16:47 <D[a]vey> good lord, what a mess he made of this.

13:17:19 <Wack> parsers are usually pretty strange code

13:17:36 <shellac> state machine code - ugh

13:18:12 <shellac> but case IN_PROPERTY_PARSE_TYPE_LITERAL falls through to IN_XML case

13:18:39 <D[a]vey> which just echos out.

13:18:48 <D[a]vey> which is *wrong*

13:19:04 <shellac> gotcha - yep

13:21:08 <shellac> so it looks like the code is missing - $this->_report_start_parse_type_literal() et al

13:21:34 <shellac> afaict

13:21:39 <D[a]vey> yeah...

13:21:44 <shellac> :-(

13:21:50 <D[a]vey> I have it in the original class on phpxmlclasses though

13:21:54 <D[a]vey> wonder if I can just rip it out... hehe

13:22:16 <shellac> that always works :-)

13:22:36 * D[a]vey diffs

13:29:13 <dajobe> that's cos it was based on repat

13:29:20 <dajobe> which did no parse type literal

13:29:23 <dajobe> and is otherwise ancient

13:30:07 <dajobe> C:an RDF API

13:30:07 <dc_rdfig> Added comment C3.

13:30:56 <dajobe> S:if I was making a new API, I'd s/model/graph/ and s/statement/triple/ everywhere

13:30:57 <dc_rdfig> Label S not found.

13:31:00 <dajobe> C:if I was making a new API, I'd s/model/graph/ and s/statement/triple/ everywhere

13:31:01 <dc_rdfig> Added comment C4.

13:32:12 <dajobe> C:since the RDF model is a set of triples, which can be drawn as a graph

13:32:13 <dc_rdfig> Added comment C5.

13:32:49 <D[a]vey> hmm... this is just too hard to hack in

13:33:07 <shellac> yes - I was just trying to work out how to do it

13:33:29 <D[a]vey> really, RAP made a *lot* of changes

13:33:51 <dajobe> C:*"Adding, removing and copying of statements." ... void add (in uTriple s, in boolean v) ... "Adds a new statement to the model. "* - umm?

13:33:51 <dc_rdfig> Added comment C6.

13:34:15 <D[a]vey> or rather a lot of changes have been added to RdfParser.php since RAP got a hold of it

13:34:44 <shellac> I think you need to a) start a new literal b) copy in xml c) report statement when it finishes

13:35:07 <dajobe> it's more complex than that

13:35:10 <dajobe> in-scope lang, base URI

13:35:14 <dajobe> XML C14N

13:35:15 <dajobe> ...

13:35:26 <dajobe> namespaces

13:35:36 <sbp> is canonicalization still a requirement?

13:35:46 <shellac> we're just trying literal support, not 100% compatibility :-)

13:35:46 <dajobe> yes

13:35:47 <sbp> oh, of course it is. I'm thinking about unicode normalization

13:36:00 <dajobe> which is nice but not a requirement

13:36:06 <sbp> yeah

13:36:10 <shellac> currently D[a]vey gets ""

13:36:33 <shellac> anything would be an improvement, I think

13:37:42 <Wack> C:I made a clear distinction between triples and statements; a triple is only statement if it's in the model/graph.

13:37:42 <dc_rdfig> Added comment C7.

13:38:06 <D[a]vey> hehe, well its not hugely broken, at least

13:38:27 <shellac> the hooks are there, I guess

13:38:49 <Wack> C:I was thinking of renaming Model to Graph as well, actually

13:38:49 <dc_rdfig> Added comment C8.

13:39:01 <dajobe> C:do it now, it helps people

13:39:01 <dc_rdfig> Added comment C9.

13:39:29 <Wack> I'm thinking of adding a BlankNode interface that inherits from Resource as well, so Property's cannot be blank

13:39:43 <shellac> jena uses both graph and model - just to make life interesting

13:39:46 <dajobe> C:a triple is still a triple [3](URI | Literal | bnode) whereer it is

13:39:47 <dc_rdfig> Added comment C10.

13:39:53 <dajobe> C10:a triple is still a triple [3](URI | Literal | bnode) wherever it is

13:39:53 <dc_rdfig> Replaced comment C10.

13:39:57 <Wack> and move the blank node 'id' attribute to that interface

13:40:11 <dajobe> shellac: but the graph's a sort-of internal thing, with warnings about not using it much as I recall

13:40:31 <dajobe> or "jjc wrote this, we don't understand it" :):)

13:40:38 <Wack> dajobe: "Adding, removing and copying of statements." ... void add (in uTriple s, in boolean v) ... "Adds a new statement to the model. " - umm? <- is that because of the triple/statement naming?

13:40:49 <dajobe> yes, and the 'v' - what's that

13:40:51 <shellac> yes - model is a wrapper around (one or more?) graphs

13:40:59 <D[a]vey> ok, I need to eat before I can carry on with this hackery, so bbs

13:41:29 <shellac> D[a]vey: see you later

13:41:59 <Wack> ah yes, volatility

13:42:46 <Wack> we make distinction between volatile and nonvolatile statements

13:43:05 <Wack> volatile statements disappear when the model/graph is closed

13:43:15 <dajobe> hmm

13:43:20 <dajobe> closed world ? :)

13:43:44 <Wack> we added this because of the way we mainly use it currently

13:44:42 <Wack> we use this database as a information abstraction layer in our website framework, volatile statements are kinda like 'request contextual' statements

13:45:01 <Wack> stuff like what node/page was/is being retrieved, what the current session is, etc

13:45:30 <Wack> and to connect request-specific output to the request-node in the graph

13:45:49 <Wack> when the framework closed the model when the http request is done, those all dissipate

13:46:57 <Wack> 2 clients that open the same model have each their own collection of volatile statements

13:47:49 <Wack> it's more like a psuedo-request-transaction

13:48:35 <Wack> internally it's just an aggregation of a in-memory and bdb graph, the in-memory graph created on graph open, and free'd on close again

13:49:25 <dajobe> ok

13:49:54 <dajobe> I'd also recommend avoiding the word context if you can

13:49:59 <dajobe> use grouping or statement sets or something else

13:50:05 <dajobe> modelIds, statementIds, ...

13:53:57 <Wack> 'context' was 'associatedModel' before, but that was just icky to use in client code :]

14:08:55 <Wack> thanks for your input dajobe, your comment on replacing model with graph pushed me over the edge to rename it as well; I tried to stick as close as possible to the older API, but since there are incompatible changes elsewhere it doesnt really matter anymore anyway

14:09:31 <Wack> just trying to find a good replacement name for 'Node::context' ..

14:10:57 <dajobe> I guess I have trouble seeing why; nodes aren't independent, they are solely parts of a triple, which is always (should be!) in a graph

14:11:37 <dajobe> so if you have a triple you should know what it's associated model(graph) is, since you have the triple

14:11:37 <Wack> it's because of the Resource:: centric statements methods

14:11:47 <dajobe> sure, that's like jena

14:12:22 <dajobe> just use graphRef then

14:12:56 <dajobe> which IDL are you using btw

14:13:03 <dajobe> logger, chump c

14:13:03 <dajobe> C:See [http://ilrt.org/discovery/chatlogs/rdfig/2003-11-19#T14-13-03|discussion]

14:13:04 <dc_rdfig> Added comment C11.

14:13:06 <Wack> and there is a dead space between creating nodes via the nodefactory to create a new statement with and asserting the statement

14:14:02 <Wack> i'm not sure if the IDL is coupled with the CORBA ORB version

14:14:36 <Wack> or whether there are noticeable differences between CORBA IDL and ISO's IDL

14:14:59 <dajobe> I guess I see IDL as a generic term. I use the SWIG IDL for example

14:16:58 <Wack> they all borrow from eachother

14:17:33 <Wack> I believe IDL was originally developed by OMG for CORBA

14:21:51 <dajobe> yeah, but for me, only SWIG targeted a decent set of languages

14:44:58 <Wack> is there any progress towards a W3C recommended RDF api, like DOM?

14:46:42 <dajobe> we discussed that last week

14:46:59 <dajobe> competitive apis at the moment

14:47:08 <dajobe> keeps you on your toes :)

14:52:58 <Wack> hehe, mysql 'bashing' haikus

14:53:03 <Wack> Once I had data

14:53:03 <Wack> My DBMS munged it

14:53:03 <Wack> But damn it was fast!

16:05:14 <_deusx> _deusx is now known as deusx

16:12:35 <roGer^work> anyone knows if jena2 has implemented or in rdql?

16:12:59 <libby> eh?

16:14:06 <jeen> ITHM if they have implemented "OR" in RDQL in Jena 2. Don't think so, but other people know better, prolly.

16:15:05 <roGer^work> k

16:16:00 <roGer^work> damnation

16:16:01 <roGer^work> :-)

16:16:29 <jeen> ...well, it's open source, so if you really need it... ;)

16:19:38 <Work-md-Work> Work-md-Work is now known as mdupont

16:29:00 <roGer^work> for example i would like to ask if a resource has one rdf:type or an other

16:29:58 <roGer^work> something like (?res, <rdf:type>, <http://my_type1>) OR (?res, <rdf:type>, <http://my_type2>)

16:32:02 <libby> oh sorry, I just didn;t understand

16:32:08 <libby> I don;t think so in jena no

16:32:37 <libby> best thing to do is multiple queries

16:32:58 <roGer^work> i know it didn't exist in jena1 but i was kinda hoping that maybe it would be implemented in jena2

16:33:34 <libby> I think andy's thinking about optional sections, which might do what you want, but not done yet I don;t think

16:34:28 <roGer^work> libby: yes i know what u mean, it would work in that example but i dont want to do that cause it would result in massive io from the database (atleast i think so) when the model-intersections are done

16:34:44 * roGer^work don't REALLY know what hes babbling about

16:34:46 <roGer^work> :-/

16:35:03 <roGer^work> libby: thanks 4 ure help anyway

16:35:12 <libby> well, unless rdql could somehow push it to the database level, chances are, optionals or ors would do the same io to the database....?

16:36:16 <roGer^work> hmm

16:36:21 <roGer^work> :-)

16:36:31 <roGer^work> i guess there is something i dont understand here

16:36:48 <libby> or perhaps something that I don;t...

16:37:19 <roGer^work> don't the rdql stuff map to sql (or something)?

16:38:22 <roGer^work> cause i dont want to half the db in to mem and process it there

16:38:44 <roGer^work> maybee i can design my datamodel better somehow

16:38:55 <roGer^work> i gotta go home now, but thanks alot

16:42:14 <libby> ah, he left

16:42:25 <libby> the or thing is really tricky

17:00:07 <mdupont> peace

17:00:46 <LotR> not on this planet.

17:20:33 <iwaiAway> iwaiAway is now known as iwaim

19:05:13 <eikeon_> eikeon_ is now known as eikeon

20:18:55 * D[a]vey is still hacking on RdfParser.php

20:52:20 <D[a]vey> ok, I got it grabbing the XML as CDATA... by enclosing it in <![CDATA[]]> and removing the r:parseType="Literal"

21:12:46 <D[a]vey> D[a]vey is now known as Davey

21:21:35 <Davey> Davey is now known as D[a]vey

21:52:38 <D[a]vey> D[a]vey is now known as Davey

21:59:46 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

22:21:01 <Davey> w00t!

22:21:21 <Davey> RAP 0.7 was just released and it supports the rdf:parseType="literal" stuff :

22:21:21 <Davey> :D

22:22:16 <_decoy_> you're parsing rdf in php? goddamn.

22:23:29 <Davey> _decoy_: yes, its lubbly :)

22:24:50 <_decoy_> frightening, morelike. 'have an app in mind?

22:35:11 <Davey> _decoy_: Annotea Server and Client

22:38:50 <_decoy_> oh. that's interesting. I tried annotea once, but at that time amaya *seriously* fought back. it'd be interesting to see e.g. a nicely engineered web interface into the underlying annotation model. (hint, hint... ;)

23:07:59 <deusx> deusx is now known as _deusx

23:11:22 <dmiles> dmiles is now known as dmiles_afk

23:47:27 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

23:49:32 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome