Semantic Web Interest Group IRC Chat Logs for 2002-10-10

00:05:19 <jordan> jordan is now known as jordanpractice

08:13:55 <kensaku> kensaku is now known as tansaku

09:13:45 <maxoff> maxoff is now known as maxf

09:13:46 <Snakker> Snakker is now known as libby

09:13:49 <dajobe> hello zetlanders

09:14:00 <maxf> hi dajobe

09:14:04 <libby> hia

09:14:45 * danbri admires the chump archive from yesterday

09:15:20 <dajobe> pull out links into a doc (with sentences) of resources, work, tools ?

09:15:28 <dajobe> or going into workshop report?

09:15:59 <danbri> bit of both. theres a lot of fodder for a w/shop report

09:16:07 <dajobe> looks like it

09:16:12 <dajobe> mtg looked to go well

09:16:20 <danbri> yeah, i think it did

09:16:29 * maxf thinks so too

09:16:34 * dajobe gives everyone involved a gold star

09:16:36 <danbri> libby++ #cally

09:16:45 <danbri> everyone-else++ too

09:16:56 <danbri> hmm, I want a commenting karmabot :)

09:17:04 <dajobe> heh

09:17:10 <danbri> now, wtf has happened to *.w3.org / MIT...

09:17:15 * danbri needs email fix

09:17:30 <dajobe> give generously, danbri needs mail

09:17:57 <danbri> send what little you can afford. Give until it hurts and give some more.

09:18:06 <danbri> (unless its about datatypes...)

09:18:34 <danbri> _actually_... Dave, do you have a pointer to what the RDF syntax for datatyping (local) will look like?

09:18:46 <dajobe> yeah

09:18:50 <danbri> Or is that on hold too for current strawpoll, I <blush/> forget...

09:18:54 <dajobe> no

09:18:58 <danbri> cool

09:19:14 <dajobe> it's waiting to decide if they have languages

09:19:24 <dajobe> see http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Sep/0198.html

09:19:39 <danbri> hmm that url worked

09:19:58 <dajobe> jjc's reply I'd consider more of an indication of which options are goers http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Sep/0211.html

09:20:29 <dajobe> so hte answer is

09:20:37 <dajobe> datatypeString ::= uriref '"' string '"' ( '-' language )?

09:20:51 <danbri> ok, that's the n-triples...

09:21:00 * dajobe re-reads

09:21:01 <dajobe> oops

09:21:05 <danbri> do you have the xml synyax too?

09:21:09 <dajobe> yes

09:21:10 <danbri> both needed, ta

09:21:27 <dajobe> <foo:prop rdf:datatype="uri-of-datatype" xml:lang="en">blah</foo:prop>

09:21:31 <dajobe> xml:lang opt of course

09:21:45 <dajobe> well, rdf:datatype is too

09:22:08 <dajobe> then it's a string literal

09:41:36 <danbri> ta

09:41:43 <danbri> can use entities for the uris...?

09:41:56 <danbri> and we've not decided which uris to use, but same as xsd would be plausible?

09:44:41 <dajobe> entities - yes

09:44:55 <dajobe> uris - xsd ones are de-rigeur as examples

09:45:23 <dajobe> I converted a dc-in-xml example to rdf+datatypes

09:45:26 <dajobe> came out ok

10:18:21 <shellac__> shellac__ is now known as shellac

12:08:54 <dmilezZ> dmilezZ is now known as dmiles

12:38:30 <danb_lap> BLURB:SOAP-RDF stuff

12:38:31 <dc_rdfig> A: SOAP-RDF stuff from danb_lap

12:39:16 <danb_lap> A:See [soap2rdf.xsl and other files|http://www.w3.org/2001/sw/interest/soaptech/xslt/]

12:39:17 <dc_rdfig> Added comment A1.

12:41:16 <danb_lap> A:swad-europe [workplan for sw/ws work|http://www.w3.org/2001/sw/Europe/plan/workpackages/live/esw-wp-4.html], [work in progress draft on soap-serialization|http://www.w3.org/2001/sw/Europe/reports/xml_graph_serialization_report/]

12:41:16 <dc_rdfig> Added comment A2.

12:42:31 <danb_lap> A:nearby in the web, some [test data in rdf and soap|http://fireball.danbri.org/people/danbri/2002/05/raa/] (currently on my own dev box as file size barfed w3.org cvs server)

12:42:31 <dc_rdfig> Added comment A3.

12:43:32 <danb_lap> A:We're looking at improved datatype support for the soap 2 rdf transformation. See [earlier discussion with dave|http://ilrt.org/discovery/chatlogs/rdfig/2002-10-10.html#T09-18-34]

12:43:32 <dc_rdfig> Added comment A4.

12:57:27 <sandro> Is anyone cognizant of RDF Core WG process handy?

12:58:24 <em_> yes

12:59:07 <sandro> ah, but I'm already scheduled to talk to *you* :-)

12:59:16 <em_> :)

12:59:29 <danb_lap> A:Stuff to do: make soap2rdf pass datatypes through into the new RDF/XML syntax for datatypes

12:59:30 <dc_rdfig> Added comment A5.

12:59:37 <em_> ok, i can wait... wanted to add this very topic to our discussion :)

13:00:41 <danb_lap> A:...and then if Max updates his [http://www.w3.org/2001/12/rubyrdf/xsltrdf/|xslt rdf parser] to consume rdf:datatype="..." syntax, we should be able to get N-Triple RDF from SOAP messages.

13:00:42 <dc_rdfig> Added comment A6.

13:01:18 <danb_lap> A:Here are a load of [soap test files|http://fireball.danbri.org/people/danbri/2002/05/raa/raa-dump/] we're using (stripped of protocol header stuff)

13:01:19 <dc_rdfig> Added comment A7.

13:01:42 <sandro> I'm thinking David Martin's complain about rdf:List should have gone to rdf-comments, but since it didn't it hasn't opened the issue it should have. Am I right? I need to report to DAML that it's an open issue, IMO. http://lists.w3.org/Archives/Public/www-rdf-logic/2002Oct/0001.html

13:01:55 <danb_lap> A:These have elements like: <update xsi:type="xsd:dateTime">2002-01-21T18:23:52Z</update>

13:01:56 <dc_rdfig> Added comment A8.

13:02:33 <danb_lap> A:soap2rdf currently drops the datatype info on the floor.

13:02:34 <dc_rdfig> Added comment A9.

13:04:44 <dajobe> sandro: I think he's still confused, it seems the construct does define a closed ordered sequence, which is what he wanted

13:05:04 <danb_lap> I forwarded those comments to rdf-comments

13:06:54 <sandro> This is a process question, dajobe. He believes this is a "a significant issue", but it has not been addressed to his satisfaction. In fact, he has received no response at all, as far as I can tell.

13:07:04 <danb_lap> A:[new home for soap2rdf.xsl|http://www.w3.org/2001/sw/Europe/reports/xml_graph_serialization_report/xslt/] in swad-e reports directory.

13:07:05 <dc_rdfig> Added comment A10.

13:07:18 <dajobe> sandro: we were discussing it further in private email

13:07:43 <em_> sandro, from a process standpoint, if he's looking for a response from the working group, he needs to send this request to the appropriate list - rdf-comments

13:08:49 <dajobe> the daml+oil description he quotes has apparently changed (was news to me); so where it used to say "unordered", in rdf:List terms it now ahs an order

13:11:39 <sandro> So you're taking this as an unofficial issue (as editor) to change the wording in the Spec to make it clear this is an ordered sequence?

13:12:12 <dajobe> the seamntics of the form is for the model theory

13:12:16 <dajobe> and that'll be there

13:12:32 <dajobe> but I'll also put more strongly that it gives you a closed and ordered sequence of nodes

13:12:35 <dajobe> in the syntax

13:13:26 <dajobe> which should answer his "RDF and DAML+OIL do not provide any convenient notation for (ordered) lists"

13:13:42 <dajobe> i.e. they do now (for some measure of convienent)

13:14:21 <danb_lap> makes sense to me; it's ordered, we just don't say much about what the ordering means.

13:14:29 <dajobe> yup

13:14:30 <danb_lap> which is fine (compare w/ the mess of rdf:Alt ...)

13:14:36 <dajobe> but I obviously said too little

13:14:55 <dajobe> can we have an rdfs:Schema class dan?

13:15:03 <dajobe> cf owl's owl:Ontology

13:16:09 <mhgrove-clone> mhgrove-clone is now known as mhgrove

13:20:52 <danb_lap> if you want to define the difference, and persuade everyone of it, sure!

13:21:01 <danb_lap> are there things that are members of the one class, but not the other, etc

13:21:11 <danb_lap> If not, why not just just owl:Ontology ?

13:21:33 <danb_lap> (apart from not liking the word... I'd prefer Vocab)

13:21:40 <dajobe> well use vocab

13:23:23 <danb_lap> why not owl:Ontology?

13:23:43 <danb_lap> doesn't seem much point in having two ways to say the same thing...

13:25:19 <dajobe> that's a new one

13:25:32 <dajobe> previously you just refused to consider adding it at all

13:26:26 <dajobe> it's for sticking description of the rdf schema 'document' and discovering it

13:26:29 <dajobe> EOR has used it

13:26:35 <dajobe> and so is daml+oil, owl

13:26:37 <danb_lap> I considered it!

13:26:38 <dajobe> probably others

13:26:47 <dajobe> for pointing isDefinedBy at

13:27:16 <danb_lap> I find RDF quite deployable without it, so as an added extra, OWL seems a fine home

13:27:27 <danb_lap> less is moreism

13:27:38 <dajobe> I'd prefer it in rdfs

13:27:40 * danb_lap can't get to w3c site...

13:27:42 <dajobe> since people have had to add it

13:27:49 <danb_lap> 'had to'?

13:28:08 <danb_lap> that's a tricky criteria. People have decided to use lots of things not in rdfs.

13:28:39 <dajobe> I've a pile of rdf, I want to find the desc of all the terms, do I a) find all the Class, Properties then reverse the isDefinedBy arcs or b) loook for rdfs:Vocab classes. b) was what EOR, daml+oil, OWL, others chose

13:28:50 * JibberJim hasn't been virtually able to get to w3.org all day, netscalibur in the UK seem to have serious routing probs.

13:28:53 <danb_lap> I'd vote for hypertext versions of rdfs:label/comment, for InverseFunctionalProperty, StaticInverseFunctionalProperty

13:29:07 <danb_lap> ...for versioning info, etc too

13:29:13 <danb_lap> but don't _need_ them in rdfs.

13:29:49 <dajobe> you don't _need_ rdfs

13:29:51 <dajobe> or rdf

13:29:52 <dajobe> or xml

13:30:06 <danb_lap> yup. hence puzzlement over your 'had to'

13:30:19 <dajobe> since you don't see the need

13:30:27 <dajobe> people had to invent their own classes

13:30:28 <danb_lap> I can see a use

13:30:30 <dajobe> to perform the same use

13:30:31 <dajobe> each time

13:30:33 <dajobe> with different ns

13:30:42 <danb_lap> a nice w3c note would do fine

13:30:50 <danb_lap> 'handy addons' namespace

13:30:51 <dajobe> and who's going to wait for the final owl class namespace to be determined

13:30:58 * dajobe groans

13:31:03 <dajobe> it's for rdf schemas/vocabs

13:31:06 <dajobe> put it in rdfs

13:31:09 <danb_lap> yes sir

13:31:23 <danb_lap> Owl ontologies are rdf schemas, vocabs.

13:32:07 <dajobe> and any language ontop of rdfs is such a thing too

13:32:12 <dajobe> and if each one has a simialr feature

13:32:17 <dajobe> s/each/many/

13:32:24 <dajobe> deserves being in the lower level

13:32:28 <dajobe> given it has use

13:32:44 <dajobe> would make the discovery of rdf schemas/vocabs, owl ontologies and derived classes easier

13:32:52 <dajobe> "give me all the subclasses of rdfs:Vocab"

13:33:25 <danb_lap> I can see a case. But we're also late. There are other things I'd rate higher for getting rdf deployed. FunctionalProperty specifically.

13:33:31 <dajobe> same reason why rdfs:Property, rdfs:Class allows discovery of terms

13:33:41 <dajobe> that's true but complex

13:33:43 <danb_lap> Unless we really believe that everything has a uri

13:33:52 <danb_lap> and that people will know those uris!

13:33:53 <dajobe> this class ahs a lower entry point

13:34:11 <danb_lap> are there things that are members of owl:Ontology but not rdfs:Vocab ?

13:34:15 <danb_lap> or vice versa?

13:35:20 <dajobe> you are the schema editor and attend webont, don't you know?

13:36:45 <danb_lap> how should I know, dave? There's no formal definition of owl:Ontology that gives necessary and sufficient conditions for class membership; and we've not defined an rdfs:Vocab at all yet.

13:36:59 <danb_lap> My expectation would be that theres no difference, and it is a word choice and which-namespace issue.

13:37:12 <danb_lap> rdfs:Ontology being the obvious compromise that nobody would like!

13:38:05 <danb_lap> A:nearby: [danbri soap 1.2 adjunct -- last call review comments|http://lists.w3.org/Archives/Public/xmlp-comments/2002Jul/0060.html]

13:38:06 <dc_rdfig> Added comment A11.

13:38:14 <dajobe> I wasn't at webont f2f I thought they discussed owl:Ontology and what that mean re' inclusion

13:38:19 <dajobe> and might mean something diff

13:38:39 <dajobe> if it has strong semantics, something weaker such as rdfs:Vocab would be useful too

13:40:04 <danb_lap> I wasn't at the f2f in the end, just the bar on tuesday...

13:41:55 <JibberJim> good to see your priorities were right there...

13:43:02 * danb_lap not a webont member, just a lurker mostly

13:44:44 <JibberJim> even better then, if you had no reason at all to be there other than the bar...

13:45:19 * danb_lap missed JimH (he left earlyish), which was a shame

13:45:54 <shellac> yeah - he left immediately after finish

13:59:49 <sandro> is the irc server having problems, or is it just me?

14:00:37 <danb_lap> mit net maybe...? are you at mit?

14:00:57 <sandro> my normal irc connection goes via a proxy at MIT, yeah.

14:08:26 <sandro> Ah, I see. (I just got the e-mail about MIT network problems.)

14:10:48 <maxf> and?

14:11:33 <sandro> and that explains why my irc connection from MIT isn't working well, I bet. (and why the web servers aren't either.)

14:11:40 <danb_lap> A:See [eg 4 from soap 1.2 primer|http://www.w3.org/TR/2002/WD-soap12-part0-20020626/#L3374] for a soap-encoded example (but no datatypes)

14:11:40 <dc_rdfig> Added comment A12.

14:48:19 <qmacro> qmacro is now known as qmacro_away

15:07:21 <Snakker> Snakker is now known as libby

15:38:17 <danb_lap> A:[xml-dist-app discussion of SOAP encoding test methodology|http://lists.w3.org/Archives/Public/xml-dist-app/2001Dec/0132.html]

15:38:18 <dc_rdfig> Added comment A13.

15:57:07 <iBitsko>http://bitsko.slc.ut.us/~ken/rss-sxs/RDF%3A%3ASimple%3A%3AXMLParser.pm

15:57:07 <dc_rdfig> B: http://bitsko.slc.ut.us/~ken/rss-sxs/RDF%3A%3ASimple%3A%3AXMLParser.pm from iBitsko

15:57:24 <iBitsko> B:|RDF::Simple Perl module

15:57:24 <dc_rdfig> Titled item B.

15:57:58 <Morbus> bookmakre.d

15:58:22 <Morbus> i'll see if i can understand your code.

15:58:31 <Morbus> maybe we can tagteam support of it.

15:58:36 <iBitsko> B:about 80% complete, I think. reads abbreviated syntax, typed nodes, and normalizes rdf:Seq into Perl arrays

15:58:37 <dc_rdfig> Added comment B1.

15:59:31 <danb_lap> B:Cool :)

15:59:32 <dc_rdfig> Added comment B2.

16:01:01 <iBitsko> B:converts RDF statements into Perl hashes. each key is the predicate name in James Clark's XML Namespace notation (similar to other Perl modules, rather than RDF norm), with the rdf:about key being the subject.

16:01:02 <dc_rdfig> Added comment B3.

16:01:45 <iBitsko> B:resolves rdf:resource references, replacing the reference with the corresponding Perl hash if it's been loaded

16:01:46 <dc_rdfig> Added comment B4.

16:03:54 <iBitsko> B:This module uses XML::Parser directly so should work standalone with only XML::Parser as a prerequisite. the RDF::Simple module should likely be factored out (and is already coded this way) so it works with both XML::Parser or SAX

16:03:55 <dc_rdfig> Added comment B5.

16:05:02 <iBitsko> shoot, I was going to change which directory that was stored in :(

16:06:07 <iBitsko> B:http://bitsko.slc.ut.us/~ken/rss-sxs/rsstest.pl

16:06:08 <dc_rdfig> Added comment B6.

16:06:18 <iBitsko> B:Example usage, using RSS

16:06:18 <dc_rdfig> Added comment B7.

16:06:35 <JibberJim> what do you mean by it resolves "rdf:resource"'s ?

16:07:16 <iBitsko> any object that has an 'rdf:resource' xml attribute is replaced physically with the Perl node that it refers to

16:07:46 <JibberJim> only internal ones, it doesn't go off outside the RDF doc though right?

16:07:55 <iBitsko> correct.

16:08:16 * JibberJim imagined it trying to download every rdf:resource mentioned in a doc...

16:08:17 <iBitsko> I had an "expand()" method in mind to cause it to go off an gather more RDF

16:08:33 <iBitsko> not automatically ;)

16:12:03 <iBitsko> how do I change a comment? I want to change B6. is it just "B6:new comment"?

16:12:53 <Morbus> yeah

16:13:05 <iBitsko> thx, I as just chatting with dc_rdfig too ;)

16:13:41 <iBitsko> B6:[rsstest.pl|http://bitsko.slc.ut.us/~ken/rss-sxs/rsstest.pl]

16:13:42 <dc_rdfig> Replaced comment B6.

16:19:14 <iBitsko> B:[rdftest.pl|http://bitsko.slc.ut.us/~ken/rss-sxs/rdftest.pl] runs the [RDF Test Cases|http://www.w3.org/TR/rdf-testcases/], but doesn't generate the corresponding output format to compare against the expected results in that format.

16:19:14 <dc_rdfig> Added comment B8.

16:21:33 <iBitsko> B:RDF::Simple is best used if you just want to slurp in RDF and use it as node-based data, similar in scope to XML::Simple, Storable, or a variety of property-list or configuration-file modules.

16:21:34 <dc_rdfig> Added comment B9.

16:24:21 <qmacro_away> qmacro_away is now known as qmacro

16:24:28 <dajobe> iBitsko: good job, ken

16:24:41 <iBitsko> oh geez. ok. just fifteen more minutes and I can wrap that up into a real module and upload to CPAN ;)

16:24:57 <iBitsko> dajobe: thx!

16:25:21 * iBitsko blows the dust off of h2xs

16:25:23 <dajobe> announce to www-rdf-interest too, when that's uploaded?

16:25:25 <dajobe> heh

16:30:36 <danb_lap>http://www.y12.doe.gov/sgml/sc34/document/0221.htm

16:30:37 <dc_rdfig> C: http://www.y12.doe.gov/sgml/sc34/document/0221.htm from danb_lap

16:30:56 <danb_lap> C:|ISO proposed work item on Topic Map Constraint Language

16:30:56 <dc_rdfig> Titled item C.

16:31:28 <danb_lap> C:Another XML-based schema language. This one for defining topic map structures / vocab...

16:31:28 <dc_rdfig> Added comment C1.

16:31:43 <larsbot> C:TM equivalent to RDFS/DAML+OIL/OWL, really

16:31:43 <dc_rdfig> Added comment C2.

16:31:59 <larsbot> C:currently stalled while the data model is stabilized

16:31:59 <dc_rdfig> Added comment C3.

16:32:27 <larsbot> C:the full story is in [http://www.y12.doe.gov/sgml/sc34/document/0323.htm|the standardization guide]

16:32:27 <dc_rdfig> Added comment C4.

16:32:28 <danb_lap> C:Yup. Interesting re binary-vs-n-ary issues w/ associations...

16:32:29 <dc_rdfig> Added comment C5.

16:32:35 <danb_lap> cool thanks for the link

16:44:30 <larsbot> C:a better link, btw, is [http://www.y12.doe.gov/sgml/sc34/document/0226.htm|the requirements document]

16:44:30 <dc_rdfig> Added comment C6.

16:45:14 <iBitsko> that link is backwards, I think it's [description|URL]

16:45:25 <dajobe> doesn't matter, the bot fixes it

16:45:28 <iBitsko> unless the code's smarter than I remember ;)

16:45:32 <iBitsko> cool!

16:45:41 <larsbot> dajobe: I was uncertain. glad the bot is smarter than me :)

16:45:42 <iBitsko> probably for just that reason ;)

16:45:45 <dajobe> turned out nobody coudl remember the order

16:45:52 <dajobe> usability wins over syntax

16:45:55 * larsbot feels slightly smarter now

16:46:13 <larsbot> having the URI first is consistent with wiki syntax, though

16:46:25 * dajobe tries to imagine a web where usability really did win over syntax

16:46:35 <larsbot> it's called HTML... :)

16:47:13 <dajobe> tag soup

16:47:31 <larsbot> sure, but easy on the users

16:49:45 <dajobe> hi AaronSw, how did your SC visit go?

16:50:17 <AaronSw> great! i'm writing about it now

16:50:22 <dajobe> cool

16:50:42 <dajobe> the reports from /. didn't seem too good for lessig, but tricky to tell I guess?

16:51:27 <dajobe> today AaronSw's blog is mostly tag soup (aka invalid html) ;)

16:52:41 <AaronSw> yeah. i got the totally opposite sense

18:03:07 <AaronSw>http://www.aaronsw.com/weblog/000650

18:03:07 <dc_rdfig> D: http://www.aaronsw.com/weblog/000650 from AaronSw

18:03:15 <AaronSw> D:|Mr. Swartz Goes to Washington

18:03:15 <dc_rdfig> Titled item D.

18:04:19 <MarkB>http://guests.evectors.it/zoe/

18:04:20 <dc_rdfig> E: http://guests.evectors.it/zoe/ from MarkB

18:04:35 <MarkB> E:|Zoe, an email indexing proxy

18:04:37 <dc_rdfig> Titled item E.

18:07:40 <MarkB> E:would love to see it integrated with [Dan's RDF/email work|http://lists.w3.org/Archives/Public/www-rdf-interest/2000Jul/0020.html]

18:07:41 <dc_rdfig> Added comment E1.

18:08:35 <qmacro> qmacro is now known as q|away

18:08:50 <MarkB> E:and then add support for [every other application protocol|http://www.xent.com/FoRK-archive/dec00/0014.html]

18:08:51 <dc_rdfig> Added comment E2.

18:13:57 <dajobe> D:great report Aaron, looks like you had a blast

18:13:59 <dc_rdfig> Added comment D1.

18:15:03 <AaronSw> thanks. yep :)

18:40:29 <iBitsko> looking at the RDF test cases, rdf-charmod-literals in this case, why does <eg:Creator eg:named="Dürst"/> create two triples?

18:40:39 <iBitsko> _:a <http://example.org/named> "D\u00FCrst" .

18:40:48 <iBitsko> <http://www.w3.org/TR/2002/WD-charmod-20020220> <http://example.org/Creator> _:a .

18:41:58 <iBitsko> second question: when _:a was generated, was that name/URI per the RDF spec or the RDF Test Cases doc?

18:42:01 <dajobe> it's a prop attr

18:42:06 <dajobe> property attribute

18:42:09 <dajobe> on a property element

18:42:15 <dajobe> so it creates an intermediate node

18:42:38 <iBitsko> ah! ok, RDF::Simple doesn't support that yet then ;)

18:43:20 <Morbus> the support calls roll in! ;)

18:43:30 <iBitsko> re the generated name, I would guess that would have to be per the RDF spec if all tools were expected to generate the same test results

18:44:31 <dajobe> i'm not sure I get Q2

18:44:51 <dajobe> are you talking about _a: ?

18:45:31 <dajobe> it's an identifier for a node in the graph with no URI

18:45:47 <dajobe> a blank node it's usually called, an empty circle in the visual pictures

18:46:05 <dajobe> so in a serialisation it gains a temporary identiifer in order to refer to it in multiple places

18:46:07 <iBitsko> yes, but where is it specified how those are generated?

18:46:09 <dajobe> _:a is such a name

18:46:16 <dajobe> it could be _:anythingyoulike

18:46:27 <dajobe> so you have to match 'the blanks'

18:46:30 <iBitsko> such that a tool can properly use the test cases

18:46:34 <dajobe> this is the graph comparison problem

18:46:57 <dajobe> there's a free tool in the test suite are you can use

18:47:04 <dajobe> in utils/ntcmp or something

18:47:10 <dajobe> C++, but there you go

18:47:48 <dajobe> it's on the web, not in the zip

18:53:50 <dajobe>http://www.w3.org/2000/10/rdf-tests/rdfcore/utils/ntc/

18:53:51 <dc_rdfig> F: http://www.w3.org/2000/10/rdf-tests/rdfcore/utils/ntc/ from dajobe

18:54:12 <dajobe> F:|ntc - free tool for comparing RDF N-triples files for graph equality

18:54:13 <dc_rdfig> Titled item F.

18:56:21 <iBitsko> heh. it occurs to me that if I am able to get the perl hashes to serialize back to N-triples I'd have made the whole round-trip possible (writing and reading its own files)

18:56:30 <dajobe> F:see also the fine paper by Jeremy Carroll: [http://www.hpl.hp.com/techreports/2001/HPL-2001-293.html|Matching RDF graphs] if you want to implement something efficient, like he did for Jena's model comparison

18:56:31 <dc_rdfig> Added comment F1.

18:56:43 <dajobe> that's the general idea

18:56:52 <iBitsko> currently, I think the import as-is is lossy

18:57:10 <dajobe> this might help: http://www.redland.opensource.ac.uk/ntriples/

18:57:24 <dajobe> an ntriples validator; or at least my ntriple parsers' claims

20:08:53 <pxh|away> pxh|away is now known as pxh

20:17:29 <iBitsko> in parsing RDF/XML, is the element depth (even/odd) always correspond exactly to description -> property -> description -> property -> ...?

20:17:42 <iBitsko> or am I misreading?

20:18:06 <dajobe> usually

20:18:13 <dajobe> if you have parseType=Resource, you skip one

20:18:36 <iBitsko> any other cases (besides parseType=Literal, too)

20:18:47 <dajobe> this is explained in the revised syntax doc, but you can google for 'rdf striping' and find danbri's doc too

20:18:55 <dajobe> no

20:19:07 <dajobe> but there are parseType=Collection

20:19:12 <dajobe> which just works different

20:19:14 <iBitsko> thx. I'm reading the current doc, but that bit didn't come out clearly

20:20:50 <dajobe> that's useful feedback

20:21:00 <dajobe> it's the bit in http://www.w3.org/TR/rdf-syntax-grammar/ sec 2

20:21:09 <dajobe> nr Example 1 and 'striping'

20:21:28 <dajobe> but I hadn't got to explaining parseTYpe=resource then

20:22:39 <iBitsko> way cool. I just got rid of the state machine in my parser.

20:23:11 <dajobe> it'll be back, I predict ;)

20:27:31 <dajobe> iBitsko: I usually point people at my rdf/xml test cases too, they were for testing individual bugs, features or tricky bits I found

20:27:43 <dajobe> (might => bigger test suite later)

20:27:56 <iBitsko> thx!

20:28:06 <dajobe> see http://cvs.ilrt.org/cvsweb/redland/raptor/tests/

20:32:55 <pxh> pxh is now known as pxh|away

21:54:57 <bijan> bijan has changed the topic to: RDF/Semantic Web 24x7 chap (aka mhgrove) - http://rdfig.xmlhack.com/ blog

22:01:57 <dajobe> dajobe has changed the topic to: RDF/Semantic Web 24x7 chap (aka mhgrove) - http://rdfig.xmlhack.com/ bloke (aka bijan)

22:02:34 * larsbot wonders what prompted this