Semantic Web Interest Group IRC Chat Logs for 2005-09-01

04:44:47 <Biblio> hi

07:37:12 <SemanticEtna> hi all

07:37:25 <Biblio> hi, etna

07:38:03 <SemanticEtna> Biblio

07:38:12 <SemanticEtna> how're you?

07:38:51 <Biblio> very well, thanks, how about you?

07:40:21 <SemanticEtna> tomorrow i'll be married

07:40:27 <Biblio> !!

07:40:32 <Biblio> congratulations!

07:40:43 <SemanticEtna> thanks

07:42:05 <Biblio> any honey moon projects?

07:42:05 <SemanticEtna> biblio would you like to come to me pvt?

07:42:19 <SemanticEtna> honey moon in florence just a week

07:42:29 <Biblio> brilliant =PP

07:45:21 <SemanticEtna> i'm so tired

07:50:04 <danbri> congratulations!

07:50:18 <SemanticEtna> thanks dan

07:55:22 <SemanticEtna> news from developer's world?

07:55:56 <danbri> no news yet, it's not 9am here ;) am thinking about possible directions for rdf/geo work ... http://www.w3.org/2003/01/geo/

07:58:51 <Biblio> hi dan

08:00:08 <Biblio> was it you who posted that url yesterday abour the w3c discussion on compatibility between rdf and topic maps?

08:00:21 <Biblio> if so, thanks. i read it, i understand much better now.

08:00:39 <SemanticEtna> so explain me

08:00:41 <SemanticEtna> hahahah

08:00:42 <SemanticEtna> ;)

08:00:46 <Biblio> hehehe

08:01:10 <Biblio> well, in rdf assertions are unidirectional, it seems

08:01:20 <Biblio> and in topic maps, they are bidirectional

08:01:32 <danbri> url, yup

08:01:49 <Biblio> thanks, then :)

08:02:41 <Biblio> also in rdf assertions always eem to have the form subjetc-predicate, while in topic maps seems to be more generic

08:04:05 <Biblio> thosee seem to be the two most important modeling differences i noticed. i could be wrong, tough, but that document was pretty good.

08:06:08 <SemanticEtna> suggest me some lectures for honey moon

08:06:25 <SemanticEtna> i think i had firstly to read something about owl

08:06:39 <Biblio> your girl could sue me :P

08:06:48 <SemanticEtna> ha ha ha

08:07:31 <Biblio> http://www.w3.org/TR/rdftm-survey/ <--this is the one dan provided, and was very informative

08:07:31 <dc_swig> A: http://www.w3.org/TR/rdftm-survey/ from Biblio

08:08:24 <Biblio> hi, libby

08:08:27 <danbri> A:|That RDF/TM survey again, in case you missed it yesterday

08:08:27 <dc_swig> Titled item A.

08:08:58 <Biblio> oops. the bot catched the url. sorry, dan

08:12:14 <Biblio> btw, some of you might remember i posted a message to the w3c html list, XHTML2: Proposal for total separation of semantics from structure

08:12:37 <Biblio> it seems to have catched on, lots of messages by now

08:14:16 <SemanticEtna> relax ng based vaklidator

08:16:11 <SemanticEtna> I was discussing yesterday night about the definition of "Presentation Layer"

08:17:25 <Biblio> in which context? with who?

08:18:53 <SemanticEtna> in a place of a little town

08:19:08 <SemanticEtna> in a square, please me

08:19:23 <[GNU]> hello all

08:19:26 <SemanticEtna> hi gnu

08:19:31 <SemanticEtna> ^name GNU

08:19:31 <julie> Query returned no results

08:19:36 <SemanticEtna> ^name [GNU]

08:19:37 <julie> Christoph Görn, Christoph Goern

08:19:40 <[GNU]> right

08:19:46 <Biblio> hi

08:19:57 <[GNU]> heya

08:20:26 <[GNU]> i am just trying to figure out what is the best way to represent a path (or track) and "attach" photos to some waypoints

08:20:35 <[GNU]> trying to document a bike ride or so

08:21:42 <SemanticEtna> danbri has just posted a link about gis

08:22:31 <[GNU]> just is like yesterday?

08:22:49 <SemanticEtna> ehm

08:22:50 <danbri> earlier, wrote: am thinking about possible directions for rdf/geo work ... http://www.w3.org/2003/01/geo/

08:22:50 <SemanticEtna> no

08:25:21 <SemanticEtna> but i think its not enough

08:25:54 <Biblio> danbri is there a list or group coordinating between all the semantic groups and the html makers, so that all needed hooks for semantic purposes are guaranteed to exist in next version of html?

08:26:10 <[GNU]> so, transfered to my example... each waypoint i record on my bike tour is a locativePacket containing a foaf:depiction ?!

08:26:32 <[GNU]> oh... read on: locative:media :)

08:29:24 <SemanticEtna> so about presentation layer

08:29:45 <SemanticEtna> the definition should be "device indipendent"

08:30:11 <SemanticEtna> and i think it must consider also user inputs

08:34:48 <[GNU]> and... any advice what class a technorati tag may be in rdf world? dc:subject i guess?!

08:36:21 <KjetilK> it isn't quite clear to me what the unit of the number in the geo:alt element is...

08:36:42 <KjetilK> meters above the WGS_84 ellipsoid?

08:39:05 <danbri> its under-specified

08:39:24 <danbri> dc:subject is fine

08:39:45 <[GNU]> thx

08:40:31 <SemanticEtna> gnu i can't help you

08:41:10 <[GNU]> i happy SE :)

08:41:46 <SemanticEtna> Are you happy?

08:42:47 <[GNU]> ja

08:44:17 <SemanticEtna> so am I

09:08:55 <ldodds> hi dajobe

09:09:23 <ldodds> did you know that librdf.org/query is broken?

09:09:55 <dajobe> what's broke?

09:10:08 <dajobe> hmm I see my removing the default namespaces broke some examples

09:10:09 <ldodds> the service

09:10:19 <ldodds> 1:The namespace prefix in "rdf:type" was not declared.

09:10:25 <ldodds> 3:syntax error, unexpected $end, expecting '}'

09:10:25 <dajobe> ditto

09:10:26 <danbri> yeah i had impression rdf: was built-in

09:10:29 <dajobe> it's not

09:10:31 <danbri> was that just a convenience

09:10:38 <dajobe> it was a mistake on my part

09:10:42 <ldodds> I thought rdf was built-in too

09:10:42 <danbri> ah ok

09:10:49 <dajobe> and if you want it built in, last call comments due today :)

09:10:53 <danbri> tripped up peter saint-andre couple days ago

09:11:44 * dajobe goes demo fixing

09:11:54 <ldodds> ta

09:12:20 <ldodds> phil wilson was looking at jabber/sparql and was playing with the examples and noted they were broken, am channelling his bug report

09:13:27 <dajobe> all fixed

09:13:44 <danbri> phil wilson... interesting. anything public?

09:13:56 <dajobe> actually I could probably s/rdf:type/a/ and remove many of them

09:14:11 <ldodds> just tinkering I believe prompted by your discussions

09:14:22 <ldodds> Phil is up at bath uni these days, met him for a beer last week

10:15:37 <bballizlife_> bballizlife_ is now known as bballizlife

10:31:04 <[GNU]> i am just trying to get a grip on sparql and cant understand why result 2 has an img that is not empty, playing at: http://librdf.org/query?uri=http%3A%2F%2Fb4mad.net%2Fmaps%2Ftracks%2Ftour3.rdf&query=PREFIX+geo%3A+%3Chttp%3A%2F%2Fwww.w3.org%2F2003%2F01%2Fgeo%2Fwgs84_pos%23%3E%0D%0APREFIX+dc%3A+%3Chttp%3A%2F%2Fpurl.org%2Fdc%2Felements%2F1.1%2F%3E%0D%0APREFIX+l%3A+%3Chttp%3A%2F%2Flocative.net%2F2004%2Fpacket%2Floc%23%3E%0D%0APREFIX+m%3A+%3C

10:31:17 <[GNU]> wow, ups, bigger than frist thought, sorry

10:32:47 <[GNU]> maybe i dont understand the OPTIONAL keyword

10:33:15 <dajobe> there's some bugs in optional with rasqal

10:33:20 <[GNU]> OPTIONAL means that it is an option that the result is empty?!

10:33:31 <dajobe> possibly

10:34:30 <dajobe> hmm, that's a bug - '.'s should be required between triples

10:34:35 <dajobe> WHERE { ?x geo:lat ?lat ?x geo:long ?long ?y l:media ?img }

10:35:29 <dajobe> in that query ?x and ?y are unrelated

10:35:35 <dajobe> unJOINed tables if it was SQL

10:35:37 <[GNU]> using your WHERE ?img does always contain something

10:37:15 <dajobe> I've looked at the query log and can't quite see what's failing

10:37:51 <dajobe> SELECT ?lat, ?long, ?img WHERE { ?x geo:lat ?lat . ?x geo:long ?long . OPTIONAL { ?x l:media ?img } }

10:37:53 <[GNU]> ja, the last one with OPTIONAL seems to deliver what i was asking for

10:38:02 <dajobe> returns 2 results, with 1 optional

10:38:04 <dajobe> good

10:38:17 <[GNU]> this OPTIONAL is confusing

10:38:30 <[GNU]> its implicite in SQL

10:38:40 <dajobe> I think of it as two stages

10:38:50 <dajobe> match all the (mandatory) triples

10:38:59 <dajobe> then add any extra info that OPTIONALs match

10:39:12 <dajobe> which give additional bindings if they do match

10:39:54 <[GNU]> hmm, ja, kind of better pov

10:50:20 <dajobe>http://meta.wikimedia.org/wiki/Semantic_MediaWiki

10:50:20 <dc_swig> B: http://meta.wikimedia.org/wiki/Semantic_MediaWiki from dajobe

10:50:27 <dajobe> B:|Semantic MediaWiki

10:50:27 <dc_swig> Titled item B.

10:51:04 <dajobe> B:ideas and plans to allow encoding of typed relationships with rdf/owl semantics inside wikitext, in the MediaWiki software used by WikiPedia and projects

10:51:04 <dc_swig> Added comment B1.

10:51:23 <dajobe> B:->wiki,mediawiki,rdf,owl,swig

10:51:23 <dc_swig> Set keywords for B.

10:54:29 <[GNU]> if i have a rdf:Seq in the model and query for ?lat and ?long the result will not preserve the ordering of the Seq aint it?

10:54:43 <[GNU]> i guess this is by intention or wrong query or both

10:55:02 <dajobe> the ordering's in the properties that are off a seq - the rdf:li (in the rdf/xml) turn into rdf:_1 ... rdf:_n

10:55:20 <dajobe> but I think the modern way is to record any ordering info off the object itself

10:55:36 <[GNU]> ?

10:55:51 <[GNU]> give an increasing node id?!

10:56:09 <dajobe> <locative:Packet> <locative:index>1</locative:index> ...

10:56:18 <danbri> <dajobe> but I think the modern way is to record any ordering info off the object itself

10:56:19 <dajobe> or something. depends what the ordering means to you

10:56:40 <danbri> if the objects have URIs, that's not so good when you merge multiple docs mentioning same things w/ different orderings

10:56:44 <danbri> rss1 orig design tried that

10:57:01 <danbri> but sometimes apps just need to hang onto provenance and keep graphs distinct. much better tool support for that now.

10:57:29 <dajobe> ordering or merge, pick one

10:57:30 <[GNU]> ordered means a linear list of points, like a path on earth

10:57:48 <dajobe> ok

10:58:05 <danbri> or apps can order on worldly properties of the thing concerned

10:58:07 <dajobe> merge doesn't seem a use case in this example, the nodes are blank nodes

10:58:09 <danbri> prices, salary, date etc

10:58:18 <danbri> ...ie merging will cause a sane re-ordering

10:58:29 <dajobe> [GNU]: so you need to describe the path and the items inside it

10:58:40 <[GNU]> right

10:58:47 <dajobe> I guess an rdf collection would work here

10:58:52 <danbri> yuup

10:59:04 <dajobe> although I don't see why rdf:Seq wouldn't either

10:59:35 <dajobe> has the geo work got a set of points concept?

10:59:43 <dajobe> seq of points/path

11:00:10 <oud> semantic mediawiki, very interesting

11:00:16 <SemanticEtna> bye all

11:00:42 <[GNU]>http://locative.net/2004/packet/packet_line.rdf

11:00:44 <dc_swig> C: http://locative.net/2004/packet/packet_line.rdf from [GNU]

11:00:46 <oud> oud is now known as Biblio-

11:00:46 <[GNU]> i guess they have

11:00:53 <[GNU]> but kept secret 404

11:01:55 <[GNU]> i think putting the Seq into the model is ok, but retrieving the points in the order of the Seq may be a problem

11:02:54 <[GNU]> i cant determine if librdf sparql just returns and unordered result or the Seq in reverse order

11:03:11 <dajobe> it's unordered without ORDER BY

11:05:50 <dajobe> SELECT ?lat, ?long, ?img WHERE { ?seq a rdf:Seq . ?seq ?prop ?point ?point geo:lat ?lat . ?point geo:long ?long . OPTIONAL { ?x l:media ?img } } ORDER by ?prop

11:06:05 <dajobe> I'm not sure if that's illegal; I order by URI of property rdf:li turns into - rdf:_1 ... rdf:_4

11:06:20 <dajobe> oops, s/?x/?point/

11:10:41 <[GNU]> so the query picks up the Seq and assignes the _n to ?prop queries for the other stuff and orders by ?prop ?!

11:11:48 <dajobe> yeah, assigns ?prop to rdf:_1 ... rdf:_4 in this case and orders by the URIs of (rdf:_1 ...)

11:12:37 <[GNU]> funny world for an sql guy

11:13:06 <ldodds> yes, see you soon

11:13:14 * ldodds types in wrong window

11:13:26 <[GNU]> now i need a redland enabled apache/php...

11:13:30 <dajobe> ok, it works with arq so I declare it legal enough

11:13:39 <ldodds> ok

11:13:41 <[GNU]> dajobe: thanks for the query

11:13:50 <dajobe> once I fixed the lax syntax

11:21:30 <[GNU]> will there be php bindings in the debian packages in the future or is that handed over to the php maintainers?!

11:22:02 <dajobe> nobody asked for them yet

11:22:29 <dajobe> it's only 1 file I think

12:59:11 <dajobe>http://sourceforge.net/forum/forum.php?forum_id=492854

12:59:11 <dc_swig> D: http://sourceforge.net/forum/forum.php?forum_id=492854 from dajobe

12:59:23 <dajobe> D:|First Complete SPARQL (for sesame) Release

12:59:23 <dc_swig> Titled item D.

12:59:38 <dajobe> D:as the google summer of code ends today, releases of software are appearing, such as this one

12:59:38 <dc_swig> Added comment D1.

12:59:52 <dajobe> D:*it successfully passed all required test cases and should be now useful for general purpose functionality*

12:59:52 <dc_swig> Added comment D2.

13:00:08 <dajobe> D:not sure what required means...

13:00:08 <dc_swig> Added comment D3.

13:00:25 <dajobe> D:the [http://sourceforge.net/project/shownotes.php?group_id=142441&release_id=353366|release notes] give more info - it does SELECT and ASK but some areas aren't done

13:00:25 <dc_swig> Added comment D4.

13:00:34 <dajobe> D:->sparql,sesame,swig

13:00:35 <dc_swig> Set keywords for D.

13:03:28 <bjoern_> "A reference implementation of SPARQL"?

14:03:27 <DanC> tim, indeed, the parseType="Collection" syntax doesn't work for literals. you have to write it out long-hand with first/rest. This was noted by WebOnt, IIRC.

19:11:15 <Wikier> Wikier is now known as WikierOFF

20:08:52 <WikierOFF> WikierOFF is now known as Wikier

20:46:16 <yosi_s> eikeon, I don't think rdflib does the right thing on http://www.w3.org/2000/10/swap/test/xml-syntax/xmllit.rdf

21:03:07 <eikeon_> yosi_s, thank you. I'll take a look

21:04:27 <eikeon_> Might have been one of the fixes that happened between 2.0.x and 2.2.x... but I'll take a look.

21:05:05 <yosi_s> the svn looks identical between the two to me

21:05:18 * yosi_s will try again

21:05:41 <yosi_s> the version I have and the version online in the svn look the same for that block of code

21:05:42 <eikeon_> I can take a look this evening too... as soon as I get home.

21:05:45 <yosi_s> ok