Experimental IRC log swig-2009-05-25

Available formats: content-negotiated html turtle (see SIOC for the vocabulary)

Back to channel and daily index: content-negotiated html turtle

These logs are provided as an experiment in indexing discussions using IRCHub.py, Irc2RDF.hs, and SIOC.

04:59:31<mhausenblas>good morning Web of Data
06:02:36<mhausenblas>LeeF, FYI, I've tried to clarify some bits re the anti-patterns (http://webofdata.wordpress.com/2009/05/24/technology-malbestpracticing/#comment-82)
09:38:37<Anchakor>anyone knows if TriG supports expressing graphs with overlaps like this? http://fotopaste.cz/full/id=643
09:40:07<Anchakor>I mean without repeating the triples
10:31:23<tommorri1>yvesr: just looking at MO, specifically mo:download.
10:31:54<tommorri1>Got a sioc:Post that will always contain a link to an MP3. In RSS terms, an enclosure.
10:32:03<tommorri1>What the? tommorri1?
10:32:22<tommorris>Yes, an enclosure. Sometimes it's music, sometimes it's not.
10:32:42<tommorris>More often than not it's not music.
10:34:19<tommorris>anyone know what the best predicate to use to represent a sioc:Post pointing to an MP3 file as an 'enclosure'.
10:35:59<tommorris>the alternative I'm thinking of is using the RSS 1.0 enclosure module at http://www.xs4all.nl/~foz/mod_enclosure.html
10:36:49<tommorris>the URL being http://purl.oclc.org/net/rss_2.0/enc#enclosure
10:37:58<tommorris>sindice says both are as unused as each other. ;)
10:54:42<mhausenblas>argh. all attachments at ESW Wiki (http://esw.w3.org/) are gone
10:55:00<mhausenblas>shit. what am i supposed to do (update them manually?)
10:55:27<mhausenblas>I see the need for Mark to prevent spammer uploading attachements, but why are all the old attachements gone
10:55:29<mhausenblas>dman
19:08:33<idmclean>'sup
19:24:56<idmclean>tobyink, you awake?
22:02:16<LeeF>mhausenblas, thanks :)
22:27:18<cygri>mhausenblas?
22:30:00<mhausenblas>si
22:30:10<mhausenblas>... is on his way home ;)
22:30:43<mhausenblas>and as you're not around anymore we have to play tomorrow, right? :)
22:32:04<kwijibo>is this the ongoing table football battle?
22:32:35<cygri>mhausenblas is always in table football mode...
22:32:43<mhausenblas>kwijibo: yes :)
22:32:56<mhausenblas>and cygri lost against jumjum, LOL :D
22:33:29<kwijibo>lol
22:33:32<mhausenblas>ok, time to go (before I get beaten up by cygri ;)
22:33:57<cygri>hausi, got a question about rdfa and bookmarklets, but can wait till tomorrow
22:34:07<mhausenblas>quick one?
22:34:14<cygri>not sure
22:34:19<mhausenblas>shoot
22:34:37<cygri>basically i want to make a bookmarklet that evaluates rdfa markup in the current page
22:34:47<mhausenblas>ACTION notes that he should ignore sentences that start with hausi ...
22:34:55<cygri>blasi?
22:34:57<mhausenblas>yes. we have that
22:35:09<mhausenblas>ACTION is not gonna say the word, no ...
22:35:20<cygri>ideally, the code should know which dom elements correspond to the RDF nodes
22:35:41<mhausenblas>hm. so, not like in lidaman,right?
22:36:19<cygri>i think lidaman just takes the statements out of the page, but does not keep a connection back to the place in the page where it was found?
22:36:43<mhausenblas>well, sort of it does
22:36:48<mhausenblas>if you look at the code in http://bitbucket.org/mhausenblas/lidaman/src/
22:36:48<cygri>i want to keep that connection... a bit like the rdfa highlighter bookmarklet, perhaps
22:37:08<mhausenblas>it decorates all the spots with a button, so you've got a hook there
22:37:43<CaptSolo>hi
22:37:47<mhausenblas>the spots == the places where it find an @about
22:38:21<mhausenblas>if you install the bookmarklet ffrom http://ld2sd.deri.org/lidaman/ and visit http://ld2sd.deri.org/lidaman/sandbox.html you'll see what I mean ...
22:38:59<CaptSolo>cygri: how would you keep that connection?
22:39:31<cygri>mhausenblas, ok will study the lidaman code
22:39:56<CaptSolo>as in: how would you represent that connection in RDF?
22:39:56<mhausenblas>best is to start with http://bitbucket.org/mhausenblas/lidaman/src/tip/lidaman.js I guess ...
22:39:58<cygri>CaptSolo, by showing some small widget thingy in the page where the RDFa occurs
22:40:38<mhausenblas>(one needs basic jQuery knowledge to read the code, but this is quite simple, I hope)
22:40:45<cygri>ACTION didn't realize that bookmarklets can load additional code from off-site
22:41:17<mhausenblas>that's the main trick
22:41:33<mhausenblas>same as we do in RDForms to load the actual form ...
22:41:37<cygri>mhausenblas, sigma uses jquery quite heavily...
22:42:24<cygri>anyway, this should be enough to get me started... thanks mhausenblas, i'll certainly have more q's tomorrow
22:42:51<mhausenblas>sure. yw.
22:42:57<kwijibo>cygri: mhausenblas what day is our voiD telecon again ?
22:43:13<mhausenblas>hm. I guess *someone* hasn't answered yet
22:43:15<mhausenblas>:)
22:43:40<kwijibo>i will be away from home, so i might need to just listen and type
22:44:07<CaptSolo>cygri: ok, a widget thingie. and you'll have the tool extracting RDFa supply you with the positional information (where this RDFa statement was found)
22:44:07<mhausenblas>looks like 28. and someone should ack that ASAP.
22:44:11<mhausenblas>someone?
22:44:34<mhausenblas>cygrione?
22:44:36<CaptSolo>i was confused how one would keep the information linking a location on the page w. RDFa dat found there
22:44:37<kwijibo>mhausenblas: who is the someone, is it me ?
22:44:52<mhausenblas>no, kwijibo. its the other one :)
22:45:06<kwijibo>good, i thought i did :p
22:45:13<CaptSolo>but if you are not trying to transfer that information anywhere outside the page, that is fine
22:45:23<CaptSolo>otherwise you will need a way to:
22:45:31<mhausenblas>no, the guy who studies lidaman code, currently
22:45:34<CaptSolo>(a) include that info w. extracted RDF data
22:45:43<Anchakor>noob question: what symbols are allowed in uri after #?
22:45:54<CaptSolo>(b) use some other method for representing this link
22:46:13<mhausenblas>Anchakor: see http://tools.ietf.org/html/rfc3986#section-3.5
22:46:19<kwijibo>CaptSolo: I imagine he isn't exporting the data, but you could do it as named graphs
22:46:34<kwijibo>using bengee's resource pointer style uris maybe
22:46:49<kwijibo>would that work? i'm tired ... :p
22:47:14<mhausenblas>ok, now really really off to bed. need some sleep to deliver good fresh dogfood tomorrow morning
22:47:21<mhausenblas>ESWC09 data upcoming
22:47:21<Anchakor>mhausenblas: heh, I was looking at it a moment ago, not sure if this is really it, will dig in it now :)
22:47:31<mhausenblas>Anchakor: it is
22:47:35<cygri>CaptSolo, if the RDFa markup uses about="#fragment", then the URI itself nicely doubles as a marker of the location in the page
22:47:44<CaptSolo>kwijibo: ok. resource pointer style uris? would they allow one to point to an arbitrary place in HTML code?
22:47:54<CaptSolo>cygri: indeed
22:48:15<CaptSolo>cygri: i guess it is not a problem then
22:48:18<mhausenblas>cygri: fill in the damn doodle :)
22:48:48<mhausenblas>good nite Web of Data
22:49:13<cygri>CaptSolo, the thing i have in mind just needs to keep track of this within the page, so it should be okay even for other uris or bnodes
22:49:25<kwijibo>CaptSolo: yeah - though you would have to choose one graph name for the triple
22:49:32<cygri>mhausenblas good night
22:49:55<kwijibo>when you might went to say wher each part of the triple came from
22:49:59<kwijibo>so, some kind of reification
22:50:16<kwijibo>maybe ultramega triple times reification
22:52:09<kwijibo>so you could reify the triples, then you could reify the reified statements, and say that the reified reified statement part had a source of an xpath to a dom node
22:52:45<kwijibo>which sounds like another RDF anti-pattern for mhausenblas's collection ;)
22:53:36<CaptSolo>kwijibo: i wonder what are you drinking :)
22:53:47<CaptSolo>what makes you come up with "ultramega triple times reification"
22:53:57<CaptSolo>sounds good, though ;)
22:54:23<CaptSolo>kwijibo: but gotta go now. take care.
22:54:38<cygri>ACTION waves to CaptSolo
22:54:53<kwijibo>bye CaptSolo
22:55:07<kwijibo>it's only double reification really
22:55:16<kwijibo>i wonder if anyone's ever really done that
22:55:29<cygri>hopefully not kwijibo
22:55:33<kwijibo>:D

Back to channel and daily index: content-negotiated html turtle