Experimental IRC log swig-2008-09-07

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.

09:58:04<mhausenblas>http://sdow2008.semanticweb.org/ is down ... anyone around that might be able to check?
09:58:30<mhausenblas>ping CaptSolo
09:59:11<mhausenblas>logger, pointer
09:59:11<logger>See http://chatlogs.planetrdf.com/swig/2008-09-07#T09-59-11
09:59:12<ghard>mhausenblas: Site seems ok from here.
09:59:20<mhausenblas>really? hm
10:00:09<mhausenblas>ghard, are you sure its not from the cache? ;)
10:00:27<mhausenblas>even if I ping it I get a timeout
10:00:42<ghard>Just did refresh.
10:00:53<mhausenblas>ACTION notes that all other pages/services are ok so can't be my access, I guess
10:00:56<mhausenblas>right
10:00:58<mhausenblas>funny
10:01:40<ghard>wget does the trick as well.
10:02:34<ghard>ACTION ducks the black helicopters.
10:02:50<mhausenblas>;)
10:03:50<ghard>Check traceroute.
10:04:06<ghard>Maybe your ISP has a specific route that is down.
10:04:09<mhausenblas>very strange. when I curl from here it can't connect however if I use my http://sw.joanneum.at/ubiq/curl.html ubiquity command it works fine
10:04:14<mhausenblas>yeah
10:09:58<Iskr>do you people use pellet?
13:30:13<tommorris>ACTION is in a RDF/Semantic Web session at barcampbrighton3
13:52:51<Anchakor>Does anyone know any good articles on rdf vocabulary interoperability?
13:55:32<Anchakor>I know you can define the relations between rdf vocabularies with rdfs and owl, but how is it all supposed to work?
14:02:58<mhausenblas>Anchakor, can you be more specific? Example use case?
14:04:49<Anchakor>like I have collected some rdf data and I prefer some vocabularies, so I want the data converted...
14:10:04<Anchakor>so where do I get the vocabulary relations defined in rdfs/owl to convert it?
14:18:33<mhausenblas>hm Anchakor I'm still not sure what excatly you're after. Maybe you mean mapping, maybe aligning, etc.
14:19:30<mhausenblas>in the end you'll end up manually inspecting the semantics and decide which class/properties from a source vocab match the target vocab
14:21:00<mhausenblas>it would be much easier, though, if you have a concrete example handy
14:24:21<Anchakor>mhausenblas: inference is the word I think, though Im really new to this so I might be wrong :)
14:24:49<Anchakor>yes inference
14:25:56<mhausenblas>hm. with inference you can make implicte information explicit, yes, but how does this help you?
14:27:00<mhausenblas>and I still don't grasp your statement 'where do I get the vocabulary relations defined in rdfs/owl to convert it'
14:27:49<Anchakor>uhm I was more thinking that I have defined that A owl:sameAs B somewhere in internet - I call it the vocabulary relations
14:27:49<mhausenblas>any *concrete* example available?
14:28:01<mhausenblas>ah! good ;)
14:28:12<mhausenblas>*now* we get closer to the issue
14:28:55<Anchakor>then I say I prefer to get A rather then B and get it... thats inferencing I hope :)
14:29:16<mhausenblas>depending on the interpetation (I'm thinking in the line of OWL-DL now) this is on the instance level
14:30:02<Anchakor>question is where do I get the vocabulary relation data so I can do the inferecing? how is it shared on internet? is there any project developing vocabulary relations and doing quality assurance of them?
14:30:10<mhausenblas>the semantics of 'A owl:sameAs B' are precisely: A is the same individual as B (having two identifiers)
14:32:01<Anchakor>this all I mean only for vocabularies which overflap each other in their coverage
14:32:01<mhausenblas>the only 'vocabulary relation data' necessary you'd need here is a reasoner that understands owl:sameAs (for example in OWL-DL) and thus merges all the statements ...
14:32:32<Anchakor>mhausenblas: no, thats the reasoner, thats the active component of the process
14:32:58<mhausenblas>sorry Anchakor I'm not following you.
14:33:16<Anchakor>reasoner takes rdf data, vocab.rel. data, my config and spits out rdf data
14:33:23<mhausenblas>agree
14:33:26<mhausenblas>and?
14:34:00<Anchakor>some vocab.rel. data will be usefull to many people
14:34:48<mhausenblas>again, if you have two vocabularies (or ontologies for that matter ;) and there is an overlap or you want to 'convert' one into the other, there is tons of research out there
14:35:32<mhausenblas>AFAIK this is all around mapping, alignment, etc.
14:35:46<mhausenblas>did you search for this kinda stuff already?
14:35:49<Anchakor>yep
14:36:03<mhausenblas>and? nothing which helps you, or?
14:36:06<Anchakor>Im interested in small part of this
14:36:21<mhausenblas>ok. but I still don't get in which one ;)
14:36:44<Anchakor>how is the vocab.rel. data shared... are there any QA projects?
14:37:10<Anchakor>or everyone has to develop their data from scratch?
14:38:00<mhausenblas>I am not aware of any site or service that you could query for 'overlapping' vocabularies/semantics
14:38:31<mhausenblas>however, that doesn't mean it doesn't exist ;)
14:38:42<mhausenblas>in that sense: a very good question
14:40:05<mhausenblas>how about: rephrase it a bit, give an example and post it to semantic-web@w3 and also discuss@lists.vocamp.org to trigger a discussion
14:41:03<mhausenblas>I'd be happy to learn if someone already works on it or if not, we could launch an according activity
14:41:23<mhausenblas>what do ya think Anchakor?
14:43:23<Anchakor>yeah
14:43:57<Anchakor>I think there should be some tool/service on picking vocabularies
14:44:36<mhausenblas>a worthwhile starting point might be http://axel.deri.ie/~axepol/presentations/20070604_eswc2007_alem-etal-2007.pdf
14:45:06<Anchakor>eventually all vocabularies and their relation should make some kind of global ontology
14:45:16<mhausenblas>and also the according paper at http://www.eswc2007.org/pdf/eswc07-alemanmeza.pdf
14:45:39<mhausenblas>what? global ontology. argh :D
14:46:03<Anchakor>though not global ontology but infinite number of possible ontologies
14:46:16<Anchakor>the user sort of composes its own ontology
14:46:17<mhausenblas>mhm
14:46:37<mhausenblas>did you skim/check the paper/presentation I just posted?
14:46:48<Anchakor>not yet
14:47:09<mhausenblas>give it a try ... IMHO good stuff re that issue
14:47:46<mhausenblas>ACTION BIAB
14:48:43<Anchakor>anyway what is reccomended practice when I find out few things are missing in a vocabulary I wanted to use?
14:50:28<Anchakor>I would just define the missing stuff as another vocabulary... or any other options?
15:49:35<danbri>https://www.idselector.com/
15:49:45<danbri>B:|OpenID ID selector
15:50:02<danbri>B:An attempt to improve user experiences with OpenID log-in
15:50:27<danbri>B:I wonder whether the CSS history hack might also be useful ... highlight/prioritse sites that CSS shows have been visited (or is that just spooky?)
16:00:03<MacTed>Anchakor - the UMBEL project is aimed exactly at this "global ontology". part of the UMBEL ontology is meant to include (as completely as possible) owl:sameAs, owl:subClassOf, owl:subPropertyOf relations to other ontologies -- such that if you say something is an UMBEL:document, any inferencer should be able to also see that it's a sioc:container, bibo:document, etc.
17:30:34<earle>drewp: the "know" relationship on identica is going to be changing - Evan hasn't applied my patch yet.
19:18:25<Anchakor>MacTed: I was reading some stuff on UMBEL website, but I understand it's a ganaral ontology which is to be used as reference when creating yours ontologies or directly using it... my problem is that people will define relations between existing ontologies which werent referenced against UMBEL, so there is need for some kind of mechanism for organising these new relations, be it website service, or another trust software
19:21:15<Anchakor>I imagine several groups (projects) will be formed to gather the relation set which fits the group's point of view
19:21:52<Anchakor>s/gather/gather relations and creat a/
19:29:11<Anchakor>the Combining RDF Vocabularies for Expert Finding paper is very interesting
19:51:59<linxeh>probably off topic, but does anyone know of any good papers that discuss question decomposition techniques?
19:52:56<linxeh>as in for question answering

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