Experimental IRC log swig-2007-07-27

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.

00:00:25<Arnia>Anyway, with description logics it is very hard to specify role-chain equality without having the logic go undecidable. Most of the stuff I'm interested in is all about role-chain equality :)
00:00:31<Arnia>Equations for the win!
00:01:46<chimezie>Yes, the expressivity tradeoff for computational tractability (which is not all that much, in theory) is often lost in the dialog about DL
00:02:15<chimezie>but I think it is mostly a problem of DL not having intersected (much) with domains where this loss in expressivity is a major issue
00:02:56<Arnia>Well, it isn't in the papers which focus on tractability. They're very aware of the loss of expressive ability and would love to know why certain constructors together cause blow ups but not with others
00:03:10<Arnia>(in general I mean; they know why for particular cases)
00:03:23<Arnia>But... I'm not convinced the nodes matter
00:03:35<Arnia>(and DLs are all about typing the nodes)
00:03:58<Arnia>I'm coming to the opinion that typing the arrows is far more important. The nodes are just there to hang arrows off :)
00:04:34<Arnia>URIRefs give you a means to share your arrows with others and agree on a (simplistic) model of identity
00:05:05<chimezie>in addition, I'm of the opinion that tableux-based calculus has a ceiling of scalability because of this constraint. A ceiling which is non existent for different reasoning algorithms over the same 'data' - which is why I'm interested in dosbarth
00:05:47<Arnia>chimezie: tableaux reasoners rely upon the ability to construct a tableaux. Not all description logic fragments admit this ability
00:05:52<Arnia>(IIRC)
00:08:00<chimezie>ACTION would love to watch smoke rise from the hardware on which a tableux based reasoner attempted to classify http://www.cyc.com/2004/06/04/cyc, for instance
00:08:11<chimezie>but I digress
00:08:39<Arnia>chimezie: Just loading Cyc into Protegé is fun to watch :p
00:08:52<Arnia>chimezie: and that's without the ThoughtTreasure additions
00:10:18<chimezie>ahh yes, TT
00:11:02<Arnia>Anyway, you know where I'm going with this. I'm interested in separating concerns in the semantics. View RDF as a category and see what properties fall out of that simple characterisation and then represent this category in order to explore different concrete semantics
00:11:19<Arnia>(such as my lovely higher-order uncertainties etc)
00:11:47<chimezie>yes, I do :)
00:12:54<Arnia>I'm hoping it will prove a profitable strategy as well as giving others some inspiration to start thinking more about the underpinnings of the system
00:13:21<Arnia>ACTION disagrees with the wedding cake model though, so I'm biased :p
00:13:36<chimezie>you aren't the only one, Arnia
00:13:50<chimezie>i wonder if you have seen: http://www.inf.unibz.it/~jdebruijn/publications/msa-ruleml05.pdf
00:16:29<Arnia>hm... thanks for the link. My main objections are the inseparability of some of the layers (and misplacement)
00:17:20<chimezie>Yes, but at least they challenge the layering and flip the DL/Rule layering with very empirical motivations
00:17:43<Arnia>The wedding cake seems to conflate social models with information trust... in reality, the confidence (and degree of belief) we have in a statement is a vital part of the meaning of the statement
00:18:05<Arnia>Oh yes, and I heartily applaud that :)
00:18:09<chimezie>modal logics?
00:18:27<Arnia>chimezie: more fine grained. Most logics have the subtlety of a sledgehammer
00:20:43<chimezie>and yes.. I agree: binary truth values and monotonic characteristics (the FOL realm) are a very purist way of modelling the world that doesn't always reflect reality
00:20:44<Arnia>chimezie: it isn't a matter of saying whether a statement is defeasible, or a default, but rather a matter of combining information smoothly from multiple sources. With model-theoretic (and particularly, boolean) semantics you often seem to throw the baby out with the bathwater
00:22:10<Arnia>chimezie: you are stuck, if there is a conflict, with choosing between keeping your valuation or throwing it out and replacing it. You can't say 'well, I am x confident in their judgement so I'll rebalance my own'
00:23:14<Arnia>chimezie: and because the space of truth values isn't continuous you can't apply lots of formal tools to the flow of belief in statements
00:23:35<Arnia>(something else which interests me, the self-organisation of cultures by sharing of knowledge)
00:23:41<Arnia>but I'm rambling again
00:23:44<chimezie>:)
00:24:58<Arnia>I should go on holiday more often... I seem to have lots of wild ideas when I'm supposed to be resting
00:25:13<Arnia>Holiday comes to an end tomorrow though and I'm back Up North
00:33:09<timbl>#swhack? what's that? Compared to #swig?
00:33:51<dajobe>their sw is not semweb
00:34:24<Arnia>Yeah. It is the sound of Monty and Phenny running into each other at full pelt
00:34:28<danbri>what is it?
00:34:41<danbri>i always assumed it was
00:34:58<danbri>in origin i mean, not actual topics
00:35:52<Arnia>"Aaron suggested chatting in "#sbp", but Sean thought that was too egotistic, so instead Aaron came up with "Swhack!". It doesn't mean or stand for anything; it's onomatopoeic (so Aaron has always claimed)" -- http://swhack.com/faq/
00:37:20<timbl>I see
00:41:17<danbri>ACTION realises we've been hanging around here in irc 6.5 years or so
00:41:21<danbri>#rdfig included
00:41:27<danbri>time to go jabber!
00:41:52<Arnia>danbri: but IRC is so retro it's cool
00:42:13<dajobe>many projects live and thrive on IRC today, it's not outdated
00:42:22<danbri>I know x2 :)
00:42:36<danbri>it's less broken than email, at least
00:42:51<Arnia>Jabber doesn't like working through my university's firewall as well
00:43:09<Arnia>At least not from the college rooms... departments etc are fine :)
00:43:12<danbri>no? interesting, i'd not heard of universitys blocking it
00:43:15<danbri>ah
00:43:23<danbri>students should be in the library ...not chatting!
00:43:31<danbri>but they let irc thru?
00:43:39<Arnia>It isn't blocked... it is just Durham's firewall is default-deny (for various historical reasons)
00:43:57<Arnia>They don't let IRC though, but that seems to be a misconfiguration they can't trace
00:43:59<danbri>ah, right
00:44:07<danbri>:)
00:44:48<Arnia>Durham's IT services can be a little odd at times
00:45:11<Arnia>I'm getting pre-nostalgic though. One more year and I'm off into the world of postdoc. Terrifying
00:45:28<Arnia>Never thought I'd get damp-eyed over the ITS ;)
00:46:28<Arnia>Has anyone worked on putting RDF into Jabber messages incidentally?
00:54:42<danbri>yup
00:55:15<danbri>recently tidied up my docs and the ageing code still kinda works - http://svn.foaf-project.org/foaftown/jqbus/intro.html
00:55:34<danbri>needs more attention though ... binding rest of sparql protcol, and dealing with result sets too big for 1 xmpp packet
00:56:42<Arnia>ah cool. *adds it to his list of things to look at*
00:57:50<Arnia>I should really get back to my paper, although I might just call it a night
00:58:45<danbri>ACTION heads off too ...gnite
12:24:33<Pexino>someone can help me with cwm for a particolar proof validation?
16:16:30<DanC>ACTION wonders if Pexino got any help or sent something to public-cwm-talk ...
16:17:05<DanC>ACTION doesn't see anything in http://lists.w3.org/Archives/Public/public-cwm-talk/2007JulSep/thread.html
19:54:44<Arnia>phenny, tell chimezie also, composite IFPs are silly easy with this semantics.
19:54:46<phenny>Arnia: I'll pass that on when chimezie is around.

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