Experimental IRC log swig-2008-07-09

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.

06:27:30<mhausenblas>ACTION waves to kwijibo
06:27:31<phenny>mhausenblas: 08 Jul 15:36Z <MacTed> tell mhausenblas there are some blurry bits in <http://www.youtube.com/watch?v=qMjkI4hJej0> that seem like they were meant be (more) legible text or screenshots (e.g., see 2:27, 2:49, 3:00); certainly the slides are more legible <http://www.slideshare.net/mediasemanticweb/quick-linked-data-introduction/>
06:27:47<mhausenblas>phenny, thanks
06:28:07<kwijibo>g'mornin
06:29:55<mhausenblas>phenny, tell MacTed re quick linked data intro: thanks and I'm gonna update it before TimBL notes my freudian slip (2006 not 1998 ;)
06:29:55<phenny>mhausenblas: I'll pass that on when MacTed is around.
06:31:26<mhausenblas>ACTION gotta reboot PC - WinXP has some new updates for me. yipee!
06:31:37<mhausenblas>ACTION BIAB
06:52:57<mhausenblas>anyone can remember the 'most-often-used-RDF-properties' on http://pingthesemanticweb.com ?
06:53:18<mhausenblas>can't find it and I'm pretty sure they were there, somewhere
06:57:46<mhausenblas>ha! found it
06:57:48<mhausenblas>http://pingthesemanticweb.com/stats/namespaces.php
06:58:09<mhausenblas>not exactly, but that is the closest I guess ...
07:34:12<shellac>semantic web on radio 4 - now
07:34:26<libby>damn, I was gonna say that
07:34:28<libby>:)
07:35:10<mhausenblas>bengee
07:35:15<mhausenblas>around?
07:49:19<kwijibo>shellac: damn! I switched off at the sports news - what were they saying?
07:54:24<shellac>nothing that is unfamiliar :-) intro, then timbl interview
07:55:16<shellac>dunno how quick the listen again and podcast go up.
07:55:24<kwijibo>sweet
07:56:22<kwijibo>ACTION wonders if we can say "semantic web" to normal people and have people recognise the phrase
07:56:32<shellac>watch this space: http://www.bbc.co.uk/radio/podcasts/today/
07:57:28<gromgull>kwijibo, have you never had the "so what do you really work on?" conversation with grand-mothers, non-techie school-friends, etc? Their eyes glaze over pretty damn quickly when you say "semantic"
07:57:50<shellac>I still try to avoid the phrase, I must confess. I tend to use web
07:57:54<kwijibo>yeah
07:58:30<shellac>ACTION sees that the bbc include a podcast link for the zune
07:58:34<gromgull>It usually end up at - "so it's like facebook? " and I agree...
07:58:48<shellac>that zune owner is treated well
07:58:54<kwijibo>gromgull: lol :)
08:00:05<gromgull>shellac: what exactly is the difference between the three rss feeds for itunes zencast and zune? A podcast rss is clearly not a podcast rss is clearly not...
08:00:14<gromgull>I guess i could download and diff them...
08:00:23<gromgull>ah never miund
08:00:24<gromgull>mind
08:00:27<gromgull>it's the protocol
08:00:36<gromgull>and the feed is the same in the end
08:00:36<gromgull>sigh
08:03:40<kwijibo>haha
08:04:34<mhausenblas>bengee?
08:21:13<EtnaRosso>morning all
08:21:35<EtnaRosso>did someone tried Oracle Semantic Technologies?
08:21:50<kwijibo>r4 Today is listenable again: http://news.bbc.co.uk/today/hi/today/newsid_7496000/7496976.stm (ta KiYanWang)
08:22:12<KiYanWang>kwijibo: no problemo
09:18:05<danja_>phenny, tell DanC which DVCS do you currently favour - mercurial, git..? (I've an existing project currently in subversion I'd like to throw into such a system)
09:18:06<phenny>danja_: I'll pass that on when DanC is around.
09:22:55<shellac>danja_: my limited experience is that mercurial is much nicer, but the git-svn integration gives git a huge advantage
09:23:24<shellac>on jruby git-svn certainly swung it for git
11:15:37<danieljohnlewis>danja_: Quick question. I just sent you a response to your questions, did you get it? and are you ok for me to post on my blog about it?
11:19:37<bblfish>ACTION you academics you
11:19:54<bblfish>ACTION ;-)
11:22:33<danieljohnlewis>ACTION looks puzzlingly at bblfish's remark... then checks the logs to see if he has just missed something interesting
11:22:56<bblfish>ACTION no just conversation on mailing list
11:23:44<bblfish>ACTION semantic-web@w3.org to be precise
11:25:12<danieljohnlewis>ahhh, ok
11:37:38<evilbblfish>ACTION you academics, all of you
11:37:55<evilbblfish>ACTION bblfish
11:38:06<bblfish>ACTION too much fun
12:24:25<fgiasson>mhausenblas: you don't have this stat, only classes
12:35:27<rho>does anyone have some info on how to map SPARQL evaluation onto mapreduce?
12:37:16<shellac>rho: not really. I'd try basic graph patterns first, and work up from there
12:39:46<shellac>ARQ (for example) will do most of the other work for you once you have the basic elements
12:40:49<rho>shellac: something like that would have been my trajectory.
12:41:31<rho>I was only wondering how to best farm out the graph first and then how much evaluation work could be brought into MR form.
12:41:38<shellac>I think there's an rdf project around hadoop, but I haven't seen much come out of it
12:41:57<rho>I'm on their channel now.
12:43:13<rho>At this stage I am more interested in the theory (translating formal semantics).
12:44:35<shellac>if anything has been done on map reduce and relational algebra (almost certainly the case) I think most of that will carry over fairly trivially
12:49:39<bblfish>it's weird the more I write UIs that deal with rdf the more it seems like using triples directly is a good idea
12:51:00<kwijibo>bblfish: what do you mean?
12:52:09<bblfish>well in sommer I wrote a java to rdf mapper, but java objects hide too much info, and are too static to be useful in a lot of cases. It's just easier to create bindings to resources and have the UI elements query the RDF DB directly
12:53:12<bblfish>I can't quite work out the right level of abstraction, but triples are pretty good
12:53:14<kwijibo>what was the motivation for writing the object mapper in the first place?
12:53:59<danbri>there's the map-reduce-merge stuff i posted about here - http://danbri.org/words/2008/05/29/322
12:54:28<shellac>kwijiba: java land is dominated by beans, so if you can make rdf look beany it could be very powerful
12:55:53<bblfish>kwijibo: It was getting annoying to write queries using statements all the time. Sometimes mapping to java objects are very nice. Sometimes it gets in the way. With UIs statements make a lot of sense though.
12:57:21<Shepard>hmm someone emptied two pages about hbase and rdf: http://wiki.apache.org/hadoop/HRDF?action=info http://wiki.apache.org/hadoop/Hbase/RDF?action=info
12:57:31<Shepard>at least http://wiki.apache.org/incubator/HRdfStoreProposal is still there
12:57:43<shellac>yeah, I noticed that when looking for links. what happened?
12:58:34<Shepard>well on the first link they simply say it was moved to incubator
12:58:43<Shepard>still, cool URIs don't change ;)
12:59:29<shellac>the uri didn't change. its informational content just plunged
13:00:15<Shepard>yeah, but I'd have left it like that and simply added a note to the top that the page is outdated and the current content is to be found elsewhere
13:00:17<bblfish>the trick is nailing down in the java components as much of the static behavior as possible, but working with triples for everything dynamic
13:01:34<rho>danbri: thx!
13:01:57<danbri>if you figure anything out, let us know here or on semantic-web@w3.org
13:03:25<rho>sure.
13:27:34<TipTop>hello
13:28:01<TipTop>AndyS: how is the safest way to completely reset a OntModel?
13:28:22<TipTop>i wish to remove all submodels, all inferences
13:28:34<TipTop>and all statements from the OntModel
13:31:35<shellac>TipTip: probably better on #jena, but I think removeAll should do it
13:59:42<DanC>ACTION scrolls back...
13:59:43<phenny>DanC: 09:18Z <danja_> tell DanC which DVCS do you currently favour - mercurial, git..? (I've an existing project currently in subversion I'd like to throw into such a system)
13:59:54<DanC>danja, mercurial. http://homer.w3.org/~connolly/projects/?sort=lastchange
14:03:57<danja_>thanks!
14:08:38<LotR>DanC: what does mercurial have over git?
14:11:02<DanC>source code that smells good.
14:11:51<DanC>I gather the windows port of mercurial is more straightforward and reliable than git
14:12:29<DanC>the mercurial command set is easier to learn, too. (reportedly; I haven't tried very hard to learn git)
14:12:55<dajobe>I still find git commands hard to remember/use (occasional use)
14:13:19<DanC>I'm a sucker for losing technologies, though. http://advogato.org/person/connolly/diary/47.html
14:13:47<DanC>"worse is better" probably applies to git vs hg too
14:14:01<dajobe>git was based on hg IIRC
14:14:39<DanC>the claims go both ways; I researched the history pretty closely and found they're completely contemporary
14:15:00<DanC>i.e. they were both inspired by the bk fiasco, but the first version of each was developed independently
14:15:22<DanC>I think wikipedia stabilized on this version of history... let's see...
14:15:22<dajobe>heh
14:17:03<shellac>they were both create for linux devel, from what I remember. I though git was a direct response to hg -- a 'nah, you don't want to do it that way' thing
14:18:01<DanC>wikipedia doesn't support that story, shellac
14:18:25<DanC>first design notes on git were in a mail message from linus April 7, 2005 ;
14:18:33<DanC>"Mackall first announced Mercurial on April 19, 2005."
14:19:41<kendallgclark>I find git far more useful in practice because of git-svn. Unless I'm doing or paying for it to be ported to Windows, I don't give two figs whether it's hard or not. :)
14:21:22<DanC>I use hgsvn quite a bit too
14:22:00<DanC>ACTION looks up git-svn, wondering to what extent it can put the toothpaste back in the tube...
14:22:33<kendallgclark>does hgsvn come with hg or is it an add-on?
14:22:51<kendallgclark>i used hg for a while and don't have hgsvn on my box and have never heard of it
14:22:55<shellac>git-svn is a great deal more useful than hgsvn
14:23:03<kendallgclark>(which is probably evidence of my ignorance more than anything else)
14:23:17<DanC>hgsvn doesn't come with hg
14:23:19<kendallgclark>git-svn makes svn bearable IMO :)
14:23:22<LotR>DanC: git-svn works pretty well two-way
14:23:44<kendallgclark>and it saved a protracted debate in my shop about whether we should move off SVN to one of the DVCSes quite neatly
14:24:17<DanC>looks like the git feature set and popularity are sufficient to overcome the "git is a mess of shell scripts" problem; I see "Eclipse IDE-based Git client, based on a pure Java implementation of Git's internals" and such.
14:24:33<kendallgclark>DanC: I agree w/ you about one thing (at least!): if i go a few days w/out using git, I tend to forgot everything and have to re-read one of the 'getting started' things...So I printed one and taped it near my desk.
14:24:53<kendallgclark>There's also a Ruby git client now, etc
14:32:12<dajobe>when the emacs build I used came with git support out of the box (tarball)) it meant I didn't even need to change keystrokes vs cvs or svn. that was nice
14:32:54<DanC>there's emacs support for hg, but it hasn't reached "out of the box" status.
14:33:35<DanC>stuff like joey's "there can be only one" suggests hg has sorta peaked. http://joey.kitenet.net/blog/entry/there_can_be_only_one/
14:35:24<fidothe>have you guys seen http://github.com/ ? It was that which made me defect from bzr and svn full time to git. The social/technical infrastructure around git seems to be really building
14:40:48<DanC>yes, github is another example of the network effects around git
14:41:51<dajobe>wasn't it down this week?
14:42:41<DanC>speaking of the risksof centralized services, identi.ca looks fun
14:43:00<DanC>meanwhile, via tim bray on twitter, " Vote on your favorite bug-tracker & SCM, & pass the word: http://tinyurl.com/5kof8c"
14:48:47<yvesr>ACTION is quite happy with git and git-svn
15:49:05<danbri>re git etc
15:49:08<danbri>Shall I pull this patch? (1/2) [ynWvpxqadjk], or ? for help
15:49:13<danbri>made be grin, trying to learn darcs
15:49:25<danbri>in fairness, the help is helpful.
19:14:25<bblfish_>is aol.com down?
19:19:19<DanC>works for me
19:20:12<bblfish_>ah well I can't ping aol.com, perhaps a local name resolution problem
19:21:44<DanC>I can't ping aol.com either; perhaps an administrative choice on their part
19:24:11<bblfish_>DanC: ah well I can't nslookup them either
19:26:08<DanC>ah. I can nslookup.
19:27:21<bblfish_>I was wondering because I can no longer use aol instant messenger
19:28:03<bblfish>ACTION very tired, got to go and rest
19:35:31<edd>bblfish: http://downforeveryoneorjustme.com/aol.com

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