Experimental IRC log swig-2008-08-25

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.

00:17:33<hhalpin>dajobe - you still around?
00:18:51<earle>doesn't seem like it, but if you need a Perl person...
00:19:05<hhalpin>Yep, I'm not much of a perl person.
00:19:11<hhalpin>But have you ever used Redland with perl bindings?
00:19:24<hhalpin>having some interesting issues re parsing lots of files at once.
00:19:46<hhalpin>In particular, these are automatically harvested files.
00:19:58<hhalpin>So, the RDF in them is not well-formed all the time.
00:19:58<earle>Haven't gotten around to, yet, no. But I can advise on Perl oddities :)
00:20:03<hhalpin>And sometimes there's not even RDF in the file.
00:21:03<earle>Is that the case at the other end, or do you suspect a harvesting problem?
00:21:26<hhalpin>The other end.
00:21:40<earle>Right.
00:21:42<hhalpin>Anyways, the problem is that sometimes it causes the Redland parser to fail with a silly error.
00:22:01<hhalpin>Such as when it attempts to parse a HTML file as RDF.
00:22:12<hhalpin>Or worse, when someone has used rdf:RDF as a property value.
00:22:17<hhalpin>Or some other silliness.
00:22:25<hhalpin>Believe it or not, that happens in the wild.
00:22:42<hhalpin>So...the question is how to get the parser to ignore these error cases and log them.
00:22:46<hhalpin>According to dajobe:
00:23:57<hhalpin>My parser uses a "parse_as_stream"
00:23:58<hhalpin>function
00:24:09<hhalpin>which " Returns an RDF::Redland::Stream of RDF::Redland::Statement objects or undef on failure."
00:24:17<hhalpin>Now, I am trying to simply exit upon failure.
00:24:25<hhalpin>s/trying/NOT trying/
00:24:40<earle>Hmm... I'd start by trying to eliminate the empty/non-RDF files first, before handing the data over to the parser - minimize the number of errors to deal with.
00:25:10<hhalpin>that's easy.
00:25:14<hhalpin>the problem is malformed RDF.
00:25:45<hhalpin>In particular, I want to check to see if the parser is undefined.
00:25:54<hhalpin>Yet it appears that the errors come up only in the stream.
00:26:00<hhalpin>Not in the initial parsing stage.
00:26:07<hhalpin>Since, it is, after all, a stream.
00:26:37<danja_>ACTION curses, too tired to natter with hhalpin
00:26:48<hhalpin>if there is an error, are the individual triples undefined.
00:27:06<hhalpin>Is there anyway of intercepting "Redland error:" and logging them, without disrupting the entire batch processing job?
00:27:13<hhalpin>Then I can go over them one by one later for clean-up.
00:27:42<danja_>not that I've seen - ping dajobe
00:27:44<earle>Hm, I suppose you could redirect and log STDERR.
00:28:03<earle>Presuming that's where the errors get printed to.
00:28:10<hhalpin>the problem is Redland errors halts the entire program.
00:28:21<earle>Oh, right, yes, I see what you mean now.
00:28:21<hhalpin>So even with redirect to STDERR I get one error at a time...
00:28:31<danja_>nn
00:28:34<hhalpin>I'd like to "intercept and ignore, yet write down" all errors.
00:28:38<earle>Okay, you need eval, I think.
00:28:50<hhalpin>Trying to figure out where to do that at in Redland Perl bindings.
00:28:55<hhalpin>eval as in Lisp?
00:28:58<hhalpin>Perl has that too??
00:29:17<earle>I'm not a Lisp hacker, but - eval is "try" for Perl.
00:29:21<earle>AIUI.
00:29:23<earle>perldoc -f eval
00:29:41<hhalpin>ah, it's sort of the same.
00:29:56<earle>eval { do some stuff; }; if ($@) { print "We had a problem! $@"; }
00:30:16<hhalpin>The question I have for dajobe is what precisely to put that "eval" around.
00:30:16<earle>Not sure if that'll stop the parser from exiting, though.
00:30:20<earle>Right.
00:30:41<hhalpin>That's another good question.
00:31:03<hhalpin>Hmmm...
00:31:05<earle>Can you bung a bit of the nearby code into swig.pastebin.com?
00:31:24<earle>Maybe I can suggest something to try.
00:31:50<hhalpin>I'll upload 3 problem files and throw all the code in.
00:32:17<earle>ACTION reads docs - "Returns an RDF::Redland::Stream of RDF::Redland::Statement objects or undef on failure."
00:32:28<earle>Right, that's what you pasted.
00:32:55<earle>ACTION checks source...
00:32:56<earle> my $stream=&RDF::Redland::CORE::librdf_parser_parse_string_as_stream($self->{PARSER},$string, $rbase_uri);
00:32:59<earle> return undef if !$stream;
00:33:13<hhalpin>The problem is that the parser seems to be returning a stream even if there is some failure parsing.
00:33:20<earle>Interesting.
00:33:21<hhalpin>then bringing up the parsing failure in runtime.
00:39:09<hhalpin>there I just put all the files online.
00:39:32<hhalpin>The program is, in extremely hacky perl:
00:39:34<hhalpin>http://www.ibiblio.org/hhalpin/homepage/notes/rdfstats.pl
00:39:47<hhalpin>Some example problem files are:
00:40:08<hhalpin>http://www.ibiblio.org/hhalpin/homepage/notes/411.rdf
00:41:02<hhalpin>which causes "Redland error: rdf:RDF is forbidden as a property element."
00:41:15<hhalpin>http://www.ibiblio.org/hhalpin/homepage/notes/411___disambiguation__.rdf
00:41:29<hhalpin>which causes the more mysterious "Redland error: rdf:RDF is forbidden as a node element."
00:41:41<hhalpin>Hmmm...
00:41:55<hhalpin>just trying to get some basic stats out of some linked data sampled from the wild.
00:42:55<dajobe>all my demos use perl and eval to catch errors
00:43:27<hhalpin>dajobe - what would you put eval around though?
00:43:49<dajobe>$parser->parse_into_model
00:44:25<hhalpin>Since I have no desire to store any of these triples anywhere, I am using $parse->parse_into_stream
00:44:53<dajobe>that'll be slow 'cos it'll create lots of objects you immediately throw away
00:45:00<dajobe>just make an memory model
00:45:09<dajobe>(I guess I should have a sink model too)
00:46:08<dajobe>actually I seem to use the error handler callbacks in perl
00:46:14<dajobe>see http://svn.librdf.org/view/librdf/trunk/demos/demo.pl?revision=14172&view=markup
00:46:34<dajobe>I mean http://svn.librdf.org/view/librdf/trunk/demos/rss-show.pl?revision=14172&view=markup - and serach for set_error_handler
00:47:43<hhalpin>looking dajobe
00:48:29<dajobe>triplr uses the eval method, but that source code isn't public
00:48:39<hhalpin>ok.
00:49:04<dajobe>ah, there's the detailed error handler too
00:49:07<dajobe>RDF::Redland::set_log_handler($handler);
00:49:15<dajobe>you get line numbers, levels, .. etc.
00:49:31<dajobe>my $handler=sub {
00:49:31<dajobe> my($code, $level, $facility, $message, $line, $column, $byte, $file, $uri)=@_; ... };
00:50:34<dajobe>see http://librdf.org/docs/pod/RDF/Redland.html
00:50:42<hhalpin>My baseline is just parsing a file into a stream of statements, and if there are any errors in parsing or in violation of the RDF spec, just log those and keep going, without exiting my entire script.
00:50:57<hhalpin>perhaps i could just set this using parser settings?
00:51:23<dajobe>eval will handle that
00:51:43<hhalpin>sounds like eval is what I want, but I am still confused about what to wrap it around.
00:52:00<hhalpin>my $stream = $parser->parse_as_stream ($uri, $uri);?
00:52:07<hhalpin>or
00:52:08<hhalpin>my $statement=$stream->current;
00:52:11<dajobe>and the entire loop tha tfollows
00:52:23<dajobe>any of them could throw the error
00:52:30<hhalpin>Ah, that's what I wanted to know.
00:53:00<hhalpin>that the initial parse of "parse_as_stream ($uri, $uri)" is, after all, streaming.
00:53:04<hhalpin>So no errors will be caught there.
00:53:12<hhalpin>Only at triple level, inspecting triples, etc?
00:53:14<hhalpin>right?
00:53:18<dajobe>that callt does no actual parsing
00:53:22<dajobe>until you call ->current
00:54:03<hhalpin>Does -current do all the parsing of a 'single triple' at once?
00:54:07<hhalpin>i.e. would errors show up there?
00:54:14<dajobe>yes
00:54:14<hhalpin>I'd ideally want to catch the errors at the triple level.
00:54:21<dajobe>that doesn't make sense
00:54:27<dajobe>a single bit of content may generate multiple triples
00:54:44<hhalpin>i thought the stream was a stream of triple statements...
00:54:54<dajobe>yes
00:55:00<dajobe>but the error isn't with the triple
00:55:04<dajobe>it's with the *source* you parsed
00:55:20<hhalpin>But that source is parsed here: 'my $statement=$stream->current;' no?
00:55:59<hhalpin>So if I catch that and can grok any before and after triples, then I can kinda guess where error came from, or is this is a bad idea?
00:56:00<dajobe>and the error that is thrown will refernece the source content
00:56:15<dajobe>there is no 'where' in triples, they are a set
00:56:33<hhalpin>s/where/where in content
00:56:34<dajobe>the structured error handler, which I mentioned above, will give the 'where'
00:56:39<hhalpin>s/where/where in content of source
00:56:51<hhalpin>ok, let me try that eval trick.
00:56:58<hhalpin>makes me happy to know perl has eval too!
01:00:20<earle>eval is an essential testing tool
01:01:07<earle>perl is more sophisticated than meets the eye :)
01:09:50<hhalpin>hmmm...still playing with it.
01:11:38<hhalpin>quick perl question.
01:11:43<hhalpin>I'm still getting some redland errors.
01:11:53<hhalpin>despite having eval {my $statement=$stream->current;};
01:12:21<hhalpin>I'd like to have the line number in the perl program that caused this error sent to me.
01:12:34<hhalpin>I know how to do this in other programming environments (java etc.) but not perl.
01:12:45<hhalpin>is this a setting in redland, or can I just do it in perl?
01:13:17<hhalpin>i.e. my silly parser is trying to parse a HTML file as RDF
01:13:32<hhalpin>and this is not being caught by $stream->current it seems
01:13:35<hhalpin>but somewhere else.
01:14:07<hhalpin>I'm not familiar with exception handling in perl...
01:15:36<hhalpin>Ah, I see exception-handling objects are now part of perl.
01:15:38<hhalpin>Sigh.
01:15:56<hhalpin>Perhaps this would just be easier if I gave up on streams and just parsed into memory, and then iterated.
01:24:19<hhalpin>hmmm..
01:25:52<hhalpin>OK, got it working.
01:26:02<hhalpin>When it doubt, it seems the rule with perl is to put eval around everything :)
01:45:22<hhalpin>earle - you still there?
01:45:26<hhalpin>Got a quick eval question.
01:45:45<dorian>heh
01:45:48<dorian>whaaat
01:46:01<hhalpin>Basically, it seems like once it goes off once, it never re-evaluates the block of code, despite it being inside a well, while loop.
01:46:06<hhalpin>I did reset $@ back to null.
01:46:15<hhalpin>Strange.
01:46:23<hhalpin>Any ideas? I am a perl newbie.
01:46:31<dorian>is your break condition for the while loop to check $@?
01:46:45<hhalpin>Nope.
01:46:55<hhalpin>I want to re-eval each time.
01:46:59<dorian>are you using strict and fatal warnings?
01:47:05<hhalpin>no.
01:47:51<earle>Hello again.
01:48:02<dorian>imho doing that catches pretty much all the dumb errors you can make in perl, if you can handle the pedantry
01:48:13<hhalpin>what would cause eval to not re-evaluate a code block if eval was in a loop?
01:48:43<dorian>is there a state within the thing you're actually evaling?
01:48:56<earle>hhalpin: a small paste will be much easier to answer that from
01:49:07<dorian>yeah toss it in a pastebin of some sort
01:49:45<earle>swig.pastebin.com is what I use in here...
01:51:01<hhalpin>http://www.ibiblio.org/hhalpin/homepage/notes/rdfstats.pl
01:51:20<hhalpin>hitting scp is easier for me than swig, especially as I am doing this in a term window on a remote machine.
01:51:42<hhalpin>warnings not helping.
01:51:49<hhalpin>And yes, there is quite a lot of state inside the eval.
01:54:22<earle>Ah - yes, that eval's really much too big, at first glance.
01:55:26<earle>my $stream = $parser->parse_as_stream ($uri, $uri); # Is it this line where the parser dies if there's a problem?
01:55:50<earle>I strongly recommend "use strict; use warnings;", by the way.
01:56:43<hhalpin>re earlier conversation with dajobe, the parsing seems to happen everywhere except parse_as_stream.
01:56:57<earle>Ah. Sorry, didn't catch the scrollback there.
01:57:14<hhalpin>The problem is that the eval, once an error is found, is never evalling again.
01:57:17<hhalpin>why would that be?
01:57:36<hhalpin>that is odd behavior for a try/catch mechanism.
01:58:04<earle>It suggests to me that you are experiencing side effects from something else within the block.
01:58:22<hhalpin>ah, for side-effect free programming :)
01:59:21<earle>Just a moment - I'm going through your code.
02:00:44<hhalpin>apologies for its messy state.
02:00:59<hhalpin>it is re-running eval.
02:03:42<hhalpin>hmmm...
02:04:45<earle>FYI: $hash_datatypes{$uri_string} = $hash_datatypes{$uri_string}+1; # $hash_datatypes{$uri_string}++;
02:06:00<hhalpin>wait a sec...
02:06:03<hhalpin>i think i found it.
02:06:11<hhalpin>once the redland parser hits an error in a streaming operation.
02:06:18<hhalpin>if *the same parser* is used again.
02:06:22<hhalpin>it refuses to parse.
02:06:51<earle>interesting
02:07:02<hhalpin>I just moved "my $parser = new RDF::Redland::Parser (undef, "application/rdf+xml")" into eval.
02:07:06<hhalpin>And it seems to work now.
02:07:21<hhalpin>Cross fingers, running it on a few hundred thousand triples now.
02:07:25<earle>I guess the error persists.
02:07:30<hhalpin>Yep.
02:07:35<earle>other minor point (while I'm reading): &foo(); # you don't need the & any more
02:07:40<hhalpin>Ah.
02:07:41<dajobe>error=falure. that's correct, wich is why parse_into_model shuld be used as it'll generate lots of errors in 1 call. if you call that good
02:08:10<hhalpin>I had worries about memory problems in this job.
02:08:12<hhalpin>which is why I used streaming.
02:08:27<hhalpin>But I guess that was perhaps a bit premature.
02:08:43<earle>For clarity, I'd suggest moving everything inside the loop into a sub.
10:39:33<bblfish>hi
10:40:23<mhausenblas>ACTION waves to bblfish
10:41:05<bblfish>ACTION waves back
10:45:14<bblfish>are the havard people back yet?
10:45:18<bblfish>from holidays
10:45:25<bblfish>mhh I mean MIT
10:45:44<bblfish>I had a question about Tabulator
10:45:48<mhausenblas>ACTION wondering. MIT chaps go on vacactions. Really?
10:46:23<bblfish>ACTION it seems to have been pretty quiet here recently
10:46:41<mhausenblas>bblfish, please ask (/me may not be able to help but happy to learn :)
10:47:56<bblfish>well I am looking forward to the default tabulator supporting N3, and the plugin working with the latest Firefox by default
10:49:06<mhausenblas>+1
10:49:21<mhausenblas>... but that was not a question, right? :)
10:51:13<bblfish>well, perhaps there is something I was missing
10:51:21<mhausenblas>not AFAIK
10:51:26<mhausenblas>ping timbl
11:04:46<bblfish>l8r perhaps, it's still a bit early now
11:06:24<mhausenblas>right. something completely different: will we meet next week at TRIPLE-I?
11:07:19<bblfish>in Austria. I don't have a ticket to go there, nor have I had time to prepare a paper
11:07:31<mhausenblas>mhm. how sad :(
11:08:16<bblfish>yes, well we can communicate well over the internet withoug polluting the world
11:08:37<mhausenblas>yeah!
11:09:06<bblfish>I'd like to go, but I am finding it already difficult enough to get ahead with my work... I think I'll have time later this year or early next year to go to a few confs.
11:09:18<mhausenblas>maybe ISWC08 in October then ...
11:11:10<drrho>ACTION has just printed company profile catalog of Triple-I 2007. Hmmm.
11:11:52<mhausenblas>drrho, I'd use TRIPLE-I08 ;)
11:12:14<drrho>ACTION likes to look into the past ;-)
11:12:19<mhausenblas>ah
11:12:29<drrho>is it worth going?
11:12:30<mhausenblas>ACTION likes to look into the future :D
11:12:52<mhausenblas>well, bit unfair from my POV to tell as I'm located in Graz
11:13:10<mhausenblas>however, yes, goof keynotes (such as Tom Heath from Talis)
11:13:21<drrho>ah!
11:13:23<mhausenblas>and some nice tracks (e.g. the Triplification Challenge)
11:13:56<drrho>In terms of academic/industry. What is the pitch? 50:50?
11:14:31<mhausenblas>hard to tell as industry track is German speaking hence addressing another bunch of people
11:15:12<mhausenblas>however, I'd say 60% academia and 40% industry (mainly German speaking countries)
11:16:48<drrho>well, I speak German pretty well (I'm Austrian ;-)
11:17:06<drrho>ACTION more interested in cooperations and product development.
11:17:52<mhausenblas>drrho, yep, in that case good choice (and Vienna - Graz is just a tiny jump, right ;)
11:18:21<drrho>right. So you are in Graz? Knowledge Center or something like that?
11:18:43<mhausenblas>drrho, you may also want to consider to join us at the Web of Data Practicioners Days in late October (in Vienna!)
11:19:07<mhausenblas>drrho, no. JOANNEUM RESEARCH
11:19:08<drrho>Aha! What's that?
11:19:17<mhausenblas>.g Web of Data Practicioners Days
11:19:17<phenny>mhausenblas: http://upcoming.yahoo.com/event/868767/
11:19:28<mhausenblas>yes
11:21:25<drrho>ACTION will definitely try.
11:21:39<mhausenblas>right :)
11:21:48<mhausenblas>ACTION BIAB ... off for lunch
11:21:51<drrho>Danny Ayers. And Andreas Blumauer, the usual suspect -;)
13:08:50<mhausenblas>danbri around?
13:09:36<mhausenblas>.seen danbri
13:09:36<phenny>mhausenblas: I last saw danbri at 2008-08-25 09:50:08 UTC on #talis
13:17:15<mhausenblas>logger, pointer
13:17:15<logger>See http://chatlogs.planetrdf.com/swig/2008-08-25#T13-17-15
13:17:19<mhausenblas>I'm wondering if it's worth creating a W3C SWIG Note re 'RDF and HTML', basically using http://infomesh.net/2002/rdfinhtml/ as a base and extending it with recent work such as GRDDL, etc.
13:19:06<mhausenblas>phenny, tell danbri: see http://chatlogs.planetrdf.com/swig/2008-08-25#T13-17-15 - what do you think :) ?
13:19:06<phenny>mhausenblas: I'll pass that on when danbri is around.
13:20:15<danbri>sounds plausible. got any time to work on it?
13:20:15<phenny>danbri: 13:19Z <mhausenblas> tell danbri see http://chatlogs.planetrdf.com/swig/2008-08-25#T13-17-15 - what do you think :) ?
13:20:39<mhausenblas>danbri, sure. would I ask otherwise ;) ?
13:21:52<danbri>http://esw.w3.org/topic/SwigAtTp2008
13:22:01<danbri>E:|Draft agenda for SWIG F2F at TPAC
13:22:25<danbri>E:We had some mail discussion on this a while back, room's booked, agenda is still somewhat fluid if you have proposals. Mail me cc: Ivan asap.
13:22:39<danbri>E:Please put SWIG in mail Subject: so I can find it!
13:22:50<danbri>mhausenblas, do you plan to be at the Tech Plenary?
13:22:50<mhausenblas>damn, I can't be there (conflicting with WOD Practitioners Days preparation)
13:22:56<danbri>yeah, feared as much
13:23:05<danbri>we can have a live irc linkup
13:23:15<mhausenblas>*that* would be great
13:23:38<danbri>one thing i wanted to talk about, is having more local meetups and scaling the community that way
13:23:48<danbri>semcamps, barcamps, foafmeets, vocamps etc
13:24:01<danbri>there are plenty of local groups but we don't do much to sync up
13:24:05<mhausenblas>right. good one
13:24:46<mhausenblas>danbri, do you think it makes sense to start work on such a note now or after the F2F?
13:25:41<danbri>drafting asap in wiki makes some sense
13:25:53<danbri>i was thinking to make a draft of using the data- prefixes
13:26:08<danbri>even if its ugly, it shows willingness to compromise and begin getting some test data together
13:27:02<mhausenblas>ok, so let me see: (1) set up draft at ESWiki, (2) send out to swig (3) see what happens, right? :)
13:27:41<danbri>sounds good to me. you might sync with ralph/ivan to be sure
13:27:53<mhausenblas>'k
13:28:14<danbri>feel free to cc me on anything
13:28:17<danbri>or nudge in irc etc
13:28:21<danbri>as always :)
13:28:46<mhausenblas>sure ;)
15:56:17<tommorris>http://getsemantic.com/wiki/Semantic_Web_in_Government
15:56:43<tommorris>F:|Semantic Web in Government (getsemantic.com)
15:58:02<tommorris>F: Thought it would be a good idea if we started putting together a document that we could use when talking to governments, public service organisations and political representatives to say "this is the sort of thing we want". Good practices, places where Semantic Web technologies (both RDF and others) can be used, case studies, links etc.
15:58:22<tommorris>F: Either add stuff to the document or hassle me and I'll do it for you. ;)
15:58:40<tommorris>F:->swig,government,publicservices,politics
16:04:23<tommorris>earle: thanks for your e-mail the other day. The answer is "hopefully". (sorry, I have an aversion to writing e-mails)
16:04:57<tommorris>My VoCamp attendance is dependent on transport and this year's university timetable.
16:06:17<tommorris>I'll probably be able to come on the Wednesday, but not sure about Thursday.
16:08:22<earle>(Semi-)Hurrah.
16:30:09<kidehen>tommorris: what about "They Work For You" ? Note we already RDFIze that
16:30:09<phenny>kidehen: 23 Aug 19:31Z <sbp> tell kidehen if you mean that you're having the OpenID problem on the DIG weblog, I'm afraid I'm not affiliated with that weblog so I can't help; but DanC or someone who is affiliated with it, and usually around on #swig, ought to be able to help you
16:30:27<tommorris>kidehen: sure, I'll add that
16:31:16<kidehen>tommorris: note Diane Abbott graph used in example at; http://ode.openlinksw.com/example.html
16:31:32<tommorris>kidehen: have you got a page up about RDFization of TWFY?
16:32:13<kidehen>tommorris: it's part of our collection (30+) RDFizer Cartridges
16:32:33<kidehen>tommorris: so if you use ODE and do "View | Linked Data Sources
16:32:47<kidehen> tommorris: on the TWFY pages you get RDF
16:33:12<kidehen>tommorris: but RDF in Linked Data form (produced on the fly as opposed to the usual warehouse)
16:33:23<tommorris>We should probably poke the TWFY people into publishing RDF.
16:33:42<kidehen>tommorris: I did, and then ended up where we are now :-)
16:34:06<tommorris>It's strange,because it should be pretty easy.
16:34:19<tommorris>just get their code, write a bunch of patches for it and deploy
16:34:22<kidehen>tommorris: ultimately, GRDDL is what these folks need with a URI into an RDFization service
16:34:47<kidehen>tommorris: yes, but that's warehousing, we need the Linked Data to be change sensitive
16:35:05<tommorris>GRDDL is a good approach for this data. I'll have a poke at the code first.
16:35:12<kidehen>tommorris: anything we can produce "on the fly" we can produce in warehouse form etc..
16:35:39<tommorris>EEEK! It's all written in Perl! And it's hosted in a CVS repository. Evil.
16:35:56<kidehen>tommorris: Yes, re GRDDL so that the first step is (dare I say it) putting the @profile bit in the <head/> and then the associated tranform URI
16:36:21<kidehen>tommorris: Yes, as I said, been there, and arrived at "on the fly" as the solution
16:36:29<kidehen>tommorris: have you used ODE at all?
16:36:40<tommorris>Not in any detail, no.
16:36:54<kidehen>tommorris: you can get the latest cut at: http://ode.openlinksw.com/example.html
16:37:03<kidehen>tommorris: sorry, http://ode.openlinksw.com/#Download
16:37:22<kidehen>tommorris: do try :-)
16:37:38<tommorris>It's on my list of things to do
16:37:53<kidehen>tommorris: this is very much the foundation layer of something much much bigger
16:38:17<tommorris>Any idea of what TWFY should be linked to?
16:39:11<tommorris>.seen iand
16:39:11<phenny>tommorris: I last saw iand at 2008-08-22 15:55:23 UTC on #talis
16:39:20<tommorris>danbri: ^^
16:40:11<tommorris>ACTION thinks it would be awesome if phenny could check twitter to see if a person still exists
16:40:24<tommorris>danbri: iand updated his twitter 3 hours ago
16:40:53<danbri>thanks
16:41:39<tommorris>kidehen: http://ode.openlinksw.com/example.html links to /index.html but the resource does not exist
16:41:51<tommorris>it should link to '/' instead.
16:44:08<tommorris>kidehen: updated wiki page, now points to TWFY and http://ode.openlinksw.com/example.html
16:54:35<kidehen>tommorris: k will have them fix it
17:00:00<kidehen>tommorris: fixed
17:00:06<tommorris>cool.
17:00:28<kidehen>tommorris: long story, we had to use a vsp (virtuso server page) to cater for IE and it's problems with xhtnl
17:00:35<kidehen>tommorris: you know the story
17:01:53<tommorris>On non-client, non-vital stuff, I've stopped worrying about IE and have just started serving XHTML - with a .xhtml extension and all
17:01:59<kidehen>tommorris: thnx!
17:03:11<tommorris>like here: http://conferences.opiumfield.com/railscamp08uk/twetter_backup.xhtml
17:03:29<tommorris>(ooh, no doctype. naughty.)
18:21:54<tommorris>I just found out that Blackboard publishes FOAF.
18:31:58<Anchakor>hmm anyone thought about semantic communication?
18:33:05<runa>heyas
18:34:59<runa>if I want to make a sparql query for a resource and related resources, I need to use union? ie, SELECT ?movieName, ?actorName {{ ?movie rdf:type movie:film. ?movie rdfs:label ?movieName. } UNION { ?movie rdf:type movie:film. ?movie movie:actor ?actor . ?actor rdfs:label ?actorName} }
18:41:45<DanC>kidehen, you had some trouble with breadcrumbs foaf/openid?
18:41:46<phenny>DanC: 23 Aug 17:56Z <kidehen> tell DanC any idea why this OpenID/FOAF profile page URL: http://myopenlink.net/dataspace/person/kidehen> no longer works with DIG? It used to work as per GGG post.
18:43:34<DanC>hmm... http://dig.csail.mit.edu/auth/breadcrumbs is down to 3 openids
18:45:35<tommorris>runa: what do you want the output to be?
18:45:54<tommorris>you may find that using CONSTRUCT then iterating through the output is easier
18:46:24<runa>tommorris: let's see what construct does
18:46:57<tommorris>CONSTRUCT lets you basically create a graph of the triples that match certain parameters
18:47:34<runa>tommorris: ua. great. I think I need to buy a SPARQL book. any recommendations?
18:48:12<tommorris>well, there really isn't a SPARQL book - just the specifications
18:48:31<kidehen>DanC: yes, having OpenID issues with DIG
18:48:39<tommorris>Semantic Web for the Working Ontologist is meant to be good, but (a)I haven't read it and (b) I don't know how much it covers SPARQL
18:49:10<DanC>looks like the openid stuff on DIG is kerflewey. I haven't looked at it in ages
18:49:12<tommorris>there's a nifty little quick reference thoguh
18:50:02<tommorris>runa: http://www.dajobe.org/2005/04-sparql/ - print it out and carry it with you everywhere, in case the need to SPARQL arises.
18:50:31<tommorris>Actually, speaking of, (dajobe: ping!) has the reference been updated to match the REC?
18:55:05<kidehen>DanC: do you know who administers it?
18:55:45<kidehen>DanC: it did work nicely around the time of timbl's GGG post, but I suspect, platforms have since changed?
18:56:13<kidehen>DanC: btw - volleyball finals at olympics were a lot of fun!
18:56:22<tommorris>.talk superfly
18:56:24<tommorris>whoops
18:56:32<DanC>back in 2007-10 when I wrote http://dig.csail.mit.edu/breadcrumbs/node/206 , Ryan Lee was keeping it going. I haven't heard from him in ages
18:57:23<DanC>yeah... last thursday, I watched the whole sand volleyball mens final (went the full 3 sets) and usa-vs-russia semis indoors (went the full 5 sets). long night!
18:57:32<DanC>and I watched the finals vs brazil last night
18:57:38<DanC>and I watched the womens final.
18:57:40<kidehen>DanC: ditto! Awesome game!
18:57:42<DanC>lots o gold!
18:57:55<DanC>I didn't watch any of the women's indoor
18:58:07<kidehen>DanC: I caught some of that too
18:58:45<DanC>I'm still scratching my head, wondering how russia and brazil blew it so badly. the USA didn't seem to have enough firepower to beat them
18:59:14<DanC>how could russia get blocked on match point?!?!
19:00:27<tommorris>ACTION has not watched even a single second of Olympics coverage
19:01:05<dajobe>tommorris: not updated it, I suppose I should
19:03:37<DanC>tommorris, is skipping the olympics normal for you? or was this year special?
19:04:24<tommorris>DanC: it's fairly normal, but the whole Chinese government organ harvesting human rights stuff sealed it
19:04:34<DanC>ah
19:05:34<tommorris>I think http://en.wikipedia.org/wiki/Reports_of_organ_harvesting_from_live_Falun_Gong_practitioners_in_China should be mandatory reading
19:06:42<DanC>yeah, I read some of that stuff when the torch protests were going on.
19:09:38<kidehen>DanC: but US did hold their heads together while the others didn't :-)
19:10:26<DanC>yup... a few mistakes here and there made all the difference
19:11:08<kidehen>DanC: still don't grok why most don't find Volleyball entertaining in the U.S., it certainly isn't boring
19:11:18<kidehen>Danc: extremely aerobic
19:12:08<bengee>tommorris, is rena for ruby, or for rails?
19:12:08<phenny>bengee: 11 Aug 16:04Z <kidehen> tell bengee why no data at: http://cb.semsol.org/explore.rdf ?
19:12:11<phenny>bengee: 11 Aug 16:05Z <kidehen> tell bengee responding to what our sponger sniffed out at: <link rel="alternate" title="RDF/XML" type="application/rdf+xml" href="http://cb.semsol.org/explore.rdf" />
19:12:39<tommorris>bengee: it's for Ruby. I'm hoping someone will do something clever with Rena and Rails.
19:12:58<bengee>'k, thx
19:13:07<tommorris>ActiveRDF integration would be nice
19:13:37<DanC>popularity of volleyball in the USA seems to be strongly impacted by performance at the olympics; e.g. following the Karch/Timmons '88 team, the AVP got pretty big... then fizzled 5 or 10 years later
19:13:48<tommorris>Also, Rena currently sucks a lot - I'm rewriting the parser to make it not suck
19:14:06<kidehen>DanC: so should have an upswing this time around, ditto swimming etc.
19:14:07<bengee>ACTION is being interviewed by crunchbase folks (who use rails), would like to hint at agile rdf options
19:14:48<tommorris>I'd strongly advise against using Rena (either my version or the original Rena) on any site that consumes RDF that you can't vouch for - because of the REXML security bug
19:14:56<DanC>I think swimming is popular enough that the olympics don't have as much impact. but I don't know for sure
19:15:15<DanC>almost every kid takes swimming lessons and gets some exposure to "swim team"
19:16:10<tommorris>If you are parsing any potentially unsafe RDF/XML in Ruby, you should probably run it through a validator first and spit it back in disgust if it fails.
19:16:15<kidehen>DanC: maybe just the "cool factor" challenge then :-) Couldn't believe the crowd in London waiting to greet Phelps.
19:17:03<bengee>tommorris, thanks. then I'll keep things wishy-washy "people are working on rdf for ruby, contributors welcome".
19:17:03<tommorris>There is a big exploit for the REXML library. Of course, the solution to this is simple: stop using REXML and start using libxml-ruby.
19:17:08<kidehen>DanC: Baltimore, understandable, but London, that's a whole new level
19:17:20<tommorris>bengee: Yes, patches welcome. ;)
19:17:37<kidehen>DanC: of course, they are the next hosts, but the crowd as simply unbelievable re. size
19:17:39<tommorris>(And bug reports.)
19:18:06<kidehen>DanC: anyway popping out, bye
21:14:43<Anchakor>Is there any activity refarding using rdf as format for communication?
21:14:54<Anchakor>ragarding*
21:17:05<Anchakor>meaning both asynchronous (mail) and synchronous (IM, group chat)
21:42:39<bblfish>Anchakor: that sounds suspiciously close to rss :-) The atom protocol is very close to mail if you think of it in a certain way
21:44:27<Anchakor>bblfish: well I was more thinking of its usage in IM protocol field
21:45:53<Anchakor>various clients supporting various features, depending what rdf vocabularies the messages contain
21:46:24<bblfish>Ask danbri, or check his blog. I think he has been playing with adding SPARQL queries to the well known xml based p2p protocol (forgot the name)
21:48:20<tommorris>xmpp
21:48:31<bblfish>yes
21:55:16<Anchakor>I started to think about this, because a project I am part of needs something like IRC but when you login you see what was said before even in your absence, and friendly frontend supporting this, some people were calling for skype (ugh), I wanted to propose some open solution
21:56:08<Anchakor>I ruled out xmpp because of clusy group chat support, which is important for us
21:57:18<Anchakor>with irc this could be done by bot but that cannot be user friendly
21:58:31<Anchakor>anyway this does not really belong to this channel, but I just wanted to say that thinking about this I though how nice it would be to use rdf in some protocol - so beautifully modular :)
22:07:22<bblfish>anchakor, well your right to look at RDF for its modularity. Solves quite a lot of problems that don't need to be re-invented.
22:07:35<bblfish>ACTION going to sleep
22:09:01<Anchakor>night

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