Experimental IRC log swig-2009-05-24

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.

06:31:07<mhausenblas>good morning Web of Data
06:35:27<drrho>goodmorning, WODezzin! :-)
06:44:23<mhausenblas>drrho: you know, TMorrow never dies :D
06:44:39<mhausenblas>ACTION wondering if this was too subtle, now ;)
06:45:34<mhausenblas>tommorris: nice SPARQL lesson at OpenHack! and people had good questions as well ...
06:56:28<drrho>mhausenblas: :-]]
07:03:21<mhausenblas>ACTION thinks drrho got it ;)
07:13:57<drrho>ACTION working on a series of blog entries: TMs for RDFreaks ;-)
07:17:26<mhausenblas>ACTION is already and RDF_a_ddicted, sorry ;)
07:17:33<mhausenblas>s/and/an
07:18:44<drrho>Well, it is not meant for convincing. I am not a missionary.
07:19:03<mhausenblas>right ;)
07:19:22<mhausenblas>so, the other day I asked about news reader for Mac here (http://chatlogs.planetrdf.com/swig/2009-05-22.html#T10-02-00)
07:19:58<mhausenblas>and, to conclude: it's NetNewsWire I finally found most usable
07:36:30<besbes>mhausenblas: i can recommend NetNewsWire on Mac+iphone, they sync your read status via newsgator
07:36:39<besbes>Ah, good morning btw
07:47:26<mhausenblas>besbes: yeah, I installed NetNewsWire on the MAc, didn't know they have it for iPhone as well - very cool
09:00:37<mhausenblas>http://webofdata.wordpress.com/2009/05/24/technology-malbestpracticing/
09:01:22<mhausenblas>A:| Technology MalBestPracticing
09:01:47<mhausenblas>A: Thoughts on how RDF is abused and hence can discourage people using it
09:02:36<mhausenblas>A: Please add your own observations - I'd like to start to collect RDF anti-patterns to raise awareness
09:13:09<drewp>mhausenblas: i already disagree with the closed-world one :)
09:13:54<drewp>ACTION says, as he hacks on a site that uses sesame for storage but has no RDF connections with the outside world
09:14:47<mhausenblas>drewp: hey, it's a free Web, sure you can :) do comment on it, but I'm really convinced that this is a strong one ...
09:14:52<mhausenblas>I'm not saying that you can't use it
09:15:12<mhausenblas>of course you can, but what is the benefit over other solutions, there?
09:15:39<mhausenblas>believe me, I've seen so many projects (ok, rather research stuff, but ...) that basically said:
09:15:46<drewp>in my case, i -hope- to connect with other data sources someday. I have twitter accounts in here, for example
09:16:08<mhausenblas>ok. but then you have a point there
09:16:46<mhausenblas>and I'm happy updating it in the sense ... 'unless you plan for serendipity future use'
09:16:55<drewp>i guess you already said "and there is no need to share/incorporate other data", but that really limits the cases IME
09:17:14<drewp>as soon as your data involves "users", there's some big potential for sharing with other systems
09:17:28<mhausenblas>well, I was referring to current, actual usage, but didn't explicitly say future use as well
09:17:35<mhausenblas>drewp: true
09:17:51<mhausenblas>please add your thoughts to the blog, it's really very very helpful
09:18:07<drewp>other reasons to go ahead with rdf, even in a closed-world case: you get generic data browsers (e.g. tabulator) for free
09:18:14<mhausenblas>(planning to collect this and put it on a Wiki, maybe semanticweb.org or so)
09:18:41<drewp>queries with lots of joins are easier in sparql than sql, at least for me
09:19:18<mhausenblas>ok, fair enough (re queries)
09:20:45<drewp>to take the other side for a minute, an RDBMS plus D2R might be suitable for connecting with generic data browsers, even if you do your "real work" entirely on the SQL side
09:20:57<drewp>we've only just started to play with D2R but it's looking nice
09:22:16<mhausenblas>re generic data browser, yes, could be, but I bet people would be able to have counterexamples at hand
09:22:20<mhausenblas>ACTION BRB, off for some coffee ...
11:36:37<gsnedders>mhausenblas: It's hard to think of stuff as something that doesn't physically exist, though (re: "Thinking of RDF on the serialisation level.")
11:47:18<mhausenblas>gsnedders: hehe. I guess you're right. please bear with me, ain't no native speaker ;)
11:48:56<kwijibo>ACTION likes mhausenblas's post, writing reply
11:48:56<phenny>kwijibo: 21 May 23:19Z <kidehen> tell kwijibo voiD Data set is typically produced by the Triple / Quad Store instance
11:48:58<phenny>kwijibo: 21 May 23:20Z <kidehen> tell kwijibo voID data set takes the form of a graph e.g. http://lod.openlinksw.com/void/Dataset
11:49:15<kwijibo>thanks phenny
11:50:37<mhausenblas>ta kwijibo ;)
11:54:45<kwijibo>ACTION hits publish, goes to check, too late, for typos and formatting blunders
12:01:35<mhausenblas>yep, kwijibo, very nice post, http://blogs.talis.com/n2/archives/470
12:01:58<mhausenblas>shall I link from my post or are you gonna leave a comment to point to yours?
12:02:16<kwijibo>thanks mhausenblas :)
12:02:28<kwijibo>i am writing a comment on yours anyway
12:02:35<mhausenblas>ah, cool, ta ma
12:02:42<mhausenblas>s/ma/man
12:02:48<kwijibo>there were som other ones i thought of that didn't fit in the post I wrote
12:03:25<kwijibo>mhausenblas: can I do html lists in your comments dyaknow ?
12:03:57<mhausenblas>hm. dunno. lemme check ...
12:05:15<mhausenblas>I can. but this might be due to being admin ...
12:05:43<kwijibo>it's ok, i just uses asterisks instead
12:06:05<mhausenblas>yep. ok. I can edit it afterwards ...
12:09:43<mhausenblas>ping bengee
12:09:43<mhausenblas>.seen bengee
12:09:43<mhausenblas>logger?
12:09:44<phenny>Sorry, I haven't seen bengee around.
12:10:09<mhausenblas>ah, yeah, sure, it's always the other bot ;)
12:14:03<kwijibo>lol
12:14:08<kwijibo>ACTION commented
12:18:05<mhausenblas>... and approved ;)
12:18:34<mhausenblas>still, the list is not rendered correctly though I edited it. hm.
12:18:34<kwijibo>:)
12:19:12<kwijibo>fkcgin wordpress eh?
12:19:27<kwijibo>it always monkeys with your markup
12:20:17<mhausenblas>mhm, certainly true ;)
14:51:41<LeeF>ACTION replies to mhausenblas in violent disagreement, as he always does with these Linked Data Web Is the One And Only Way positions :-p
14:53:15<mhausenblas>ACTION thanks LeeF - the post was meant to provoke different opinions and I'm glad to read your (great!) reply
15:40:46<LeeF>ACTION :-)
15:59:02<mhausenblas>phenny, tell timbl that I've updated http://esw.w3.org/topic/WriteWebOfData now - please review/comment
15:59:02<phenny>mhausenblas: I'll pass that on when timbl is around.
17:32:03<mhausenblas>good nite Web of Data
18:16:29<idmclean>Greetings semanticians, What is new?
18:21:52<idmclean>Are there any ontologies out there which describe hypotheses?
19:24:21<LeeF>idmclean, you probably would be interested in the SWAN work that Tim Clark is doing
19:29:43<idmclean>Leef, thank you. I do believe SWAN is something I am looking for.
19:39:25<idmclean>LeeF, I was reading a textbook on Elementry Statistics yesterday, and I realized while reading hypothesis testing that using RDF, we could automate a large amount of hypothesis formation and testing using existing data sources. It would be something like folding@home but more generally applicable.
20:11:46<Anchakor>idmclean: hi, I am working on the example, progressing quite well, but now I got idea - do you think it is possible to extract function definition from programs themselves? basically function is just a piece of code using some built-in functions with few undecided variables = arguments...
20:12:40<idmclean>Anchakor, you mean reverse engineering from an executable?
20:12:53<Anchakor>idmclean: no from the semantics of the code
20:13:54<idmclean>Anchakor, extract function definitions from the abstract syntax tree, parse tree, or the abstract semantic graph?
20:14:23<Anchakor>idmclean: no, I am not talking about existing systems...
20:14:45<Anchakor>but maybe abstract semantic graphs is what I mean, I dunno what it really is
20:15:38<idmclean>Anchakor, I need you to elaborate on some more. If I am comprehending your question, I know you could do it in Java. java.lang.reflect is just for that purpose.
20:18:04<Anchakor>I mean like instead of defining a function, you write a code, a part of it you tag as some reusable complex, leaving some variables to fill in as arguments, defining a function in a different way...
20:54:14<idmclean>Anchakor, reusable complex?
20:55:38<Anchakor>idmclean: sorry I meant "concept" :)
20:57:24<idmclean>Anchakor, I grok your meaning. In Java, you can grab a method, you can grab the method's variables, you can grab the method's parameters, but I don't know if you can grab the block that represents the body of the method.
20:58:47<idmclean>Anchakor, you can also add annotations and serialize the methods as objects. I have no idea how to cleanly grab the method/function definition, but I imagine it is probably possible.
21:01:35<idmclean>I would research how the Java developers managed to do that. I suspect it would unveil a method of doing what you describe.
21:03:32<Anchakor>I think the method wouldn't be to hard to figure out, I am more concerned if it really would prove as useful for something...
21:05:34<idmclean>Anchakor, it would be useful for automating software design, development, and testing, so I would find it very useful. It means I could describe what I want and let the computer figure out the details.
21:08:48<Anchakor>yeah, I imagine you could host your open source code on your website as linked data
21:09:22<Anchakor>that alone makes it useful
21:09:35<idmclean>Yup
21:12:39<idmclean>Also, we could run source code through analyzers to perform algorithm analysis, breed new algorithms, and perform hypothesis testing on different iterations and variations of the source code.
21:43:25<idmclean>Anchakor, people could call your functions remotely via URI namespaces.
21:56:10<idmclean>No need to download libraries locally if you can pick off the pieces you need remotely as you need them.
23:04:20<Anchakor>idmclean: yeah exactly
23:05:08<Anchakor>I guess caches will become more important then
23:14:58<idmclean>Anchakor, It depends on how you use that capability. If I were to write a program that was compiled into an executable, nothing would change about the executable. At the development level, I would simply download the relevant code from the libraries online as part of compiling my source code.

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