Friend of a Friend (FOAF) IRC Chat Logs for 2005-04-07

12:54:20 <Esaj_> Esaj_ is now known as Esaj

15:10:26 <Cloud_> Cloud_ is now known as Cloud

15:32:55 <balbinus6> balbinus6 is now known as balbinus

18:07:30 <JimJibber> JimJibber is now known as JibberJim

18:13:02 <crschmidt> libby: oops, apologizes for misreading DOAP bulliten board

18:13:50 <libby> s'ok, we just get the final decision is all

18:14:09 <libby> probably worth creating another more general one on the foaf wiki

18:14:17 <ldodds> url?

18:14:43 <crschmidt>http://esw.w3.org/topic/SemanticWebDOAPBulletinBoard

18:15:16 <ldodds> oh ta, the esw wiki changes somehow slipped out of my rss reader.

18:15:25 <libby> we're teying to simplify the waywie collect apps and demos together in swbp wg

18:15:44 <libby> so we want dioap files for semweb apps and demos specifically

18:15:48 <libby> doap

18:15:58 <libby> a more general page woudl be cool

18:17:31 <ldodds> what do you class as "RDF, RDFS and OWL applications"?

18:17:35 <ldodds> are editors, query tools included?

18:18:45 <libby> yep :)

18:19:02 <ldodds> ok, cool

18:19:04 <libby> very much so

18:19:50 <ldodds> is there an RDF version of that bulletin board available somewhere?

18:20:16 <ldodds> IIRC my XSLT hack doesn't work on MoinMoin (html too broken)

18:20:21 <crschmidt> yeah

18:20:28 <crschmidt> jim's got stuff to do it for the FOAF bulliten board

18:25:20 <crschmidt> bleh, esw's HTML output is more friendly than FOAF

18:25:51 <libby> not currently anything, should be....

18:26:50 <crschmidt> ^ns rdf

18:26:51 <julie> The URL for the rdf namespace is http://www.w3.org/1999/02/22-rdf-syntax-ns#

18:27:03 <crschmidt> ^ns rdfs

18:27:03 <julie> The URL for the rdfs namespace is http://www.w3.org/2000/01/rdf-schema#

18:29:31 <crschmidt> one liner for generating from FOAF or DOAP bulliten board on FOAF wiki:

18:32:25 <crschmid1> crschmid1 is now known as crschmidt

18:32:30 <crschmidt> echo '<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"><rdf:Description>' > temp;w3m -dump_source http://rdfweb.org/topic/FOAFBulletinBoard | grep external | perl -pe 's/.*?external" href="(.*?)"/\1\n/g' | grep http |perl -pe 's/^(.*?)$/<rdfs:seeAlso rdf:resource="\1" \/>/' >> temp; echo "</rdf:Description></rdf:RDF>" >> temp

18:33:24 <crschmidt> oh, that uses w3m: can equally use wget -O -

18:34:58 * bengee reads the adtf inclusion criteria

18:38:35 <libby> it's still vague bengee

18:39:05 <libby> basically you're fine if it's free/open source in some version and it uses/creates rdf or owl

18:39:20 <libby> note that these are not finalized yet. i should put a inote to that effect

18:40:09 * libby hurries to do that

18:40:22 <bengee> I was wondering if being an ADTF project could be auto-detected or something. The semanticweb.org site could be used to create/host adtf doap descriptions...

18:41:44 <bengee> I could prolly offer you to create a custom adtf-a-matic template there, too. would be happy to get some stuff in the store..

18:43:01 <libby> interesting bengee :)

18:43:17 <libby> all this needs more thought really ...we need to discuss it in our taskforce

18:47:16 <bengee> should be possible to query for adtf demos when the criteria were RDFy (e.g. "all doap:Projects with a doap:license of foo, and a download-page")

18:48:34 <libby> the aim is that we just pick the ones we think are suitable for the task force's collection, and then others can put them in a facetted browser or whatever

18:48:46 <libby> a subgoal being to persuade people to create doap files :)

18:50:41 <libby> hm, the booklet on http://esw.w3.org/topic/SemanticWebBestPracticesTaskForceOnApplicationsAndDemos could be hacked to put things in the wiki of course

18:51:13 * libby delighted you can put almost any old js crap in bookmarklets - didn;t realize how cool they were. like loops

18:52:12 <bengee> bookmarkcraps ;)

18:52:20 <JibberJim> bookcrapmarks!

18:52:41 <crschmidt> hey jim, how are you doing the FOAFBullitenboard thing?

18:52:54 <JibberJim> er, some regexps like you

18:53:14 <libby> and in fact, better still, we hack chris's firefox extenssions so a click on the button puts it in teh wiki for you

18:53:32 <crschmidt> hm, can you do that?

18:53:37 <crschmidt> Can javascript POST things?

18:53:41 <danbri> Here's an extension idea...

18:53:49 * libby goes to put together new mac mini

18:53:59 <libby> I hink you can more or less do what you want with extensions

18:54:18 <JibberJim> yes javascript can post things

18:54:19 <danbri> dredger... scoop up rss feeds and foaf files directly as you navigate, without asking the user about saving them

18:54:45 <danbri> can extensions communicate? it could feed into Sage, Piggybank, Libby's and Chris's stuff...

18:55:08 <crschmidt> Well, I personallly think piggybank would be much more useful if it could do that anyway

18:55:22 * crschmidt kind of dreads heading back into extensionland, javascript is scary

18:55:41 <kasei> heh

18:56:20 <danbri> do you know if the moz .js debugger makes much difference w.r.t. speed of writing extensions?

18:56:48 <JibberJim> speed of writing?

18:59:37 <danbri> pain of development, i mean

18:59:41 * danbri finds http://kb.mozillazine.org/Dev_:_Extensions

19:00:51 <crschmidt> danbri: well, not much to compare it against, is there? only firefox has extension, and it also always has the debugger...

19:02:03 <crschmidt> whoa, http://kb.mozillazine.org/Dev_:_Extensions is useful

19:06:11 <crschmidt> Didn't know it was possible to not have to restart firefox when testing extensions

19:06:35 <JibberJim> well if you like a debugger, venkman is pretty good.

19:06:37 <danbri> which bit explains that?

19:06:48 <danbri> libby was restarting it 3 times per edit, swearing as she went...

19:07:53 <crschmidt> "nglayout.debug.disable_xul_cache = true. Disables the XUL cache so that changes do not require a restart (more)."

19:08:23 <crschmidt> (you have to do some dicking with it first, setting the locations of files manually, can't just install a new one while it's running, but you can edit the files it uses)

19:13:03 * bengee wonders if there are any common FunctionalProperties other than foaf:primaryTopic..

19:13:26 <crschmidt> ^q select ?t where (?t rdf:type owl:FunctionalProperty)

19:13:27 <julie> irc:software, http://www.purl.org/net/ontology/beer.owl#hasAlcoholicContent, http://www.purl.org/net/ontology/beer.owl#hasOriginalWortContent, irc:topic, cyc2004:guid, irc:hostname, foaf:gender, foaf:primaryTopic, irc:username

19:13:43 <crschmidt> ... foaf:gender is a functionalproperty?

19:14:03 * bengee trying to avoid to smush on datatype props

19:15:07 <bengee> the beer ones are booleans, I guess?

19:16:46 <bengee> primaryTopic should do then for a first test..

19:17:10 <sh1m> sh1m is now known as sh1mmer

19:22:33 <crschmidt> I'm not sure what a FunctionalProperty is supposed to be, I guess

19:23:08 <crschmidt> Should be "A Subject has one and only one value for any property which is a FunctionalProperty", I guess/

19:25:21 <bengee> does julie have an "un-smusher" or are the subjects/objects overwritten?

19:25:58 <crschmidt> Overwritten

19:28:56 <danbri> dateOfBirth is functional

19:28:59 <bengee> so if you wantd to clean up the store, would you remove single triples then, or remove all triples from a given source, or is there any other common practice?

19:29:01 <danbri> for eg

19:29:11 <danbri> or the idea is anyway

19:29:16 <danbri> things that have at most one true value

19:29:33 * crschmidt nods.

19:29:45 <bengee> if dob were functional, why did I give up on my three-prop proposal then? ;)

19:30:00 <crschmidt> bengee: I haven't done anything with it yet. I've turned it over in my head a lot

19:30:07 <crschmidt> the only thing I've ever done so far is remove single triples

19:30:10 <bengee> I see

19:30:19 <crschmidt> But I'm thinking about something that would remove all triples into or out of a given subject

19:30:30 <crschmidt> ^q select ?p where (?p foaf:nick "danbri")

19:30:31 <julie> http://www.cs.umd.edu/~golbeck/daml/golbeckFOAF.rdf#danbri, http://kasei.us/about/foaf.xrdf#danbri

19:30:42 <crschmidt> ^q select ?p where (?p foaf:name "Dan Brickley")

19:30:45 <julie> (r1095214475r19), http://www.ideaspace.net/users/wkearney/foaf.xrdf#danbri, http://kasei.us/about/foaf.xrdf#danbri, (r1095208694r13)

19:30:59 <crschmidt> So I could so, for example, "delete http://www.ideaspace.net/users/wkearney/foaf.xrdf#danbri"

19:31:25 <crschmidt> and delete the fact that wkearney knows danbri, and all the things attached to that copy of danbri

19:31:38 * crschmidt wonders why wkearney didn't include a sha1sum for danbri in that file

19:33:04 * bengee wonders what's wrong with that rdf..

19:34:05 <crschmidt> which?

19:34:18 <bengee> that wkearney one

19:34:53 <crschmidt> nothing in particular, although it makes my path finder fail, because there are two danbris: one that only wkearney knows, and one that everyone else knows (which is the kasei.us uri)

19:35:18 <bengee> ah, 'k

19:35:53 <crschmidt> well, it would succeed eventually, but i ran it for 30 minutes and gave up, because it had to search all the paths from libby, which number in the thousands

19:35:54 <bengee> hm, yeah, removing unsmushable resource descriptions is a use case I haven't considered yet..

19:36:11 <crschmidt> the only problem with wkearney's is that it won't *ever* be smushable

19:36:20 <crschmidt> unless I add a triple manually... why didn't I think of that before?

19:36:40 <bengee> esp. as he assigned a URI to danbri ;)

19:37:36 <crschmidt> ^addturtle <http://www.ideaspace.net/users/wkearney/foaf.xrdf#danbri> foaf:mbox <mailto:danbri@w3.org>.

19:37:37 <julie> Model size increased by 1 to 2123720 via turtle statements.

19:37:40 <crschmidt> There we go

19:37:45 <crschmidt> now i can smush a danbri away! hooray!

19:37:51 <kasei> hahaha

19:37:56 <kasei> crschmidt++

19:38:18 * crschmidt is kind of evil, but that's okay

19:38:50 <crschmidt> ^q select ?p where (?p foaf:name "Dan Brickley")

19:38:50 <julie> (r1095214475r19), http://kasei.us/about/foaf.xrdf#danbri, (r1095208694r13)

19:38:56 <crschmidt> woot

19:39:08 <crschmidt> now i've gotta do it for those blank nodes too. that's more difficult

19:40:53 <crschmidt> kasei: hm, one of those is your fault

19:41:04 <crschmidt> [http://kasei.us/about/foaf.xrdf#simonprice], [http://xmlns.com/foaf/0.1/knows], (r1095214475r19)

19:41:44 <crschmidt> oh, hmph.

19:42:27 <crschmidt> old data, i think

19:42:28 <kasei> hrm... i'll look into cleaning my end up (including getting rid of the uri)

19:43:01 <crschmidt> I don't seem to see a simonprice knows danbri nodei n that file anymore

19:43:09 <kasei> neither do I

19:43:15 <crschmidt> oh, silly me

19:43:19 <crschmidt> it doesn't *have* to be in that file

19:43:31 <crschmidt> but simonprice-the-bnode became simonprice-the-kasei-uri

19:44:07 <crschmidt> yep, looks like that's what happened, and at some point, simon price had danbri listed with an mbox as a literal, rather than a resource

19:44:16 <crschmidt> (r1095214475r19), [http://xmlns.com/foaf/0.1/mbox], "mailto:danbri@w3.org" <-

19:44:22 <crschmidt> which, clearly, won't smush

19:46:10 <crschmidt> ^q select ?p where (?p foaf:name "Dan Brickley")

19:46:10 <julie> (r1095214475r19), (r1095208694r13)

19:46:32 <crschmidt> (slowest way ever of cleaning up a megatriple database)

19:47:09 <kasei> i don't know about that. you could be doing it without julie :)

19:47:19 <crschmidt> heh

19:47:24 <crschmidt> except i have nice query functions in my wrapper

19:47:32 <crschmidt> so, it wouldn't make much difference :)

19:47:57 <crschmidt> The primary reason for my wrapper is to make it so I can just m = get_model(), and not have to look up the Storage() init params, which I always forget

19:48:17 <kasei> me too.

19:48:33 <crschmidt> ^available http://www.digitalskunk.com/blog/

19:48:34 <julie> Query returned no results

19:48:47 <kasei> steve davie

19:49:05 <crschmidt> ^q select ?r where (?p foaf:homepage <http://www.digitalskunk.com/blog/>) (?p rdfs:seeAlso ?r)

19:49:14 <julie> Query returned no results

19:49:26 <kasei> ^name obigabu

19:49:27 <julie> Steve Davie

19:49:33 <crschmidt> hm.

19:49:54 <crschmidt> er. looking at the wrong person.

19:49:57 <kasei> that should totally be connected.

19:49:57 * crschmidt needs greg, not steve :)

19:50:01 <kasei> although there's no seeAlso

19:50:16 <crschmidt> oh, greg is you

19:50:40 <crschmidt> weird, the second bnode is pretty uninformative: must have picked it up somewhere and not spidered the seealsos or something

19:51:35 <crschmidt> ^q select ?r where (?p foaf:homepage <http://www.digitalskunk.com/blog/>) (?p rdfs:seeAlso ?r)

19:51:35 <julie> Query returned no results

19:51:44 <crschmidt> ^q select ?p where (?p foaf:name "Dan Brickley")

19:51:44 <julie> (r1095214475r19)

19:51:46 <crschmidt> there we go!

19:51:59 <kasei> hmm?

19:52:07 <crschmidt> only one danbri now :)

19:52:21 <kasei> ooh. right. nice.

19:52:27 <bengee> yay! on to the next name ;)

19:52:41 <kasei> only a few hundred thousand left.

19:53:18 <crschmidt> hey, there's not even 200,000 people!

19:53:23 * crschmidt gestures to his blog

19:53:42 <kasei> really?

19:53:43 <crschmidt> Total foaf:Persons: 129,932

19:54:15 <crschmidt> man, my recent flickr pictures are odd

19:54:38 <kasei> i guess my recent spidering got a bit out of hand.

19:54:58 <crschmidt> 1. local ice cream shop, 2. E&R cleaners truck, for some random reason, 3. deep fried green beans from lunch today, 4. jess at lunch, wearing the "there's no place like 127.0.0.1" shirt

19:55:20 <crschmidt> kasei: it's really easy to head overboard once you hit LJ

19:55:36 <crschmidt> There's a very good reason I limit my spider to only go out two levels

19:55:38 <kasei> yes it is

19:56:01 <kasei> i like having the firehose of data to play with, though :)

19:59:10 * bengee gives his FP smusher a try..

19:59:34 <crschmidt> bengee: so, are you just tossing older triples or something?

19:59:43 <bengee> hm, "merging 0 resources"

20:00:01 <bengee> I just make sure that nothing is smushed. no need for an undo, huh?

20:00:22 <crschmidt> I mean, you can't take two subjects and say they're the same... oh, I suppose you can take two objects and say they're the same

20:00:34 <crschmidt> hm - shouldn't foaf:based_near be an FP then?

20:00:56 <crschmidt> You can only be based_near one location at a time...

20:01:07 <kasei> no!

20:01:12 <bengee> nearestAirport, perhaps

20:02:04 <kasei> you can be near lots of things at the same time

20:02:16 <kasei> I would assume the spec says as much, although I'm not sure.

20:04:52 * bengee manually creates some primaryTopics for a resource..

20:08:55 * bengee re-smushes on a document describing two no ifp sharing libby bnodes..

20:09:34 <bengee> aha, "merging two resources"..

20:10:05 <crschmidt> I think foaf:knows should be an FP

20:10:06 <crschmidt> :)

20:10:13 <crschmidt> "I can only know one person, damnit!"

20:10:27 <crschmidt> (would make my model a lot smaller... paths easier to calculate...)

20:10:42 <JibberJim> as long as it's wn:beer

20:10:45 <bengee> guess that makes you a "poor lonesome OWLboy"..

20:11:38 <kasei> it would be a very tall, narrow graph

20:12:02 <JibberJim> hmm, very wide, very short graph here :-)

20:12:21 <kasei> heh

20:25:23 * bengee can't believe that the smusher works. not sure if it'll scale, though..

20:25:41 <crschmidt> pfft. scaling.

20:26:34 <bengee> hehe

20:27:51 <crschmidt> howdy zwnj/benham

20:31:49 <Mutiny> smusher?

20:32:02 <Mutiny> sounds interesting. :)

20:32:10 <crschmidt> .g rdf_smush

20:32:12 <zwnj> crschmidt: hi

20:32:22 <phenny> rdf_smush: sorry, no results were found.

20:32:26 <crschmidt> hm

20:32:52 <bengee> .g smushing

20:33:05 <phenny> smushing: http://rdfweb.org/topic/Smushing

20:33:51 <crschmidt> .g rdf-smush redland

20:33:55 <crschmidt> was what i was pointing towards

20:34:03 <bengee> ah

20:34:08 <phenny> rdf-smush redland: http://crschmidt.net/svn/rdfpython/trunk/rdf-smush.txt

20:34:33 <Mutiny> ahhhhh!

20:34:47 <Mutiny> i just had to write some code to do the same thing, interesting it's got a name. ;)

20:35:16 * crschmidt keeps meaning to write one in Python, just so he can tweak it, since he doesn't understand C

20:35:29 <crschmidt> all the hard work is in Redland stuff anyway, so I don't expect it'd be much slower

20:35:36 <Mutiny> just using InverseFunctionalProperties?

20:36:00 <crschmidt> yep

20:38:16 * danbri wonders if dave would be interested in bundling rdf-smush as a standard redland utility, so it could have script wrappers built w/ the other redland stuff

20:44:06 * libby watches 'hackers' with some hilarity

20:44:28 <crschmidt> hackers is wonderful :)

20:44:36 <libby> "use your best viruses to buy us some time!!!"

20:44:38 <crschmidt> (hilariously inaccurate, but still wonderful)

20:45:56 * danbri wires a spybot virus to libby's irc cyberinbox

20:46:32 <libby> damn you!!!!

20:49:34 <Mutiny> hack the planet! haaaack the plaaaaneeeeeet!

20:50:01 <libby> hacking makes lightning bounce from one machine to another - cool!

20:50:06 <JibberJim> hmm - http://rdf4food.asemantics.org/AccomoDation the "plan" link is a 404

20:50:20 <libby> yeah...I can't fix atm

20:51:22 <libby> heheh. just had the "hack the plaaaaneeettt!" moment

20:51:34 <crschmidt> They're Trashing our rights! Trashing! trashing!

20:52:38 <mattmcc> Heh. They ran that on one of the cable networks on Sunday.

20:52:46 <Mutiny> that was like angelina jolie's first movie, too.

20:52:56 <libby> they all look so young...

20:53:44 <Mutiny> this WAS 95 after all.

20:53:59 <crschmidt> hm, never got kasei's email on the list

20:54:14 <libby> dunno what http://imdb.com/name/nm0000498/ is doing with those plaits

20:54:22 <Mutiny> it'll be happy day when matthew lillard gets hit by a bus.

20:55:07 <libby> he was ok in scraem

20:55:35 <crschmidt> i liked cereal killer!

20:56:05 <Mutiny> he's definately not a versatile actor at all.

20:56:41 <mattmcc> What always seemed weird to me was Jonny Miller going from that to Trainspotting.

20:56:51 <Mutiny> i like the #1 thing under imdb's list for 'goofs' on Hackers.

20:56:51 <Mutiny> Goofs for

20:56:52 <Mutiny> Hackers (1995)

20:56:52 <Mutiny> * Factual errors: Generally ill-informed and ridiculous to the extreme regarding the capabilities and characteristics of computers, technology, and phone systems.

20:57:01 <libby> +1 mattmcc

20:57:07 <libby> zac fixed http://www.asemantics.com/n/images/planz.png

20:59:06 <crschmidt> "The game being played in the arcade is a high-quality prototype of the Playstation game "Wipeout" by Psygnosis. It is done on a high-end SGI server and allowed the development team to try out tracks and gameplay, before porting it to the Playstation. As a result, there are features and graphics in the movie that do not exist in the actual game, including the "high score smashing" sequence."

20:59:12 <crschmidt> Didn't know that

21:03:50 * dajobe returns from watching hackers, what a laugh

21:04:11 <crschmidt> have you people not seen it before?

21:05:28 <dajobe> I've forgotten it already

21:05:46 <mattmcc> Yeah, it had that effect on a lot of people.

21:05:58 <crschmidt> heh, heh

21:07:34 <zwnj> crschmidt: seems foaf-current.png doesn't work online

21:09:33 <crschmidt> zwnj: foaf-current.png isn't current

21:09:40 <crschmidt> hasn't been for quite a while

21:11:57 <dajobe> you can parse svg easier than in http://crschmidt.net/blog/archives/51/parsing-svg-metadata/

21:12:26 <dajobe> set parser feature scanForRDF to true

21:12:41 <crschmidt> ah

21:12:51 <crschmidt> I don't know how to set a parser feature

21:13:03 <crschmidt> is it just a string?

21:13:16 <dajobe> key: uri string, value string

21:13:30 <crschmidt> yeah, but I don't know what the uri I'm supposed to use is

21:13:36 <dajobe> true

21:13:49 <dajobe> I'm looking to see if I'd documented it, and no I didn't

21:14:48 <dajobe> set feature key: uri http://feature.librdf.org/raptor-scanForRDF value: 1

21:15:30 <crschmidt> asesome

21:15:32 <crschmidt> er

21:15:33 <crschmidt> awesome

21:15:55 <dajobe> if you read rapper -f help you can see what the others are

21:17:13 <crschmidt> does that work in non XML docs too?

21:17:24 <dajobe> no

21:17:32 <dajobe> I was thinking about adding a parser for that

21:18:04 <dajobe> to scrape things like CC metadata out of <!-- <rdf:RDF ... > -->

21:23:11 * crschmidt updates post to reflect the fact that Dave was way ahead of him

21:23:30 <dajobe> can't blame you for not knowing an undocumented feature

21:24:11 <crschmidt> well, now it's slightly more documented :)

21:27:50 * bengee looks for a better word for "un-smush"..

21:28:38 <crschmidt> dajobe: while we're talking documentation: q2 = RDF.Query("- - -", name="triples") is in the pydoc, but the "name" param doesn't seem to be in in the init function: is that supposed to be query_language, or is it a deprecated example?

21:29:37 <dajobe> ah, wrong example

21:29:50 <dajobe> it's query_language now

21:30:00 <crschmidt> is there a "triples" query_language anymore?

21:30:20 <dajobe> inside redland yes, but it's pretty useless and was just for testing things

21:30:52 <crschmidt> hm, q2 = RDF.Query("- - -", query_language="triples"), q2.execute(m) doesn't seem to return any results

21:31:36 * dajobe shrugs, must have rotted

21:31:43 * crschmidt nods

21:55:32 * bengee has a de-smusher :)

22:00:18 <kasei> bengee: what does that do?

22:02:36 <kasei> oh. he left.

22:12:50 <zwnj> is there any {orkut,360,etc}-to-foaf library around?

22:13:13 <danbri> orkut ppl were playing with one

22:13:20 <danbri> and there are various orkut scrapers/crawlers

22:13:31 <danbri> but nothing obvious to use right now for that purpose

22:13:46 <danbri> the foafnet list has some examples of what orkut were looking to expose

22:13:55 <danbri> i think they have their hands full currently re scalability

22:41:26 <zwnj> danbri: thanks

22:41:39 <zwnj> what about clients for blog/foaf/etc?

22:42:09 <zwnj> i mean a real application that also works offline