Semantic Web Interest Group IRC Chat Logs for 2003-04-18

01:37:18 <DanC>http://xmlsoft.org/xmlreader.html

01:37:18 <dc_rdfig> A: http://xmlsoft.org/xmlreader.html from DanC

01:37:28 <DanC> A:|Libxml2 XmlTextReader Interface tutorial

01:37:28 <dc_rdfig> Titled item A.

01:37:50 <DanC> A:by Daniel Veillard

01:37:50 <dc_rdfig> Added comment A1.

01:37:52 <DanC> A:v 1.6 2003/04/17

01:37:52 <dc_rdfig> Added comment A2.

01:38:00 <DanC> A:seems there's life after SAX after all.

01:38:00 <dc_rdfig> Added comment A3.

01:46:59 <DanC> hoot! [[

01:47:01 <DanC> Computer game for real geeks: Russian roulette for your Unix box.

01:47:01 <DanC> [ $[$RANDOM % 6] = 0 ] && rm -rf /

01:47:07 <DanC> ]] -- http://www.advogato.org/person/ask/

03:57:48 <lilo> [Global Notice] Hi all. Two fairly low main rotation servers are about to be rehubbed. Total user count is about 1900 users. Apologies for the inconvenience, and we'll send a notice to each server. Any further info should be on WALLOPS per http://freenode.net/news.shtml .... thanks.

03:58:27 <mea_culpa> mea_culpa is now known as Radical_Edward

04:22:06 <lilo> [Global Notice] Hi all. Two more events and they'll both be very large. One will be a full server whose users will be dumped completely; the other will be a major hub dropping its connections, no users lost but quite a few split temporarily. We'll try to leave about five minutes between events.

04:35:32 <lilo> [Global Notice] Okay, folks, last one of the night. No users will be dumped, other than possibly from flood; it's just going to be a big netsplit. Apologies in advance, and please bear with us through this last event. Thanks.

04:37:58 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

09:39:00 <danbri>http://www.scandemonium.com/photodata.htm

09:39:00 <dc_rdfig> B: http://www.scandemonium.com/photodata.htm from danbri

09:39:09 <danbri> B:|PhotoData

09:39:10 <dc_rdfig> Titled item B.

09:39:17 <danbri> B:"PhotoData is a system of permanently associating information with an image, and for making that information universally accessible."

09:39:18 <dc_rdfig> Added comment B1.

13:40:13 <danbri> libby, http://www.w3.org/2001/sw/WebOnt/webont-mem-pub.rdf

13:44:09 <danbri>http://www.w3.org/2002/01/tr-automation/

13:44:10 <dc_rdfig> C: http://www.w3.org/2002/01/tr-automation/ from danbri

13:44:17 <danbri> C:|Automating the publication of Technical Reports

13:44:17 <dc_rdfig> Titled item C.

13:46:16 <danbri> C:includes editor markup, eg: <editor rdf:parseType="Resource"><contact:fullName>Brian McBride</contact:fullName></editor>

13:46:16 <dc_rdfig> Added comment C1.

13:46:51 <danbri> C:No email addresses or homepages for editors though, data merge w/ other sources might be fiddly.

13:46:51 <dc_rdfig> Added comment C2.

13:47:07 <danbri> C:See [http://www.w3.org/2002/01/tr-automation/tr.rdf|generated RDF]

13:47:07 <dc_rdfig> Added comment C3.

13:48:17 <danbri> C:Nearby: (via WG homepage, memberlist section), [http://www.w3.org/2001/sw/WebOnt/webont-mem-pub.rdf|WebOnt WG membership in RDF].

13:48:17 <dc_rdfig> Added comment C4.

13:50:02 <zool> quick question; i'm graphing something in an owl:Class Building ... in plain RDF that would be a rdf:type space:Building ... i'm not sure how to represent that (on the graph) correctly in OWL

13:50:16 <danbri> should be the same

13:51:00 <zool> just rdf:type ? not owl:instanceOf or anything like that?

13:51:09 * danbri thinks of OWL as OWL Full; I don't remember quite what restrictions OWL DL etc might place...

13:51:12 * danbri hopes not

13:51:17 <zool> that is what i mean, what is the predicate, sorry that is not clear

13:51:18 <zool> okay

13:51:27 <danbri> rdf:type

13:51:42 * zool has a lightbulb moment

13:51:54 * Morbus gets his peashooter ready.

14:05:52 <danbri> danbri has changed the topic to: Semantic Web wish factory, http://rdfig.xmlhack.com/ 'When you go meta, don't forget to come back'

14:06:25 <danbri> (topic from http://www.gca.org/attend/2000_conferences/Extreme_2000/friday.htm via http://esw.w3.org/topic/SeedApplications)

14:31:58 <DanC> danbri, SeedApplications is a draft of my action from the cal/geo meeting on Weds

14:32:52 * danbri merges two mental nodes

14:42:39 * DanC noodles on revision of SeedApplications:

14:43:23 <DanC> 1. SeedApplications should be connected to value-networks of real-world data. e.g. web server logs, RSS feeds, email, quicken records, geourl. Real life data.

14:44:08 <DanC> 2. developers of SeedApplications should endeavor to document DesignPatterns as they work

14:44:46 <DanC> 3. developers of SeedApplications should look for opportunities for sharing and economies of scale in the VocabularyMarket.

14:45:09 <DanC> whaddya think?

14:46:54 <ilowe> What is a SeedApplication? Googling comes up blank...

14:46:59 <danbri> I like the seed metaphor, and the general approach

14:47:09 <DanC> cf http://esw.w3.org/topic/SeedApplications

14:47:20 <ilowe> Thanks

14:47:34 <DanC> note that it's a wiki. i.e. we're deciding, together, what a SeedApplication is

14:47:40 <danbri> it feels quite inclusive, something RSS-DEV or foaf or DC guys could buy into, without necc feeling very much part of W3C's SW/SWAD/RDFIG...

14:48:24 <DanC> point 2 is borderline exclusive, if we take it to mean 'document in this wiki the design patterns...'

14:49:32 <DanC> well, not 'exclusive' but not completely minimally constraining

14:50:01 * DanC is still trying to get a handle on who are the DC guys

14:50:21 <danbri> I switched the foaf/rdfweb/mumble wiki to use Moin too, eg http://rdfweb.org/topic/ScutterSpec

14:51:12 <danbri> I think of DC as being in an awkward migration period from data format workshop series to being more a general metadata conference and set of mailing lists.

14:51:33 <danbri> I don't see much more work on actual vocab coming out of DC. Personal view.

14:52:21 <danbri> ...its good as a meeting ground for practitioners, very good; lots of enthusiasm. But the DC-arch WG has been chairless for nearly two years, suggesting that tech work isn't a priority.

14:52:45 <DanC> but I'd like to close the loop in some ways; e.g. if DC is to be used with OWL, folks must choose either { dc:creator rdfs:range rdfs:Literal } or { dc:creator rdfs:range cyc:Agent}.

14:53:14 <DanC> with OWL DL, that is

14:53:35 <danbri> yes... I feel bad every time I write <dc:creator><foaf:Person><... etc knowing that others are writing <dc:creator>Name Goes Here</dc:creator>

14:53:54 <danbri> but the former is so much more useful... (to me)

14:54:12 <danbri> and... not using DC seems, i dunno, rude.

14:54:20 <DanC> I'd like to know what the value network around dc:creator looks like. What do the Adobe tools think the range of dc:creator is? what does chefmoz think? dmoz?

14:54:34 <danbri> good question... dunno...

14:55:05 <DanC> "2003-09-28: DC-2003

14:55:05 <DanC> Dublin Core 2003 conference in Seattle, Washington." -- http://www.daml.org/events

14:55:37 <DanC> I guess whoever the DC folks are, they'll be there then. (except for exceptions)

14:55:55 <DanC> i.e. that's a target for closing the loop

14:56:27 <DanC>http://dc2003.ischool.washington.edu/

14:56:27 <dc_rdfig> D: http://dc2003.ischool.washington.edu/ from DanC

14:56:36 <DanC> D:|Dublin Core 2003

14:56:36 <dc_rdfig> Titled item D.

14:56:56 <DanC> D:argh! where/when info on homepage is in a gif. get WaiHappy

14:56:56 <dc_rdfig> Added comment D1.

14:57:19 <DanC> D:28 Sep - 2 Oct 2003 in Seattle, Washington

14:57:19 <dc_rdfig> Added comment D2.

14:58:48 <libby> danc, shellac suggesed that a set of conventions for meetings and conferences in the chump might be an idea - reckon it might be a goer?

14:59:06 <libby> e.g. start date on the first line, enddate on the second or something...

14:59:40 <libby> not worked out detail

15:00:14 <DanC> D:close the loop on dc:creator and [http://esw.w3.org/topic/ThingsVersusTheirNames|ThingsVersusTheirNames]?

15:00:15 <dc_rdfig> Added comment D3.

15:00:39 <DanC> yes, just give me a tool that rewards me for using the conventions and I'll use them.

15:00:50 <DanC> I already chump events relevant to my life regularly.

15:01:25 <DanC> (reward can be as small as showing me the data in RDF/XML or a confirmation by an IRC bot or some such)

15:02:53 * DanC is having a bad net day

15:05:09 * DanC noodles on backlinks from the wiki to the chump/weblog/scratchpad... wonders if google's backlink service suffices... wonders what the latency is

15:05:30 * DanC is having a very, very bad net day

15:06:07 * DanC neglected to attend RDF core nor send regrets

15:07:30 * DanC can't get the agenda in hypertext-UI time

15:10:31 <danbri> no meeting today - uk national holiday

15:10:47 <DanC> whew. thx.

15:11:40 * danbri runs ayftest.rb (scutter script), wonders how many RDF docs to expect to find...

15:11:50 <danbri> libby? anyone else? what's a normal find?

15:12:23 <danbri> INIT: s.left.size=817 s.seen.size=98 current: http://www.purl.org/net/inkel/inkel.foaf.rdf

15:12:23 <danbri> ERROR: rdfget: HTTP GET failed. Returning empty graph. error:302 "Found"

15:12:23 <danbri> I

15:12:24 <dc_rdfig> Label INIT not found.

15:12:25 <dc_rdfig> Label ERROR not found.

15:12:30 <danbri> hmm, not dealing w/ redirects yet

15:13:33 * DanC scribbled Miami expenses into my PDA in N3... extracted them with python the other day... starts recording the steps in http://www.w3.org/2003/04dc-mia/Makefile

15:13:56 <DanC> pythonPlug: python's stdlib groks redirects for free

15:14:41 * DanC noodles on how to get danbri to return to cwm-based scuttering...

15:15:31 <danbri> I don't know Cwm's APIs, but plan on learning them. I don't believe I can do what I want with commandline / N3 interface to Cwm

15:16:15 <danbri> I want, after each doc is parsed, to stash it in SQL store, replacing older content if from same src.

15:16:36 <DanC> plan... no time like the present!

15:16:38 <danbri> I think I know enough of EricP's stuff to attempt it using the annotea/perllib codebase too. Not yet same w/ Cwm.

15:17:35 <danbri> heh, I have plenty mission for today already. Go outside before the sun goes in, tidy my room, and get my head around using debugger in Emacs.

15:17:39 <libby> danbri, i've got 101001 triples from a crawl a couple of days ago

15:17:49 <danbri> That said, if I wanted to learn Cwm API in 5 minutes, what should I read?

15:18:09 <DanC> toIcal.py

15:18:28 <DanC>http://www.w3.org/2000/10/swap/pim/toIcal.py

15:18:29 <dc_rdfig> E: http://www.w3.org/2000/10/swap/pim/toIcal.py from DanC

15:18:39 * mattb been spending a little time playing with sesame today

15:18:46 <mattb> anyone know if the authors frequent any irc places?

15:18:49 <DanC> E:toIcal.py -- -- convert RDF to iCalendar syntax

15:18:49 <dc_rdfig> Added comment E1.

15:18:50 <danbri> how'd you find it mattb?

15:18:55 <mattb> i like it a lot

15:18:58 <DanC> E:|toIcal.py -- -- convert RDF to iCalendar syntax

15:18:58 <dc_rdfig> Titled item E.

15:18:58 <danbri> Jeen stops by here when we talk query...

15:19:05 <mattb> been using the new SeRQL engine from the cvs version

15:19:50 <danbri> is that the newer QL mentioned on www-rdf-rules recently?

15:20:54 <danbri> lib, INIT: s.left.size=141 s.seen.size=789 current: http://www.ecademy.com/module.php?mod=network&op=foafrdf&uid=8

15:21:06 <DanC> E1: an example of using the cwm api

15:21:06 <dc_rdfig> Replaced comment E1.

15:21:25 <danbri> 141 left, seen 789 docs. Libby, how many distinct RDF docs do you find?

15:21:40 * danbri looks at toIcal.py

15:22:32 <mattb> danbri: yeah, it has "construct" queries to specify what graph you would like built from your results

15:22:55 <mattb> possible to do things like normally {depicts, depiction} -> {depiction} that way for example

15:22:56 <danbri> nice for composing bits of multiple merged sources?

15:23:10 <mattb> yeah, not sure about actual smush-merging

15:23:25 <mattb> it's more powerful than the usual rdql or similar, but not actual inferencing

15:23:33 <mattb> the store also has a "truth maintenance engine" or something

15:23:50 <mattb> which follows inferences and maintains consistency after deletes and so forth

15:24:02 <mattb> the thing i like most is it's very HTTP

15:24:04 <mattb> very REST

15:24:05 <danbri> DanC, "sts = load(addr)", returns a graph aka Formulae?

15:24:10 <mattb> update with POSTs, query with GETs

15:24:32 <mattb> i think it might be the web-based RDF storage i've been looking for to make picdiary rdf properly mutable

15:24:36 <mattb> so i can get back to annotation bots and the like

15:24:40 <libby> danbri, I seem to have about 915

15:24:51 <DanC> returns a graph... I think so. something pretty close, anyway. maybe a 4-ple store. you can call each/any methods on whatever it is.

15:24:55 <libby> not sure how distinct

15:25:33 * DanC sometimes wishes python had interfaces

15:25:50 <DanC> C# has interfaces, no?

15:26:22 <danbri> OK, I have an idea for converging. I'll ape those bits of the cwm api in my code, draft a scutter around that, then the python port sohuld just be a matter of learning enough python

15:26:39 <ericP_> ericP_ is now known as ericP

15:26:54 <danbri> ...the plan to learn cwm api was so i could implement something around my ruby code that'd be familiar and reasonably well thought out

15:27:15 <DanC> idea for converging: ok, I'll count that as progress

15:27:20 <danbri> for cal in sts.each(pred = RDF.type, obj = ICAL.Vcalendar):

15:27:29 <danbri> ...is this string-centric, or nodes/terms?

15:27:49 <DanC> not sure what you mean... nodes/terms, I think. RDF and ICAL are Namespace objects

15:27:57 <danbri> are there factories for ensuring we don't have multiple representations of the same resource?

15:27:59 <danbri> ok

15:28:16 <DanC> factories: I think so...

15:28:19 <danbri> is a node/term dependent on a particular graph/formulae? or free-floating?

15:29:05 <DanC> ... it would be great to have all these details documented (in the form of doctest module unit tests). but I think you can monkey-see-monkey-do based on toIcal.py pretty well, for now.

15:29:55 <DanC> dependent: very good question! timbl/eikon/I discuss that often.

15:30:16 <danbri> I think I probably could, yeah. The details could lead to unexpected behaviour, but I'm sorta expecting that so know what to watch for.

15:30:56 <danbri>http://www.w3.org/2001/12/rubyrdf/pack/tests/api/tc_basicapi.rb

15:30:57 <DanC> I hope I didn't get the impression that the cwm API is all worked out and stable and documented and all. But already we're benefitting from your willingness to converge... we're getting anothe set of eyeballs on the current design issues.

15:30:57 <dc_rdfig> F: http://www.w3.org/2001/12/rubyrdf/pack/tests/api/tc_basicapi.rb from danbri

15:31:20 <danbri> F:|I made a start at documenting the (somewhat emergent ;) behaviour of my RDF code

15:31:21 <dc_rdfig> Titled item F.

15:31:35 <danbri> F:descended from previous perl, and before that java, implementations.

15:31:35 <dc_rdfig> Added comment F1.

15:32:25 <DanC> E:I hope the cwm api will get documented using [doctest|http://www.python.org/doc/current/lib/module-doctest.html] sometime soon.

15:32:25 <dc_rdfig> Added comment E2.

15:32:26 <danbri> Yes, I'm finding this helpful... I don't mind coding in another language but I think coding much more without looking at the Cwm/Rdflib design would be dumb.

15:33:04 <DanC> E:and convergence with rdflib (LinkMe) is a goal too.

15:33:05 <dc_rdfig> Added comment E3.

15:34:19 <danbri> E:I was [http://esw.w3.org/topic/RdfPerlLib|looking at EricP's perllib] too. He has explicit 'atom dictionary' mechanism for ensuring quickly-comparable nodes.

15:34:20 <dc_rdfig> Added comment E4.

15:36:17 <DanC> E:re comparing symbols as pointers rather than as strings, I hope the cwm API uses [weakref|http://www.python.org/doc/current/lib/module-weakref.html] dictionaries soon

15:36:17 <dc_rdfig> Added comment E5.

15:36:29 <DanC> oops...

15:36:43 <DanC> no, E is right

15:37:02 <danbri> E:Jan Grant wrote a [http://www.w3.org/2001/sw/Europe/reports/rdf_api_comparison_report/|report] on RDF API comparision. It mentions provenance, didn't see much on terms/factories/etc

15:37:02 <dc_rdfig> Added comment E6.

15:37:32 <danbri> Yes, c# ha interfaces (er, I think; I'd be v suprised if not, being such a java clone)

15:38:17 <danbri> "Inheritance

15:38:17 <danbri> Both Java and C# doesn't supports multiple inheritance and provides interfaces as an alternative to it. However C# covers much more advanced issues in interfaces. Please check out the topic titled New Additions given at the end of this article. "

15:38:21 <danbri> --http://216.239.53.100/search?q=cache:7FKH9CJOlPcC:www.csharphelp.com/archives/archive96.html+c%23+sharp+java+interfaces&hl=en&ie=UTF-8

15:38:28 <DanC> C# has even more good ideas from M3 than Java does. I hope they applied M3 lessons to the stdlib as well as the language (i.e. don't repeat Java's locking contention in critical places nightmares)

15:38:30 <danbri> (dead article c/o google)

15:39:12 <DanC> any more money/person-months in the wp7 budget?

15:39:47 * DanC almost missed the link to WP7 description cuz of InYourFaceUrl

15:40:11 <danbri> ask libby, WP7 is all ILRT...

15:40:20 <DanC> can't I ask the web somehow?

15:40:32 <DanC> I see it starts in "Month 1" per http://www.w3.org/2001/sw/Europe/plan/workpackages/live/esw-wp-7.html

15:40:48 <DanC> can I find what month it's scheduled to end? and what month we're in?

15:41:13 * DanC is frustrated at lack of links to context from esw-wp-7.html

15:41:14 * danbri loses that coding-on-his-day-off feeling

15:42:11 <danbri> yeah, we should fix all those things.

15:42:14 <danbri> logger_1, pointer?

15:42:14 <danbri> See http://ilrt.org/discovery/chatlogs/rdfig/2003-04-18#T15-42-14

15:42:18 <danbri>http://ilrt.org/discovery/chatlogs/rdfig/2003-04-18#T15-42-14

15:42:19 <dc_rdfig> G: http://ilrt.org/discovery/chatlogs/rdfig/2003-04-18#T15-42-14 from danbri

15:42:25 <danbri> G:|More swad-e website todos

15:42:26 <dc_rdfig> Titled item G.

15:43:36 * DanC was sure he saw a schedule/chart somewhere; can't find it now, though

15:44:44 * DanC does some url-hacking...

15:44:46 <DanC> aha! http://www.w3.org/2001/sw/Europe/plan/workpackages/live/_gantt.html

15:44:52 * danbri files in http://esw.w3.org/topic/SwadEuropeTodos

15:45:15 <danbri> libby, to go back to original q, how long is WP7 work continuing? JanG especially?

15:45:28 <DanC> "Month 1 is May 2002."

15:47:18 <danbri> INIT: s.left.size=0 s.seen.size=1030 current: http://jibbering.com/rdfweb/1016746423488.rdf

15:47:18 <danbri> RDFDOC: count='690': uri:http://jibbering.com/rdfweb/1016746423488.rdf gave RDF graph #<Graph:0x40c7c41c> with 23 triples

15:47:18 <dc_rdfig> Label INIT not found.

15:47:19 <dc_rdfig> Label RDFDOC not found.

15:47:31 <DanC> does the 7.3 row in _gantt.html correspond to WP7 somehow? it seems to go thru month 12, i.e. May 2003. and 7.4 goes thru month 16 (counting on fingers...) Sep 2003.

15:47:52 * DanC reads http://www.w3.org/2001/sw/Europe/plan/workpackages/live/_delivtable.html

15:48:05 <danbri> yes, n.* deliverables are in WPn

15:48:30 <DanC> ooh... [[ 2003-04-28 (12) 7.3 7 Public release of reference implementation of an RDF API which may build upon existing implementations. Open Source/free software. (6 months, software, Pub.) ]]

15:49:17 <danbri> hmm was that jang or dave? jang i think...

15:49:51 * danbri goes out for a bit, back later

15:49:54 <DanC> 10 days should be plenty to sync with the cwm api. ;-)

15:50:12 <danbri> heh, sounds plausible to me

15:50:41 <DanC> the agenda for this tuesday's swad meeting (22Apr) is kinda empty now that we've moved proof to 29Apr.

15:50:55 <danbri> what's the Cwm API again?

15:50:58 <danbri> i) sts.each(pred = RDF.type, obj = ICAL.Vcalendar)

15:51:05 <danbri> ii) if sts.statementsMatching(RDF.type, comp, ICAL.Vevent):

15:51:19 <danbri> iii) kb = llyn.RDFStore()

15:51:28 <danbri> vi) sts = load(addr)

15:51:51 <DanC> that's a pretty good summary. (timbl might suggest iii is a lower-level interface)

15:52:06 <danbri> Trying to document mine and Eric's code a bit, I was left wondering about a set of tasks we could document across Cwm, PerlLib, RubyRdf...

15:52:11 <danbri> "How do I..."

15:52:15 <DanC> quite.

15:52:17 <danbri> -load a graph from a URI

15:52:31 <danbri> -create a new SQL-backed RDF store

15:52:33 * eikeon lurks

15:52:36 <danbri> -stash some RDF there

15:52:40 <DanC> esw:RdfApiFaq maybe...

15:52:54 <danbri> -refresh a named portion of the DB with this graph

15:53:00 <danbri> -serialize to RDF/XML

15:53:11 <danbri> -get a graph that matches (some blanked out statement)

15:53:16 <danbri> -do fancier queries

15:53:36 <danbri> -create a... statement... graph... term...

15:53:38 <danbri> etc

15:53:47 * DanC doesn't believe in SQL-backed triple-stores; sees SQL stores like documents.

15:53:47 <danbri> hi Eikeon!

15:54:20 <danbri> we needn't be in-your-face about their SQL-ness, except for that it usually requires some sysadmin/DBadmin tasks to get up and running

15:54:28 <DanC> hmm... was thinking of inviting eikon to 22Apr SWAD telcon, but remembered timbl won't be there, so we can't really effect major changes to the cwm api

15:54:34 <danbri> logging the transactions would be good

15:54:41 <eikeon> Hi Danbri

15:54:55 <danbri> tuesday night I'm going out (had tentatively regret'd already; now confirmed).

15:54:58 <DanC> ScheduledTopicChat on RDF apis?

15:55:19 <DanC> eikon, wanna take the lead on such a chat?

15:55:21 <danbri> Sounds like a good idea. DotGnu folks were interested too...

15:55:34 * danbri volunteers Eikeon too ;)

15:55:42 <DanC> timeframe: 3 or 4 weeks from now?

15:55:50 * DanC wonders if that bumps into Budapest travel

15:56:05 <DanC> timeAndSpaceBot, what's happening 3 or 4 weeks from now?

15:56:24 * danbri leaves for BUD 17th May

15:56:30 <eikeon> danbri: Sure.

15:56:37 <danbri> great!

15:57:23 <danbri> Weds April 30th perhaps?

15:57:37 <DanC> following the http://esw.w3.org/topic/ScheduledTopicChat pattern, that means you should say (now/soon) what mailing list(s) you'll send an announcement to a week in advance. I suggest www-rdf-interest

15:57:50 <danbri> yup, www-rdf-interest at least

15:57:55 * DanC asks hipAgent what I'm doing 30Apr...

15:58:48 <eikeon> April 30th works for me.

15:59:42 <DanC> [[ <http://calendar.sidekick.dngr.com/day?date=2003-04-30> foaf:topic [ xsdt:date "2003-04-30";

15:59:42 <DanC> is k:startingDate of [ rdf:value "Ian weekly "; is foaf:topic of <http://calendar.sidekick.dngr.com/event?id=71&event-date=2003-04-30>];

15:59:42 <DanC> is k:startingDate of [ rdf:value "Family time "; is foaf:topic of <http://calendar.sidekick.dngr.com/event?id=62&event-date=2003-04-30>];

15:59:42 <DanC> ].

15:59:43 <DanC> ]]

15:59:59 <danbri> Eikeon, can you scribble something in http://esw.w3.org/topic/ScheduledTopicChat (date/topic/etc)

16:00:20 * danbri ponders wikiname for api design writings

16:00:21 <DanC> no, scribble it in RdfApi or whatever the real topic is, with a link to ScheduledTopicChat

16:00:26 * eikeon is looking at ScheduledTopicChat now.

16:00:31 <danbri> . http://esw.w3.org/topic/RdfApiDesign ?

16:00:58 <danbri> yes, just the date topic in the ScheduledTopicChat page. Agenda, issues etc in (whatever name we choose)

16:01:00 <DanC> any names to steal from wp7?

16:01:04 <danbri> Don't think so

16:01:12 <danbri> any objections to RdfApiDesign ?

16:01:15 <eikeon> Or RDFApiDesign? ;)

16:01:19 <danbri> its a verb and a noun

16:01:22 <danbri> caps work badly :(

16:01:27 <eikeon> ;(

16:01:29 <DanC> RDFApiDesign isn't a MoinMoin:WikiName

16:02:04 <DanC> (I consider that a bug, not a feature. DanC is clearly a WikiName in the esw community)

16:02:30 <danbri> oh, that doesn't count either? ick

16:02:49 <DanC> of course, why should it be RDF but Api?

16:03:02 <danbri> cos we're talking about apis?

16:03:05 <eikeon> RDFApplicationsAndProjects is a WikiName

16:03:09 <DanC> ResourceDescriptionFrameworkApplicationProgrammerInterfaceDesign ;-)

16:03:18 <DanC> no, it's not, eikon

16:03:23 <eikeon> Or is that because it has two other caps besides RDFA

16:03:47 <DanC> MoinMoin supports arbitrary strings as topic names. you have to write ["RDFApplicationsAndProjects"]

16:03:59 <eikeon> Ah... true. re: RDF but Api

16:04:19 <danbri> RdfSoftwareDesignPatterns?

16:04:22 <DanC> KnowledgeBaseInterfaceDesign

16:04:33 <DanC> KnowledgebaseInterfaceDesign

16:04:33 <danbri> no, too AI

16:05:01 <DanC> I'

16:05:12 <DanC> I'll take 'knowledge base' over 'triple store' any day

16:05:23 <libby> danbri, WP7 finishes Aug I think, end aug

16:05:25 <danbri> WebDataInterfaceDesign

16:05:50 <DanC> WebDataInterfaceDesign +1, i think...

16:05:52 * eikeon likes interface over API

16:06:03 <danbri> any objections to WebDataInterfaceDesign?

16:06:18 <eikeon> Nope.

16:06:22 <DanC> though WebDataInterfaceDesign suggests network interfaces, ala joseki, too

16:06:25 * danbri creates it

16:06:28 <danbri> I can live with that

16:06:44 <danbri> just a little matter of writing now. gimme a few mins.

16:09:17 * DanC wishes for increased consistency in what we call http://rdfig.xmlhack.com/ . it calls itself scratchpad, but we often call it the chump and/or the weblog

16:09:48 <DanC> ^re "# Convene the "meeting" in IRC by chumping (LinkMe) the meeting call/agenda."

16:11:34 * DanC wrestles with an urge to document MeetingRecords, MeetingChair, ActionItem and such in esw space

16:14:12 * danbri doesn't worry, is crappy, saves http://esw.w3.org/topic/WebDataInterfaceDesign

16:14:27 <danbri> hack away, i'm offline for a bit.

16:16:17 <DanC> cool.

16:16:46 * DanC wonders why WebDataInterfaceDesign doesn't show up in backlinks from ScheduledTopicChat

16:18:57 <danbri> logger_1, pointer?

16:18:57 <danbri> See http://ilrt.org/discovery/chatlogs/rdfig/2003-04-18#T16-18-57

16:19:55 * danbri adds a link

16:20:44 * DanC saves 1st

16:22:49 <danbri> hmmmmmmmmmmmmmmmm

16:22:57 * danbri added a section 1 'meetings'

16:23:05 <danbri> "Thank you for your changes. Your attention to detail is appreciated.

16:23:05 <danbri> Nobody subscribed to this page, no mail sent."

16:23:07 * DanC hates headings in wiki pages

16:23:27 <danbri> they're a bit in yr face, at least with their css and the numbering...

16:23:35 <danbri> did you remove my text or did we lose an update?

16:23:52 <DanC> we seem to have lost an update

16:23:59 <danbri> crud

16:24:06 * danbri pretends it didn't happen for now

16:24:44 <DanC> it's more than just the formatting that I don't like. it's the tendency toward ListMode

16:25:03 * DanC looks the other way too

16:25:22 * eikeon should not have hit refresh... could have added it back in.

16:25:35 <danbri> its ok, danc said the same thing more concisely

16:31:04 * DanC considers getting rid of all the headings and putting the whole thing in FaqMode

16:31:25 * DanC leaves it be for now

16:31:57 <danbri> fine by me

16:32:46 * DanC goes back to extracting Miami expenses...

16:35:38 * DanC noodles on getting SWAD software described in RDF ala eikon's deely

16:43:27 <DanC> .google foaf namespace

16:43:28 <datum> foaf namespace: http://xmlns.com/foaf/0.1/

16:43:39 <DanC> .google rdf namespace

16:43:40 <datum> rdf namespace: http://www.w3.org/1999/02/22-rdf-syntax-ns

17:00:38 <DanC> ugh! notation3.py is a nightmare!

17:00:57 * DanC is trying to add support for unprefixed names

17:01:08 <DanC> clashed with integer parsing

17:01:41 * eikeon is interested in knowing how DanC's noodling goes... re: describing SWAD software in RDF.

17:07:10 <DanC> hmm... @namepattern "\d\d?[A-Z][a-z][a-z]\d\d\d\d" <http://www.w3.org/2002/12/cal/date#>.

17:08:17 <DanC> or @datatype "[+-]?\d+" <http://...#integer>.

17:14:06 <sbp`> how would RDF/XML and NTriples parsers be handed that knowledge? i.e. what's the equivalent construct in those serializations? shouldn't that information be stored in triples?

17:14:47 <sbp`> or is datatype munging something that needs to be built in?

17:15:12 * sbp` recently added datatype parsing to yet another NTriples+ parser...

17:24:22 <danbri> anyone got a Debian box they're prepared to risk running my code on?

17:24:33 * danbri just regen'd debian package

17:24:44 <mortenf> it'll be a cold day in h*ll... :)

17:24:53 <danbri> don't badmouth Hull!

17:25:00 <mortenf> nope, redhat only, sorry.

17:25:06 <danbri> chicken ;)

17:25:19 <mortenf> Hmmm, Hull...

17:25:35 <mortenf> is that convex or concave?

17:25:39 * mortenf shuts up now...

17:27:36 <DanC> sbp`, I just elaborated a bit in http://esw.w3.org/topic/ShorthandRDF

17:28:06 <sbp`> DanC: many thanks; reading...

17:29:21 <sbp`> ah, I see. it's a kind of macro system for datatyped literals. thanks for the explanation!

17:29:58 <DanC> sure.

17:31:50 * DanC really needs to get this expense report done; is really bummed that notation3.py is such a nightmare

17:32:14 <danbri> I thoughtyou had a nice grammer-generated n3 parser nowadays?

17:32:25 <DanC> almost

17:32:39 <DanC> that is: I do have a nice generated parser, but it's not integrated into cwm.

17:32:44 <DanC> never quite got it passing all the tests or something.

17:32:51 <danbri> you wouldn't have a few mins to test my debian packaging would you?

17:33:07 <danbri> it *should* just be 'vi /etc/apt/sources' / update / install...

17:33:20 <DanC> how many points to I get for testing your debian package?

17:33:43 <DanC> i.e. how many points are you offering? ;-)

17:34:02 <DanC> what goes in /etc/apt/sources?

17:34:41 <DanC> logger_1, pointer?

17:34:41 <DanC> See http://ilrt.org/discovery/chatlogs/rdfig/2003-04-18#T17-34-41

17:34:54 * DanC noodles on the fact that /etc/apt/sources is the ultimate expression of trust

17:35:06 <danbri> see section "9.1" in http://esw.w3.org/topic/RubyRdf

17:35:10 <danbri> deb http://www.w3.org/2001/12/rubyrdf/debian/packager/ ./

17:35:14 <danbri> apt-get update

17:35:14 * DanC would really like to put public keys in /etc/apt/sources.list

17:35:19 <danbri> apt-get install rubyrdf-ruby

17:35:30 <danbri> then this one should be runnable: http://www.w3.org/2001/12/rubyrdf/pack/tests/scutter/ayftest.rb

17:37:54 <DanC> apt-get update done

17:38:20 <DanC> Need to get 1136kB of archives. After unpacking 5648kB will be used.

17:39:20 <DanC> update-alternatives: unable to make /usr/lib/mozilla-cvs/plugins/javaplugin_oji.so.dpkg-tmp a symlink to /etc/alternatives/javaplugin_oji-mozilla-cvs.so: No such file or directory

17:39:25 <DanC> (not related to rubyrdf)

17:39:31 * danbri phews ;)

17:39:46 <DanC> er... seems to be installed. how do I run the tests?

17:39:48 <danbri> -rw-r--r-- 1 danbri danbri 39832 Apr 18 18:21 rubyrdf-ruby_0.91-1_all.deb

17:40:13 <danbri> Ah, the tests... for now, grab http://www.w3.org/2001/12/rubyrdf/pack/tests/scutter/ayftest.rb and see if it runs.

17:40:17 * DanC finds nothing in /usr/share/doc/ruby-rdf ; is pretty sure that's a debian-no-no

17:40:43 <danbri> I didn't debianise the tests as I wasn't sure yet where to install them. They're in the tar.gz. Ah, /usr/share/doc is for tests too?

17:40:56 * DanC exits superuser window...

17:41:03 <danbri> Had same questions w.r.t. Debianizing Cwm... Need to read the packaging guide...

17:41:04 <DanC> ok, I have /tmp/ayftest.rb

17:41:06 <DanC> now what?

17:41:12 <danbri> ruby ayftest.rb

17:41:36 <DanC> er... it's spewing stuff on my screen

17:41:37 <danbri> if it starts emitting a harvester log, we won. If not, hmmm.

17:41:43 <danbri> oh, ok. we won.

17:41:47 * danbri smiles

17:41:52 * DanC kills it

17:41:56 <danbri> awww

17:42:00 <DanC> INIT: s.left.size=9 s.seen.size=11 current: http://www.aaronsw.com/about.xrdf

17:42:00 <DanC> RDFDOC: count='4': uri:http://www.aaronsw.com/about.xrdf gave RDF graph #<Graph:0x40368fac> with 45 triples

17:42:00 <dc_rdfig> Label INIT not found.

17:42:01 <dc_rdfig> Label RDFDOC not found.

17:42:18 <danbri> Okay, that's encouraging. An RDF harvester that's apt-getable.

17:42:35 <danbri> I need to get some more debian clues though. Where/what re docs and tests.

17:43:32 <DanC> what does it do? i.e. if I let it run, what could I do?

17:44:07 <DanC> can I run queries over all the data out there?

17:44:49 <danbri> have a look at the code. See in http://www.w3.org/2001/12/rubyrdf/pack/tests/scutter/ayftest.rb I create a bunch of code fragments with Proc.new...

17:45:00 <danbri> they get called by the harvester at various lifecycle points.

17:45:23 <danbri> Right now, it just lets the data wash through it. My next thingy to write is one that stashes the graph into Postgres

17:45:41 <danbri> (it also installed code that does Squish2SQL translation for postgres storage)

17:45:55 * DanC doesn't understand the urge to use an SQL server as an RDF cache

17:46:15 * mortenf doesn't understand the problem?

17:46:36 <DanC> "have a look at the code" i.e. "UTSL" is not a very satisfying answer to "why should I let this code I barely trust run on my machine for a while".

17:47:09 <danbri> have a look at the tiny little script...

17:47:48 <zoyd> hi folks.

17:47:49 <DanC> I'm not in "because it's there" mood re ruby-rdf, I guess. I was basically asking if there was some use case that it addresses, but I gather no, it's just a "let's play with the data" sorta thing.

17:48:13 <danbri> If you let it run, it will (a) make lots of http connections to URIs it finds from crawling RDF (b) emit things to STDOUT (eg. it snoops for Airport descriptoins) (c) possibly consume more and more memory (but I think that was just on windows).

17:48:26 <danbri> OK, be that way.

17:48:28 * danbri wanders off

17:48:38 <DanC> ruby seems to entice folks to do the same sort of stuff that perl does.

17:48:42 <DanC> 1st line of uncommented code is

17:48:44 <DanC> $:.unshift '../../lib/'

17:48:57 <zoyd> wanted to know what are the general steps to create a module for RSS/FOAF or a new spec altogether.

17:49:01 <Morbus> then you've got the wrong folks, not the wrong langauge ;)

17:49:40 <mortenf> zoyd, generally step 0 is looking for something already existing...

17:51:46 <zoyd> ok.

17:51:56 <mortenf> what's your subject?

17:52:38 <zoyd> mortenf: a business directory.

17:53:16 <mortenf> aha, like modelling organizations, or a collection?

17:53:25 * nb is away: afk

17:53:31 <zoyd> collection.

17:53:44 <mortenf> k.

17:53:51 <zoyd> decentralized business network.

17:54:08 <mortenf> but "connected"?

17:54:44 <zoyd> mortenf: yes, just like FOAF, if that's what you're referring to.

17:55:03 <mortenf> there's foafcorp and other related activities going on.

17:57:26 <zoyd>http://rdfweb.org/foaf/corp/intro.html?

17:57:27 <dc_rdfig> H: http://rdfweb.org/foaf/corp/intro.html? from zoyd

17:57:42 <mortenf> yep, and "they rule".

17:58:38 <mortenf> H:|foafcorp

17:58:39 <dc_rdfig> Titled item H.

17:59:15 <mortenf> there's also another project, that I can't seem to find at the moment, perhaps danbri remembers?

18:00:42 <zoyd> brb

18:03:40 * zoyd_ is back

18:04:45 <mortenf> another place to look for vocabs is http://rdfschema.info/ and of course cyc - http://www.cyc.com/cycdoc/vocab/vocab-toc.html

18:07:33 <zoyd_> ok.

18:07:47 <danbri>http://hierarchies.org/

18:07:48 <dc_rdfig> I: http://hierarchies.org/ from danbri

18:07:54 <danbri> I:|Hierarchies.org: Who Owns Who

18:07:55 <dc_rdfig> Titled item I.

18:07:55 <mortenf> thx.

18:08:13 <danbri> I:They showed up in #foaf a few weeks back, I forget the IRC nicks offhand. Were working on an RDF view...

18:08:13 <dc_rdfig> Added comment I1.

18:08:36 <danbri> I:Talked about extending foafcorp vocab to cover more of what their app does...

18:08:36 <dc_rdfig> Added comment I2.

18:12:40 <zoyd_> mortenf: and what's step 1?

18:12:46 <mortenf> heh.

18:13:34 <mortenf> read and learn from others' schemas, and send a pointer in here or to the mailing list.

18:13:58 <mortenf> danbri, any other suggestions?

18:14:25 <mortenf> that is, other than drafting a schema for comments?

18:17:23 <zoyd_> hmm

18:20:08 <danbri> its good to accompany schema drafts with instance data examples, ideally with a story alongside giving motivation/context so people can understand what you're trying to achieve.

18:20:45 <danbri> having those be valid RDF/XML helps too; some combination of http://www.w3.org/RDF/Validator/ and reading the RDF syntax spec and asking here should help...

18:21:31 <zoyd_> fine.

18:22:44 <zoyd_> zoyd_ is now known as zoyd

19:04:38 <mortenf>http://rdfweb.org/pipermail/rdfweb-dev/2003-April/002945.html

19:04:39 <dc_rdfig> J: http://rdfweb.org/pipermail/rdfweb-dev/2003-April/002945.html from mortenf

19:05:09 <mortenf> J:|Announcement of proposal: Names in FoaF

19:05:09 <dc_rdfig> Titled item J.

19:06:29 <mortenf> J:Proposal: [http://www.wasab.dk/morten/2003/04/foaf-names-proposal-1.html|Names in FoaF]

19:06:29 <dc_rdfig> Added comment J1.

19:06:57 <DanC> sigh...

19:06:59 <DanC> File "RDFSink.py", line 224, in newLiteral

19:06:59 <DanC> if dt != None: raise ValueError("This sink cannot accept datatyped values")

19:06:59 <DanC> ValueError: This sink cannot accept datatyped values

19:08:57 * DanC reads J

19:11:03 <DanC> J:using rdf:li for sortname strikes me as massive overkill

19:11:03 <dc_rdfig> Added comment J2.

19:11:51 <mortenf> J:Yep, if it was for only that.

19:11:51 <dc_rdfig> Added comment J3.

19:17:21 <DanC> J:I spent some time studying naming in [http://esw.w3.org/topic/TopicMaps|TopicMaps] last night. basename/sortname/displayname are fundamental parts of the topicmap model

19:17:21 <dc_rdfig> Added comment J4.

19:17:40 <DanC> J:and they have contexts for names

19:17:40 <dc_rdfig> Added comment J5.

19:18:24 <DanC> J:that appeals to me. e.g. (in LTM): [texas = "TX" / usps

19:18:25 <dc_rdfig> Added comment J6.

19:18:33 <DanC> J6:that appeals to me. e.g. (in LTM): [texas = "TX" / usps]

19:18:33 <dc_rdfig> Replaced comment J6.

19:19:52 <DanC> J:see also [http://esw.w3.org/topic/PropertiesForNaming|PropertiesForNaming]

19:19:52 <dc_rdfig> Added comment J7.

19:20:58 <DanC> J:and recent mail to public-webont-comments (LinkMe) about icelandic names and such.

19:20:59 <dc_rdfig> Added comment J8.

19:22:51 <DanC> J5:and they have scopes for names

19:22:52 <dc_rdfig> Replaced comment J5.

19:24:55 <DanC> J:and [http://lists.w3.org/Archives/Public/public-webont-comments/2003Apr/0003.html|recent OWL requirements comments] about firstname/lastname; thread goes into icelandic names and such.

19:24:55 <dc_rdfig> Added comment J9.

19:25:33 * mortenf follows J9

19:25:53 <DanC> writes john cowan: [[[ The only really safe scheme is "full presentation name" vs. "sortable

19:25:53 <DanC> name". This does not leverage the typically high redundancy between

19:25:53 <DanC> these two values, but it does work in all cases. ]]]

19:26:31 <DanC> I wonder what icelandic folks do when publishing in international journals and such.

19:27:16 <DanC> i.e. what do they call themselves in international contexts

19:30:34 <DanC> J:there's a basic question of "what are you trying to model?" if you're trying to find an existing international convention for modelling people's names, there doesn't seem to be a ubiquitously deployed one. If you're making up something new for the foaf community, that's a whole different ballgame.

19:30:34 <dc_rdfig> Added comment J10.

19:31:01 <DanC> J:I wonder what ldap and vcard do.

19:31:01 <dc_rdfig> Added comment J11.

19:32:00 <mortenf> J:Searches high and low yield nothing but single-strings or first/last/given/family - nothing out of the ordinary.

19:32:00 <dc_rdfig> Added comment J12.

19:33:52 <DanC> J:I have some use for surname. see [meeting people|http://rdfig.xmlhack.com/2003/04/17/2003-04-17.html#1050544833.016718]

19:33:52 <dc_rdfig> Added comment J13.

19:34:23 <DanC> # In our model of the world, a person has just one surname.

19:34:23 <DanC> foaf:surname a owl:FunctionalProperty.

19:34:26 <DanC> ^cheating?

19:35:31 <mortenf> perhaps cheating, but in general not really useful (for anything but limited tests).

19:36:02 <DanC> *not* useful to assume people have at most one surname???

19:36:25 <DanC> surname is one of the most relaible way to tell people apart, in my experience.

19:36:27 <mortenf> that as well, but also to assume that no two people at a meeting share surname.

19:36:50 <DanC> I'm not assuming surname is unambiguous

19:37:02 <DanC> only that it's unique. i.e. different surnames -> different people

19:37:44 <mortenf> yeah, ok, that's true, but still not enough?

19:37:56 <DanC> not enough for what?

19:38:13 <DanC> I *was* able to detect a conflict in the meeting record.

19:39:19 <mortenf> but you wouldn't have been able to determine all conflicts - as I said, it's fine for testing.

19:40:35 <mortenf> for your use case, I'd suggest using foaf:name instead. Even though it's not an owl:FP, it's much closer.

19:44:05 <DanC> yes, I can find all conflicts of the form "an action was assigned to somebody with surname ?X while the list of attendees includes noone with that surname"

19:44:24 <DanC> no matter how big the meeting.

19:44:27 <mortenf> of course.

19:46:21 <DanC> but foaf:name won't work; I consider "Dan Connolly" and "Daniel W. Connolly" both foaf:names for me.

19:46:37 <DanC> so it's not a FunctionalProperty

19:47:17 <DanC> when a (traditional U.S.) woman gets married and takes her husband's family name, do they call that her surname?

19:48:20 <mortenf> good questions.

19:49:25 * DanC googles, finds a bunch of geneology stuff

19:49:48 <mortenf> I didn't put surname in the proposal, since it seems to be (close to) a synonym for family name

19:49:55 * DanC checks vcard...

19:51:21 * mortenf eats, back in a bit...

19:51:54 <DanC> FN:Mr. John Q. Public\, Esq. -- http://www.ietf.org/rfc/rfc2426.txt

19:51:54 <dc_rdfig> Label FN not found.

19:52:02 <DanC> "Type special notes: This type is based on the semantics of the X.520

19:52:02 <DanC> Common Name attribute."

19:52:42 <DanC> "N:Stevenson;John;Philip,Paul;Dr.;Jr.,M.D.,A.C.P."

19:53:13 <DanC> I guess that's much like using <rdf:li> in foaf:sortname

20:13:09 <DanC> J:niftiest part of this proposal is the statistical data, which, for example, justifies sticking with surname over family_name

20:13:09 <dc_rdfig> Added comment J14.

20:15:42 <mortenf> heh, you really must like surname!

20:16:32 <mortenf> apart from capitalization etc., why do you prefer surname over family name?

20:19:13 <DanC> not sure.

20:19:41 <DanC> partly because family name seems like it should be a property of my family, not a property of me

20:20:21 <mortenf> yeah, ok, but that's where it originated?

20:20:46 <DanC> like I said, I'm not really sure why I like surname better than family name

20:21:09 <mortenf> that's ok, i'm just trying to find out if there's a meaning to the word that i don't get/know.

20:21:39 <DanC> partly because surname feels more precise...

20:21:43 * DanC checks dictionary..

20:21:53 <DanC> "family name

20:21:53 <DanC> n.

20:21:53 <DanC> See surname."

20:21:58 <mortenf> heh.

20:21:59 <DanC> -- http://dictionary.reference.com/search?q=family%20name

20:22:12 <DanC> so that dictionary prefers surname to family name too

20:23:01 <DanC> the etymology of surname is also clearly western european.

20:23:18 <DanC> surname has fairly wide applicability, but doesn't pretend to be a really world-wide concept.

20:23:21 <mortenf> "added ... and becomes a family name."

20:25:08 <DanC>http://dublincore.org/documents/2003/03/04/dcmi-terms/

20:25:08 <dc_rdfig> K: http://dublincore.org/documents/2003/03/04/dcmi-terms/ from DanC

20:25:19 <DanC> K:|DCMI Metadata Terms

20:25:19 <dc_rdfig> Titled item K.

20:25:39 <DanC> K:issued 2003-03-04 by DCMI Usage Board

20:25:39 <dc_rdfig> Added comment K1.

20:26:01 <DanC> K:"Typically, the name of a Creator should be used to indicate the entity."

20:26:01 <dc_rdfig> Added comment K2.

20:28:39 <DanC> K:where did agentName go?

20:28:40 <dc_rdfig> Added comment K3.

20:28:52 <mortenf> from where?

20:29:02 <mortenf> I don't remember seeing it in DC.

21:09:32 <danbri>http://www.wired.com/news/technology/0,1282,58422,00.html?tw=wn_ascii

21:09:32 <dc_rdfig> L: http://www.wired.com/news/technology/0,1282,58422,00.html?tw=wn_ascii from danbri

21:09:47 <danbri> L:|Planning for the Next Cyberwar, Wired News 2003-04-18

21:09:47 <dc_rdfig> Titled item L.

21:13:01 <libby> re software ontologies, there's some stuff in http://www.aktors.org/ontology/portal

21:13:25 <danbri> L:cites [http://dtsn.darpa.mil/ixo/|Introduction to IXO] (Information Exploitation Office), which points to [http://dtsn.darpa.mil/ixo/wishner/summary.htm|Director's IXO Overview]. Under [http://dtsn.darpa.mil/ixo/wishner/current_projects.htm|Current Projects] we find 'DARPA Agent Markup Language'.

21:13:26 <dc_rdfig> Added comment L1.

21:14:15 <danbri> L:"IXO is a new office at DARPA, launched in part because the Department of Defense, and related groups, have fully recognized C4ISR as a major contributor to battlefield effectiveness. T he President of the United States has publicly endorsed the need for technologies that continuously locate, track, and kill moving land targets ? a central focus of our office."([http://dtsn.darpa.mil/ixo/Wishner/info_exp.htm|Information Exploitation])

21:14:15 <dc_rdfig> Added comment L2.

21:14:46 <danbri> L:[+http://dtsn.darpa.mil/ixo/Wishner/images/Fig2.GIF]

21:14:46 <dc_rdfig> Added comment L3.

21:15:02 <danbri> L:+[http://dtsn.darpa.mil/ixo/Wishner/images/Fig2.GIF]

21:15:02 <dc_rdfig> Added comment L4.

21:15:12 <danbri> L4:=""

21:15:13 <dc_rdfig> Replaced comment L4.

21:15:19 <danbri> L3:+[http://dtsn.darpa.mil/ixo/Wishner/images/Fig2.GIF]

21:15:19 <dc_rdfig> Replaced comment L3.

21:15:38 <mortenf> ouch.

21:15:43 <libby> joy

21:17:03 <mortenf> well, darpanet led us to where we are now, so perhaps we should wait and see...

21:17:15 <libby> neat with the foaf names proposal morten...will take a look tomorrow properly...

21:17:21 <libby> night

21:17:57 <danbri> L:"In the Information Integration area, we are working on a DARPA agent mark-up language called DAML, which is now being used to provide semantics-based markups of intelligence reports. In the future, DAML will enable the Services to rapidly construct dynamic, agent-enabled, semantic, web-based C4KISR networks. We are developing other agent technologies and applying them to applications such as survivable logistics information systems in

21:17:57 <danbri> raLog project."

21:17:57 <dc_rdfig> Added comment L5.

21:18:51 <danbri> L:Possibly the only RDF application I've no enthusiasm for...

21:18:51 <dc_rdfig> Added comment L6.

21:19:43 <mortenf> L:Perhaps that should be "negative" instead of "no"?

21:19:44 <dc_rdfig> Added comment L7.

21:20:46 <mortenf> hmm, on the other hand i should probably shut up now, if they're already running tests on wild data...

21:20:49 <danbri> Perhaps that should be "revolted by"...

21:21:54 <danbri> L:Another poor taste KillOMatic GIF... +[http://dtsn.darpa.mil/ixo/Wishner/images/4.GIF]

21:21:55 <dc_rdfig> Added comment L8.

21:22:27 <mortenf> that's sick.

21:25:00 <danbri> L:I wonder if anyone yet has been killed via RDF-based technology. You know you've hit the bigtime when... :(

21:25:00 <dc_rdfig> Added comment L9.

21:55:31 <earle> L:if they have, at least you can know [how to classify it|http://www.isx.com/~phaglic/horus/ontologies/public_onts/horusweaponont.daml].

21:55:31 <dc_rdfig> Added comment L10.

21:57:26 <earle> L4:""

21:57:39 <dc_rdfig> Deleted comment L4.

22:01:18 <mortenf> hmm, re the onion model; from http://www.isx.com/~phaglic/horus/ontologies/public_onts/horuslocusplusont.daml: <daml:DatatypeProperty rdf:ID="proximityRadiusUnits">...

22:01:48 * earle nods; PeteHaglich's stuff is interesting

22:03:57 <earle> I like <rdfs:label>Area Of Uncertainty</rdfs:label> .

22:06:22 <mortenf> ouch, rdf:ID="EthnicCleansingIssue", i think i'd be sad going home from work every day working there...

22:06:46 <earle> Yes, there are some worrying things, like in the ontology for electronic warfare.

22:07:26 <earle> Still, this is one of the things we have to face as developers of a technology - that it will be used for a very wide range of purposes, not all of which we may approve of.

22:08:15 <earle> if I was making a wiki page about it, I'd call it the TechnologyProgenitorEthicalDilemma...

22:08:21 <mortenf> yeah, it just seems easier not knowing about it - hmm, erh, actually i wouldn't want them to hide too much, but still.

22:09:14 * danbri would be amused to see someone try to define terrorism over a DAML/OWL ontology

22:09:26 <earle> Heh.

22:10:06 <mortenf> ... with test cases including "events" from all nations and administrations...

22:10:22 <earle> Grief, imagine the working group.

22:10:26 <mortenf> ;)

22:11:07 * danbri been meaning to get in touch with the UN website folks...

22:11:10 <earle> "Bush Administration Accuses $country Of Developing Weapons Of Mass Classification"

22:11:45 <mortenf> for $country in (world\us) do;

22:12:08 <danbri> hmm http://www.isx.com/~phaglic/horus/turkorg.htm PKK are terrorists not freedom fighters...

22:13:07 <earle> Difficult issues.

22:13:36 <danbri> to put it mildly...

22:14:45 <earle> "The Horus Project" seems to be a popular name for projects.

22:15:09 * DanC returns from big-trash-day domestic duties, resumes noodling on PropertiesForNaming

22:15:54 <danbri>http://www.sfbg.com/MediaBeat/192.html

22:15:54 <dc_rdfig> M: http://www.sfbg.com/MediaBeat/192.html from danbri

22:16:06 <danbri> M:|On reuters re the term 'terrorism'

22:16:07 <dc_rdfig> Titled item M.

22:16:24 <danbri> M:'In a clarifying statement, released Oct. 2, the top execs at Reuters explained: "Our policy is to avoid the use of emotional terms and not make value judgments concerning the facts we attempt to report accurately and fairly."' (source?)

22:16:24 <dc_rdfig> Added comment M1.

22:16:50 <danbri> M:Same issue comes up in RDF/OWL vocabulary design, but hasn't to date proved too urgent / controversial.

22:16:50 <dc_rdfig> Added comment M2.

22:17:19 <DanC> does anybody use foaf data for their real address book? e.g. sync their cellphone addressbook with the web?

22:17:36 <DanC> or do auto-completion in their mail user agent based on foaf data?

22:17:40 <danbri> not me, my cell is woefully out of date

22:17:53 <danbri> i just remember email addresses or use reply-to...

22:17:55 <mortenf> not yet, still no rdfdb running.

22:18:22 <DanC> why would you need rdfdb? a python hack to sync with an ldap server was what I was gonna try

22:18:51 <danbri> I started towards an import-from-foaf for ldap, but never managed to get import-from-ldif working w/ slapd, failed at first hurdle

22:19:01 <mortenf> sounds interesting.

22:19:03 <danbri> i agree that'd be an interesting deployment path...

22:19:20 <danbri> much as w/ calendar, merge / mangle in rdf, then turn back into app-specific formats for actually using

22:19:22 <DanC> I'd prefer to do it completely serverlessly, but I think the easiest way to update evolution's autocomplete lists on my laptop and desktop is to share an ldap server between them.

22:20:40 <danbri> if you find a recipe for feeding ldif files to a freshly apt-got ldap server, i'm all ears...

22:20:59 <mortenf> [OT] tanlines: http://www.isx.com/~phaglic/humor/tanlines.jpg

22:21:12 <DanC> it's tricky to think about when 10^7 foaf:mbox triples are available a few hops from my homepage. I've been thinking about rules ala "if there's an event in my calendar involving ?WHO, make sure ?WHO's mailbox is in my autocomplete list"

22:21:58 * DanC played the apt-get ldap server game a while ago; didn't reach real closure, but kept notes in http://dm93.org/z2001/MailStorePuzzle

22:22:02 <mortenf> i've done partial merges of n3 files with a "deepcopy" sub (in Perl)...

22:22:08 * danbri hmms re http://www.salon.com/tech/wire/2003/04/18/darpa/index.html

22:22:14 <danbri> "DARPA cancels open-source software project after anti-war comments"

22:24:17 <DanC> hmm... maybe I didn't make ldap notes in MailStorePuzzle after all

22:29:44 * danbri only saw imap

22:30:11 <danbri> in http://dm93.org/z2001/SearchPage?source=jump&expr=ldap 0 hits

22:30:35 <danbri> <DanC> it's tricky to think about when 10^7 foaf:mbox triples are available a few hops from my homepage.

22:30:46 <danbri> that fwiw is my motivation for using an sql store as an RDF document cache

22:30:56 <DanC> slapd:

22:30:57 <DanC> Installed: 2.0.27-3

22:30:57 <danbri> I also use ntriples in local filesystem

22:31:43 <DanC> 10&7 foaf:mbox triples doesn't make me want to use an sql store. it makes me want to keep just the data I'm interested in my evolution addressbook

22:31:46 <DanC> 10^7

22:32:03 <danbri> I want to hack my harvester to allow rules such as 'traverse to docs that describe calendars' or 'traverse to CalendarDocs', ie the subject/object of rdfs:seeAlso statements, to focus the crawling a bit.

22:33:08 <danbri> yup, we're just fumbling around trying to find deployment/use models. My current hunch is 'crawl a lot of small but useful rdf docs and store them in medium performance RDF db, do selects against that to find things to index in application-optimised stores (addressbooks, calendars)'

22:34:31 <DanC> there's a time and a place for an SQL server. I think it's for the real source of the data, esp. when it's updated transactionally (i.e. concurrently).

22:35:10 <DanC> I suppose SQL is kinda handy cuz it's powerful enough to do graph matching.

22:35:46 <danbri> yes, there are Squish/RDQL/etc to SQL-self-join filters that are dead handy....

22:36:13 <danbri> I really don't like keeping data in SQL as its primary src, but as a medium-performance melting pot triplestore, it seems useful.

22:36:28 <DanC> have you played with any of the in-process SQL thingies? gadfly and the like? I saw one chumped recently.

22:37:18 <danbri> no, i haven't. i think i chumped another that was recommended to me. an in-memory C thing w/ py, ruby etc bindings.

22:37:39 <DanC> yeah, that's the one I'm thinking of.

22:37:57 <danbri> sql backing nice for admin stuff, you can drop down to the sql layer for query facilities the RDF query languages don't have yet.

22:38:12 <danbri> count and suchlike

22:38:21 * DanC really wants to get unstuck on this expense report

22:38:38 <DanC> [[[

22:38:39 <DanC> <http://calendar.sidekick.dngr.com/edit?id=451> foaf:topic [

22:38:39 <DanC> dngr:e_title """Marriott Biscayne Bay""";

22:38:48 <DanC> ...

22:38:49 <DanC> dngr:e_notes """3 nights, one meal

22:38:49 <DanC> in [ a HotelRoom; number 2718;

22:38:49 <DanC> in [= _:hotel; work <tel:+1-305-374-3900>];

22:38:49 <DanC> payee _:hotel;

22:38:49 <DanC> price 347.12; db VISA9688; cr MIT2000;

22:38:50 <DanC> checkIn 7April2003;

22:38:52 <DanC> checkOut 10April2003;

22:38:54 <DanC> Government Rate under DAML: $98.00+tax/night""";

22:38:56 <DanC> ].

22:38:58 <DanC> ]]]

22:39:47 * DanC needs to make .qif out of that and wants to make MIT-expense-report.html out of it

22:40:22 <DanC> [[[

22:40:23 <DanC> <http://calendar.sidekick.dngr.com/edit?id=452> foaf:topic [

22:40:23 <DanC> dngr:e_title """Quicksilver""";

22:40:24 <DanC> ...

22:40:35 <DanC> dngr:e_notes """

22:40:35 <DanC> from home; to MCI;

22:40:35 <DanC> price 28.50; db Visa9688; cr MIT2000;

22:40:35 <DanC> conf 940177.""";

22:40:35 <DanC> ].

22:40:35 <DanC> ]]]

22:41:54 <DanC> oh... also:

22:41:55 <DanC> dngr:year_start """2003""";

22:41:55 <DanC> dngr:month_start """4""";

22:41:55 <DanC> dngr:day_start """07""";

22:41:55 <DanC> dngr:e_allday """f""";

22:41:55 <DanC> dngr:hour_start """7""";

22:41:57 <DanC> dngr:min_start """15""";

22:41:59 <DanC> dngr:use_duration """true""";

22:42:03 <DanC> dngr:e_hour_dur """1""";

22:42:05 <DanC> dngr:e_min_dur """0""";

22:43:20 <dajobe> I think mattb has some experience with that mini SQL stuff, but IIRC eventually he ripped it out and replaced with MySQL since it just bloated memory, ran to a crawl on moderate data

22:43:43 <danbri> maybe it was him that recommended it to me...

23:10:16 * DanC got unstuck using str:scrape...

23:10:25 <DanC> qif:amount "347.12";

23:10:25 <DanC> qif:memo "3 nights, one meal";

23:10:25 <DanC> qif:payeeName "Marriott Biscayne Bay";

23:20:55 * DanC is winning...

23:20:56 <DanC> D2003-04-07

23:20:57 <DanC> M3 nights, one meal

23:20:57 <DanC> PMarriott Biscayne Bay

23:20:57 <DanC> T347.12

23:20:57 <DanC> ^

23:21:13 * DanC heads out to the Royals baseball game...

23:32:13 <earle> earle is now known as grault