Friend of a Friend (FOAF) IRC Chat Logs for 2004-09-15

00:06:31 <pjz> hrm

00:07:06 <pjz> so now I'm generating the list of stuff that can be added to a foaf file.

00:07:20 <pjz> but the ordering leaves something to be desired.

00:07:31 * pjz ponders.

00:08:42 <crschmidt> pjz: I'm not quite sure what you mean by "can be added to a foaf file" - you mean all of RDF?

00:08:51 <kpreid> heh heh.

01:40:44 <pjz> crschmidt: no, I just mean the bits that directly make sense inside a foaf:person

01:41:16 <crschmidt> pjz: ah, okay

01:41:20 <crschmidt> i ran the oppisite of that the other day

01:42:46 <crschmidt>http://crschmidt.net/foaf/properties.txt

01:42:57 <crschmidt> Things that a foaf:Person fits inside of in my data set

01:44:15 <pjz> right, and yes, I understand that I'm talking about something relevant to some particular set of schemas that I'm paying attention to

01:46:19 <kpreid> ^query SELECT ?p WHERE (?p <http://www.w3.org/2000/01/rdf-schema#domain> <http://xmlns.com/foaf/0.1/Agent>)

01:46:20 <julie> Query SELECT ?p WHERE (?p <http://www.w3.org/2000/01/rdf-schema#domain> <http://xmlns.com/foaf/0.1/Agent>) returned no results

01:47:10 <Arnia> ^query SELECT ?p WHERE (?p <http://www.w3.org/2000/01/rdf-schema#domain> <http://xmlns.com/foaf/0.1/Person>)

01:47:11 <julie> Query SELECT ?p WHERE (?p <http://www.w3.org/2000/01/rdf-schema#domain> <http://xmlns.com/foaf/0.1/Person>) returned no results

01:47:53 <Arnia> ^put http://rdf.netalleynetworks.com/ilike/20040830

01:47:55 * kpreid ^adds the foaf schema

01:48:00 <kpreid> ^query SELECT ?p WHERE (?p <http://www.w3.org/2000/01/rdf-schema#domain> <http://xmlns.com/foaf/0.1/Agent>)

01:48:01 <julie> [http://xmlns.com/foaf/0.1/holdsAccount] , [http://xmlns.com/foaf/0.1/made] , [http://xmlns.com/foaf/0.1/tipjar] , [http://xmlns.com/foaf/0.1/weblog] , [http://xmlns.com/foaf/0.1/msnChatID] , [http://xmlns.com/foaf/0.1/yahooChatID] , [http://xmlns.com/foaf/0.1/icqChatID] , [http://xmlns.com/foaf/0.1/aimChatID] , [http://xmlns.com/foaf/0.1/jabberID] , [http://xmlns.com/foaf/0.1/gender] , [http://xmlns.com/foaf/0.1/mbox_sha1sum] , [http:

01:48:01 <julie> //xmlns.com/foaf/0.1/mbox] ,

01:48:31 <kpreid> hmm. crschmidt: how about automatic common-prefix shortening in the output?

01:48:40 <Arnia> ^query SELECT ?p WHERE (?p <http://www.w3.org/2000/01/rdf-schema#domain> <http://xmlns.com/foaf/0.1/Person>)

01:48:41 <julie> [http://xmlns.com/foaf/0.1/pastProject] , [http://xmlns.com/foaf/0.1/currentProject] , [http://xmlns.com/foaf/0.1/publications] , [http://xmlns.com/foaf/0.1/topic_interest] , [http://xmlns.com/foaf/0.1/interest] , [http://xmlns.com/foaf/0.1/knows] , [http://xmlns.com/foaf/0.1/schoolHomepage] , [http://xmlns.com/foaf/0.1/workInfoHomepage] , [http://xmlns.com/foaf/0.1/workplaceHomepage] , [http://xmlns.com/foaf/0.1/myersBriggs] , [http://

01:48:41 <julie> xmlns.com/foaf/0.1/img] , [http://xmlns.com/foaf/0.1/plan] , [http://xmlns.com/foaf/0.1/family_name] , [http://xmlns.com/foaf/0.1/surname] , [http://xmlns.com/foaf/0.1/firstName] , [http://xmlns.com/foaf/0.1/geekcode] ,

01:48:53 <crschmidt> kpreid: Tell dajobe. I just use redland ;)

01:49:06 <kpreid> crschmidt: uh.

01:49:12 <kpreid> crschmidt: you don't format the output?

01:49:25 <pjz> I've been doing all my playing with just rdflib's TripleStore

01:49:31 <crschmidt> kpreid: I *could*, but I don't.

01:51:04 <pjz> who's the guy who maintains rdflib ?

01:51:18 * pjz remembers that he shows up here sometimes

01:51:41 <crschmidt> kpreid: I just attach each key in the returned statements, then stick a comma after each statement

01:52:19 <kpreid> i think in a few days we'll have three rdf query bots :)

01:52:30 * pjz isn't writing a query bot :)

01:52:40 <Arnia> kpreid: Don't forget, I'm writing mine too :p

01:52:40 * kpreid didn't say pjz was

01:52:54 <kpreid> well, four then!

01:53:04 <Arnia> Although mine is a little different

01:53:13 <crschmidt> easy enough ;)

01:53:14 <crschmidt> ^part

01:53:15 <kpreid> oh?

01:53:33 <kpreid> how is it different?

01:53:38 <crschmidt> It's natural language

01:53:41 <Arnia> My bot is a natural language understanding system

01:53:45 <kpreid> oho

01:53:59 <kpreid> i'll stick to unnatural language, then

01:56:12 <Arnia> The ultimate aim of my bot is for it to sit in a channel and annotate a channel log based on what's said

01:56:33 <kpreid> ook.

01:56:52 <Arnia> If I mention to someone that I wrote a document, it should understand that I created the document

01:57:42 <Arnia> Long way off... but do-able

01:58:09 <pjz> cool goal

01:58:33 <Arnia> To be honest, the hard bit is the lack of a decent reasoner

01:58:51 <Arnia> (or at least, a lack of a decent open-source reasoner for RDF)

01:59:45 <Arnia> The natural language bit is straightforward... my mapping from RDF to and from NL is straightforward... its just the mapping from ontologies to and from cognitive semantics that's tricky

02:02:23 <Arnia> Please note, I'm not saying the first two things are trivial :) They're hard, but they're well-defined. I have a lot of existing information and implementation to draw from

02:04:33 <kpreid> Hmm: I'm designing a RDF API. I have a pattern match function: given a pattern graph and a specimen graph, return possible mappings from bnodes in the pattern to bnodes in the specimen. If the pattern is empty, is it more appropriate to return no matches or one empty one?

06:05:34 <dngor> @find () (is) (human)

06:05:57 <workbench> (I'm not sure sungo) (is) (human)

06:05:57 <workbench> (to err) (is) (human)

06:05:57 <workbench> (but the racism, genocide, etc it) (is) (human)

06:05:57 <workbench> (human) (is) (human)

06:24:09 <balbinus_> balbinus_ is now known as balbinus

08:05:42 <libby> heyup pshab

08:13:30 <pshab> hi libby, you at ECDL?

08:15:14 <libby> nah, I'm staying home today

08:15:28 <pshab> how was yesterday afternoon/

08:15:45 <libby> not too bad

08:15:52 <dajobe> I'm there now

08:15:55 <libby> dinner was nice :)

08:16:01 <libby> oh, hi dave!

08:16:09 <dajobe> ih

08:16:10 <dajobe> hi

08:16:32 <dajobe> let me summarise this paper - I made a registry of OAI things by googing.

08:16:37 <dajobe> googling

08:49:37 * edd returns to england

08:50:22 <ephidrina> hi edd! where are you?

08:50:38 <edd> ephidrina: back home at york now. been travelling in california last week

08:50:56 <edd> ephidrina: that's where my body is. where my brain is, i'm not sure

08:51:01 * edd gets floaty sensations

08:51:21 <libby> heya edd, ephi

08:51:26 <edd> the libster!

08:51:44 <ephidrina> hey libby! svg con been & gone?

08:51:50 <libby> yep, home now

08:52:51 * ephidrina has the beginnings of a v.nice project at college ;)

08:53:05 <libby> lovely :)

08:53:06 <ephidrina> am tryinmg to steer in semantic-web-wards

08:53:31 <ephidrina> yesterday i unleashed the power of the t-shirt

08:54:16 <Cloud> morning all

08:56:06 <Cloud> quick question about foaf:dateOfBirth - has anyone used 0000 for the year?c

08:56:27 <edd> ephidrina: the power of the t-shirt is mighty!

08:57:22 <ephidrina> edd: it certainly gave them something to think about ;)

08:57:42 <edd> i wore my original foaf shirt at foocamp. people sure love it

08:57:49 <ephidrina> we are trying to build a template for multi-user non-linear narratives

08:57:58 <libby> foaf:dob is not officially in spec. I'd prefer foaf:birthday really...

08:58:09 <edd> ephidrina: that sounds very interesting

08:58:15 <edd> ephidrina: have you talked much with paul ford?

08:58:20 <edd> ephidrina: ftrain.com is sort of that.

08:58:32 <Cloud> i just saw it used by foafspace.com, so when i went to look for it that was the first 'unofficial' one that came back

08:58:35 <JimH> libby: isn't dob or similar in bio?

08:58:42 <ephidrina> edd: strangely enough, i have been corresponding with one p.ford

08:58:45 <Cloud> would your dob have elements for year, month, date

08:58:59 <edd> ephidrina: that is the bestest then

08:59:09 <libby> yeah we should probably just se that if it is jimh

08:59:24 <edd> ephidrina: you know XML Europe is returning to AMS next year?

08:59:39 <edd> ephidrina: except we're calling it "XTech" and broadening it out in topic matter.

08:59:41 <ephidrina> edd: that is excellent news - when is it?

08:59:54 <edd> ephidrina: so, there'll be a strong Web Dev track,,,

09:00:03 <libby> hm, you can do dob with event and date I guess in bio: http://vocab.org/bio/0.1/

09:00:06 <edd> ephidrina: and I'm wondering how to treat FOAF and friends.

09:00:22 <edd> ephidrina: last week in May

09:00:25 <JimH> it would be worth specifying which calender system is being used for the date too, for anyone born more than 2-300 years ago

09:01:04 <libby> cliud: http://www.google.com/search?q=site%3Ardfweb.org+dateofbirth&sourceid=mozilla-search&start=0&start=0&ie=utf-8&oe=utf-8

09:01:04 <edd> ephidrina: I'm not sure I'm content to divide a track between topic maps and rdf as before. it seems to me that FOAF and semwebby topics could possibly sustain an entire 2 day track by themselves

09:01:12 <libby> cloud, even

09:01:59 <ephidrina> edd: you could be right - perhaps a separate sub-conference with workshops?

09:02:45 <edd> yeah. some workshoppy stuff would be great.

09:02:57 <edd> there's the tutorial day before the conference, where i can devote a room to workshops

09:03:48 <edd> i suppose cost is the only issue for some foafers

09:03:53 <edd> how much was the reg for the Galway thing?

09:04:00 <Cloud> okay so a birthday for just the month/day

09:04:25 <libby> galway was free

09:04:33 <shellac> sweet EU cash

09:04:34 <libby> yeah maybe cloud

09:04:41 <libby> foafcamp was 100 euros

09:04:50 <edd> libby: how many campers came?

09:04:54 <libby> (including 2 breakfast, 2 lunch, one dinner)

09:05:02 <libby> 30-ish edd

09:05:16 <edd> maybe i should keep it as a subtrack then, hmmm...

09:05:17 <ephidrina> how was galway?

09:05:29 <libby> 100 people came when it was free :)

09:05:32 <edd> i'd definitely like the conf to explicitly support FOAFy things tho, rather than have it as incidental

09:05:37 <libby> cool :)

09:05:47 <ephidrina> good idea

09:06:57 * ephidrina wonders how to tempt an rdf expert to amsterdam when she has no budget ...

09:07:32 <Treenaks> ephidrina: seduction..

09:08:00 <ephidrina> thank u treenaks, for some age-old wisdom ;)

09:08:08 <Treenaks> :)

10:33:26 * Cloud has fun with foafspace.com

10:33:58 <Cloud> submitted my url on boards.ie, it's crawled through me to 1707 other members so far

10:34:34 * Cloud watches boards.ie move up through the foafspace top 20 hosts

10:37:58 <Zenethian> I really have to learn RDF and FOAF.

10:38:01 <Zenethian> I really do.

10:38:23 <Zenethian> I just need a good tutorial on it (or better yet a GOOD book on it. Which I believe doesn't exist)

10:40:43 <dajobe> semantic web primer by harmeln and ?sp grigoris

10:40:56 <dajobe> a relative newbie to rdf told me it was great

10:41:09 <Zenethian> Hmmm

10:42:01 <dajobe> if a book's what you wanted http://mitpress.mit.edu/catalog/item/default.asp?tid=10140&ttype=2

11:11:50 <Cloud> did anyone get the community graph on foafspace to work or is just me?

11:12:19 <eaon> didn't work here

11:30:40 <danja_> libby/danbri - I'm getting a 500 when I try to comment on planb

11:31:12 <danja_> comment is:

11:31:38 <danja_> Basildog says "whoofwhoofwhoofwhoof!"

11:36:27 <CaptSolo> danja, cloud - hi :)

11:36:35 <CaptSolo> dajobe :)

11:38:08 <Cloud_> Hi CaptSolo

11:57:53 <libby> I've taken all the commenting off on rdfweb.org and planb and moblog. I was haviong to remove comments every day. shoulda said something...

12:16:04 <CaptSolo> libby: yes, somment spam is a painful issue

12:16:14 <CaptSolo> i remember shelley had to remove comments as well

12:16:30 <CaptSolo> as for me, i do not thing removing comments is a solution

12:16:42 <libby> we were getting 100s a day, tiresome. couldn't be bothgtered with it after a wuile

12:17:15 <CaptSolo> ... that would be as if you left the city just because people were smoking on the streets and you had to breathe the smoke

12:17:43 <CaptSolo> 100s a day would mean some automated scripts which could be detected and blocked

12:17:59 <CaptSolo> (not that it woukd not take time to write that code)

12:18:17 <libby> yeha I was using blacklist. but that was too much effort. we very rarely get any comments form real people anyway. and people can always comment on their own blogs

12:18:29 * CaptSolo is probably lucky not getting so much comments / spam

12:19:15 <CaptSolo> libby: commenting on their own blogs is not completely the same thing. although if the trackback is provided, it helps

14:08:22 <crschmidt> Hi danbri_dna

14:15:27 <danbri_dna> hi

14:15:29 <danbri_dna> so at galway we talked about an irc meeting today, but i never announced it properly and am now jetlagged somewhat...

14:15:34 <danbri_dna> reckon it'd be ruder to cancel or to have an unannounced meeting?

14:15:36 <danbri_dna> it was 5pm if i recall...

14:15:41 <danbri_dna> uk time i mean

14:15:48 <danbri_dna> ppl can talk anyway of course...

14:15:52 <danbri_dna> how did last week's chat go?

14:16:54 <crschmidt> Last week's chat didn't go - 3 people signed up, and no one showed up :)

14:17:07 <danbri_dna> aw

14:17:13 <crschmidt> I know that some people plan on showing up around noon or a bit later, so you may want to be around to lead discussion

14:17:31 <crschmidt> IT was badly organized (my fault). I need to get more discussion going on one for next week, but have been busy and have a sick child in the house

14:18:17 * danbri_dna meant to org this week's one from japan, but wasn't online enough

15:39:44 <mortenf> danbri, do we have a verdict on the chat meet?

15:40:12 <mortenf> personally, i'd like a postpone, since i can't be here, but for "the project" it seems better to keep it

15:40:28 <danbri> there's no agenda or anything, and i'm jetlagged to hell, but chris says ppl were planning to turn up so i feel bad for going to sleep...

15:40:44 <mortenf> right :/

15:41:40 <danbri> I'd be interested to talk about what W3C might do re standards in this area...

15:42:01 <danbri> not the foaf ns as a REC so much as issues like trust, annotation, provenance, reputation, single signon, etc...

15:42:16 <mortenf> ... addresses, names...

15:43:12 <mortenf> in any case, a chat about the FOAFCommunityProcess doc seems like a good idea, to also get the people not present at galway involved

15:43:33 <danbri> hi chris

15:43:52 <crschmidt> howdy

15:44:54 <danbri> hi chris

15:45:13 * crschmidt thinks danbri's record player is broken.

15:45:24 <danbri> hi chris

15:45:33 <danbri> ;)

15:45:52 <danbri> sorry, i don't know if i hit 'up arrow' and return by accident or spaced and said the same thing twice :)

15:46:10 <danbri> re http://rdfweb.org/topic/FOAFCommunityProcess I'd be interested see what folks make of that

15:51:45 <edd> hi danbri

15:51:52 <danbri> heya

15:52:20 <balbinus> hi edd

15:52:40 <edd> heyh

15:53:01 * danbri blinks sleeplessly

15:53:14 <edd> danbri: sympathy. i have califorbian jeylag

15:53:22 * mortenf heads out -- http://www.commonme.org/blogs/001314.html

15:53:58 * ndw chuckles at califorbian

15:59:02 <libby> wow, cool, blogger dinner

16:00:00 <crschmidt> .t Z

16:00:03 <phenny> 2004-09-15T16:00:01Z

16:00:41 <bengee> is dat 5pm bst now?

16:00:43 <crschmidt> Well, it's 5pm BST, which was the decided time for the FOAFCommunity PRocess meeting today. Is there anyone here that's looking to participate?

16:00:57 <bengee> heh, how quick ;)

16:01:07 <edd> I'm looking to lurk :)

16:01:11 <crschmidt> Danbri was mentioning how organization didn't happen quite as it could have becaus of his trip to Japan, so I'm not sure if anyone is here for that :)

16:01:48 <pjz> I can lurk thru anything

16:02:01 <Cloud_> Hi

16:02:05 * inkel is lurking

16:02:08 <crschmidt> Well, I can sleep through anything, but that doesn't mean I don't prefer an evening in the country to a rock concert :)

16:02:20 <libby> heh, is anyone actually *not* going to lurk?

16:02:24 <danbri> Let's all lurk, see if anyone says something?

16:02:25 * bengee here for the chat, but it's ok if dan prefers some sleep..

16:02:27 <crschmidt> Well, we have some people here, so I"m just going to introduce myself, and start talking :P

16:02:54 * danbri interested in two things --- thoughts on http://rdfweb.org/topic/FOAFCommunityProcess and ideas on what W3C might consider doing in this-ish space...

16:02:58 <crschmidt> I'm Christopher Schmidt, http://crschmidt.net . I created the LJ FOAF output 6 months back or so, and since then have done lots of little hacky semweb projects

16:03:35 <crschmidt> Anyone else want to introduce themselves?

16:04:43 <danbri> I'm Dan Brickley, http://danbri.org/ and http://www.w3.org/ and started the foaf project w/ libby a while ago now, and lately have been doing foafish things in my dayjob at W3C (like havign the recent workshop in Galway w/ support from SWAD-Europe)

16:05:30 <pjz> I'm Paul Jimenez (http://www.place.org/~pj/ and am a not quite complete newcomer. I'm working on wrapping my head around RDF and FOAF and writing a real FOAFster.

16:05:47 <libby> libby miller, foaf, calendaring, image annotation, query, bots etc. jetlaggeed, just woke up, can;t type...

16:05:57 <bengee> bengee is now known as a_w

16:06:30 <crschmidt> Anyone else?

16:06:31 <a_w> Im a.w. and you may be interested in the Internet Business Logic system ;)

16:06:40 <a_w> a_w is now known as bengee

16:06:42 <bskahan> I'm Brian Skahan (http://bskahan.etria.com/foaf.rdf), working on foaf visualization tools in Zope

16:06:53 <bskahan> generally just lurking in here

16:06:55 <crschmidt> ^add http://bskahan.etria.com/foaf.rdf

16:06:56 <julie> Adding that to my database...

16:06:57 <julie> Added 72 statements from http://bskahan.etria.com/foaf.rdf. Model size is 377658.

16:07:06 <danbri> :) good bot

16:07:09 <bskahan> yeah

16:07:18 <balbinus> :)

16:07:33 <bengee> benjamin nowack, trying to build tools around the FOAFCommunityProcess doc

16:08:02 <inkel> I'm Leandro Mariano Lopez, aka inkel. I've created the speaks, reads and writes vocabulary, and I'm very interested in FOAF and Web Semantic development & research in Latin America (I'm from Argentina)

16:08:30 <Cloud_> I'm John Breslin, postdoctoral researcher at the Digital Enterprise Research Institute at NUI Galway, co-owner of community site boards.ie and founder of boards.jp (both FOAF enabled), co-organiser of the recent FOAF Workshop Galway 2004

16:09:09 <danbri> I stumbled across boards.jp lately, wondered if you were involved...

16:09:19 <Cloud_> Cool

16:09:57 <Jibbler> I'm Paul Mutton, http://www.jibble.org/ I jibble about a bit and wrote IRC Hacks for O'Reilly and I'm just finishing my PhD on Force Directed Layout of Diagrams and about to enter the real world of work and earn some real money...

16:10:15 <libby> hey jibbler, not seen you around for a bit

16:10:20 <Jibbler> been uber busy :(

16:10:21 <pjz> good luck with that last bit, jibbler :)

16:10:25 <Jibbler> but always here :)

16:12:10 <danbri> OK so I've a change to the foaf spec to put past you folks...

16:12:21 <danbri> it's on my laptop, switching ... [...]

16:12:48 <danbri_dna> back

16:13:00 <Cloud_> :)

16:13:20 <danbri_dna> Context: In RDF 1997-2001, the notion of the class rdfs:Resource and its relationship to literals was a bit... vague

16:13:39 <danbri_dna> nowadays, if I understand the formal semantics right, everything is considered a 'resource'

16:14:06 <danbri_dna> and the RDF/RDFS/OWL specs happen to lack a definition for the class of things that are the "non-literal resources"

16:14:21 <danbri_dna> ie. things which can be represented directly in RDF as the subjects of statements

16:14:34 <danbri_dna> ...ie URI-namable, or bnodes

16:14:47 <danbri_dna> (don't ask me why literals can't be subjects, i get confused!)

16:15:14 <danbri_dna> so anyway, various bits of the FOAF spec, the RDF version, define properties as having a 'domain' or 'range' of rdfs:Resource

16:15:28 <danbri_dna> ...although the intended meaning was a bit more restricted, meaning non-literal resources

16:16:00 <danbri_dna> after talking w/ kanzaki and libby and jimley, I think this should be changed

16:17:02 <danbri_dna> ...so the new way will say 'the domain of this property is a class that is disjoint with rdfs:Literal'

16:17:18 <danbri_dna> <rdfs:Class rdf:nodeID="NonLiteralResource">

16:17:18 <danbri_dna> <owl:disjoint rdf:resource="http://www.w3.org/2000/01/rdf-schema#Literal"/>

16:17:18 <danbri_dna> <!-- from a lunch discussion with Masahide Kanzaki -->

16:17:18 <danbri_dna> </rdfs:Class>

16:17:26 <libby> makes sense to me. would clarify the spec. I didn;t even realize it was an issue until kanzaki mentionned it

16:17:33 <danbri_dna> ...named inline in the foaf spec like this, rather than an addition to FOAF

16:17:49 <danbri_dna> ...since it is pretty core to RDF, I'd rather not clutter the foaf namespace with it.

16:17:58 <bengee> what about typing the rdf:Propertties as owl:ObjectProperties?

16:18:13 <libby> what will it screw up? e.g. protege etc? or maybe that'll just ignore it

16:18:19 <danbri_dna> <rdfs:range rdf:nodeID="NonLiteralResource" />

16:18:25 <danbri_dna> ...etc

16:18:59 <danbri_dna> hmm that'd work for the range of properties, but not the domain. Is there a way to do the inverse beyond saying it via an anonymous inverse?

16:19:29 <danbri_dna> how is this done in OWL? range of something that's an object property?

16:19:43 <danbri_dna> (we were offline when we discussed this, have barely been online since)

16:22:26 * bengee wonders if adding an rdf:type="&owl;"ObjectProperty wouldn't be enough. (although in owl full it might not work..)

16:24:11 <bengee> and why exactly is it an issue. are there reasoners that allow literals as subjects?

16:24:37 <bengee> afaik, there is no way to serialze that in rdf/xml..

16:24:45 <CaptSolo> damn, missed begining of the foafmeet again :/

16:24:59 <CaptSolo> (not that previous foaf ircmeet did happen at all)

16:25:00 <CaptSolo> hi all

16:25:04 <sh1mmer> Sorry I am late

16:25:08 * sh1mmer rushing back from work

16:25:30 * danbri_dna checks http://www.w3.org/2002/07/owl# to see what the namespace says

16:25:40 <danbri_dna> i find the owl specs a little inpenetrable

16:26:13 <bengee> they shoul have published an owl in owl spec..

16:26:22 <bengee> rater than just rdfs

16:26:34 <bengee> s/rater/rather

16:26:49 <danbri_dna> <rdfs:Class rdf:ID="ObjectProperty">

16:26:49 <danbri_dna> <rdfs:label>ObjectProperty</rdfs:label>

16:26:49 <danbri_dna> <rdfs:subClassOf rdf:resource="&rdf;Property"/>

16:26:49 <danbri_dna> </rdfs:Class>

16:27:16 <danbri_dna> yeah, I don't see how the meaning of ObjectProperty gets exposed except by building it in

16:27:23 <danbri_dna> so we can do both, I think.

16:27:50 <danbri_dna> for object properties, say so using the type mechanism and explicitly, via range being disjoint w/ Literal

16:28:12 <danbri_dna> similar for domain

16:28:57 <danbri_dna> except maybe its implicit for domain already, cos of the rdf constraint about properties not being directly applicable to literal nodes?

16:29:04 * danbri_dna doesn't see any harm in being more explicit

16:29:26 <danbri_dna> (but then i'm not feeding this stuff to owl engines at runtime, just trying to make the vocab definitions clearer where possible)

16:29:39 <sh1mmer> danbri_dna when you are ready i had a spec q

16:29:46 <danbri_dna> sure

16:30:20 <sh1mmer> GivenName, FirstName

16:30:23 <sh1mmer> whats the craic?

16:30:36 <danbri_dna> craic?

16:30:37 * CaptSolo is Uldis Bojars (CaptSolo) - http://kaste.lv/~captsolo/semweb/foaf-captsolo.rdf - interested in Semantics, RDF, FOAF - has created Resume RDf schema

16:30:52 <sh1mmer> danbri_dna how should we use them?

16:30:59 <Cloud_> Should givenname be givenName and family_name be familyName

16:31:04 <sh1mmer> I was thinking that maybe GivenName in an ordered list would be preferable

16:31:07 <danbri_dna> so there's two issues at least

16:31:21 <sh1mmer> I have been using firstname for my firstname and givenname for my middle name

16:31:41 <sh1mmer> but because of the lack of clarification some tools use my given name not my first name, so i am henry croucher

16:31:49 <danbri_dna> one is, that we have two pairs of properties. the ugly westerncentric first/last and the more i18n-friendly given/family

16:31:51 <Cloud_> Hi Henry

16:31:58 <sh1mmer> Hi John

16:32:23 <Cloud_> Actually it's Gerard in your world :)

16:32:25 <danbri_dna> next is that there is incosistency in the use of _, case/capitalization

16:32:29 <Cloud_> Yep

16:32:30 <sh1mmer> heh.

16:32:55 <sh1mmer> oh yes, didn't spot that, but yes

16:33:25 <danbri_dna> My view is that we can leave both pairs in, usefully, and flag first/last as being for direct represention of legacy data, and family/given being a bit more modern

16:33:45 <Cloud_> There's a couple of things like that, based_near -> basedNear, topic_interest (if kept) -> topicInterest or topicOfInterest

16:34:01 <danbri_dna> and then include a bunch of disclaimers saying that this is all very minimalistic compared to the full challenge of doing this in a way that works for everyone

16:34:22 <danbri_dna> so generally I was moving away from camelCase

16:34:49 <sh1mmer> danbri_dna well I still think that we need ordering on givennames

16:34:52 <balbinus> "legacy" is bad... everyone will use the way they prefer, whether it's "deprecated-but-legacy" or not... (to my mind)

16:35:00 <sh1mmer> loosing the context of their order is bad

16:36:16 <CaptSolo> choosing something one would be preferable - otherwise as balbinus says everyone will use legacy

16:36:21 <Cloud_> Well I guess if you have both, and have some legacy ones, not so bad - givenname is neither camelCase nor _ separated so maybe needs fixing

16:36:21 <danbri_dna> 'cos it reads funny (a) when we have acronyms / multiple capitals, or when there are two letters the same

16:36:24 <danbri_dna> so, this will be lossy

16:36:26 <danbri_dna> I'd be interested to see a W3C group or taskforce do a richer naming vocab

16:36:30 <danbri_dna> but until then, the main thing to rely on is foaf:name, which you can write just how you like

16:37:35 <danbri_dna> so, who has the most recent deployment stats we could look at, to judge cost of fixing/changing the _/case/etc structure?

16:38:00 <sh1mmer> danbri_dna I just don't see how legacy firstname/lastname can't be easly represented by using givenname/familyname

16:38:17 <danbri_dna> thats why we leave first/last in there!

16:38:24 <danbri_dna> cos you can't assume a mapping, in general

16:38:28 <sh1mmer> ah.

16:38:36 <sh1mmer> can we deprecate them?

16:38:41 <danbri_dna> PIM tools, palmtops etc., very very often have fields called 'first name', 'last name', ...

16:38:49 <danbri_dna> which makes them legacy

16:38:55 <danbri_dna> but not same as depracated

16:39:07 <danbri_dna> well, perhaps there are two senses of depracated

16:39:13 <danbri_dna> one is that this might get removed from the vocab

16:39:15 <sh1mmer> So this is about a direct sense of mapping?

16:39:23 <danbri_dna> the other is that it isn't a great way to represent stuff

16:39:35 <CaptSolo> why isn't first_name => given_name and last_name => family_name?

16:39:57 <danbri_dna> first/last imho can stick around so long as there is data managed using those fields

16:40:02 <danbri_dna> which basically means... forever

16:40:19 <danbri_dna> captsolo, 'cos that only holds in some countries/cultures

16:40:41 <danbri_dna> japanese folk often write the family name first, for eg.

16:41:00 <danbri_dna> its considered a v western-centric way of describing name parts

16:41:04 <CaptSolo> :/

16:41:24 <danbri_dna> see http://rdfweb.org/pipermail/rdfweb-dev/2003-August/011756.html

16:41:30 <CaptSolo> that leaves us with 'bad' or inconsistent data

16:41:46 <danbri_dna> [[

16:41:46 <danbri_dna> >

16:41:47 <danbri_dna> Hi Dan,

16:41:47 <danbri_dna> I seem to recall that when I last looked, foaf had taken

16:41:49 <danbri_dna> some of the features vCard got right (like family name

16:41:53 <danbri_dna> and given name) and replaced them by broken ethnocentric

16:41:57 <sh1mmer> danbri_dna I still think it would be preferable to encourage people to map from legacy data into the preferable format

16:42:01 <Cloud_> Yep, could given name include middle name (or second names)

16:42:01 <danbri_dna> constructs like surname and first name). Looking now at:

16:42:03 <danbri_dna> http://xmlns.com/foaf/0.1/

16:42:05 <danbri_dna> I see that both the culturally neutral forms and the

16:42:07 <danbri_dna> Western-centric forms appear to be legal. I don't know

16:42:09 <danbri_dna> whether that is a recent change. In any event, by

16:42:11 <danbri_dna> allowing the biased forms, foaf still appears to be

16:42:13 <danbri_dna> broken. It would be just as legitimate to consider:

16:42:17 <danbri_dna> first name = family name

16:42:23 <danbri_dna> as to consider:

16:42:27 <danbri_dna> last name = family name

16:42:31 <danbri_dna> Thus, assertions such as:

16:42:35 <danbri_dna> the person's first name is Misha

16:42:37 <sh1mmer> danbri_dna well china and india do that.

16:42:39 <danbri_dna> the person's last name is Wolf

16:42:43 <danbri_dna> tell me only that the words "Misha" and "Wolf" each

16:42:45 <danbri_dna> occur somewhere in the person's name.

16:42:47 <danbri_dna> ]]

16:42:49 <danbri_dna> oops big paste

16:42:53 <danbri_dna> I agree, we should encourage that

16:42:53 * CaptSolo quickly looks at what fields mail program has ( first-name, middle-name, last-name, nick, displayed-name (as diplayed in mailer), gender, photo, groups (folder belonging to) )

16:42:57 <danbri_dna> *however* by restricting the expressiveness of foaf to just the more culturally neutral properties, we risk encouraging auto-mapping which will be innaccurate

16:43:06 <danbri_dna> better to represent bad data within the sandbox of first/last than to have folk shovel it out as family/given without really being careful

16:43:23 <sh1mmer> danbri_dna but I think you just proved that first / last names risk more ambiguity

16:43:40 <sh1mmer> family / given name aren't ambigious

16:43:42 <danbri_dna> generally you need to know more about a dataset

16:43:56 <danbri_dna> Yes, I am very much in favour of family / given being the preferred form

16:44:16 <sh1mmer> right but if I am in india I can map my firstname to family name and my lastname to given name

16:44:19 <antont> this discussion somehow reminded me of the history of the Internactional Classification of Diseases (IDC), which is currently in version 10 afaik, but making the 1 version in the late 1800s was reportedly strange

16:44:23 <danbri_dna> I'm just arguing that there are tons and tons of databases etc that do it wrong, and that it will be valuable to be able to pass that data between applications using rdf/xml

16:44:24 <sh1mmer> in the uk i would do it the other way around

16:44:38 <sh1mmer> if we used only family and given names that wouldn't be an issue

16:44:43 <danbri_dna> ...not necc to publish 'in my foaf file', but foaf can be used for import/export in a variety of envs

16:45:08 <JimH> even family name isn't necessarily simple, the Spanish use both parents' "family names"

16:45:09 <sh1mmer> danbri_dna we can't fix bad data conversion, but we can fix potentially ambigious properties

16:45:31 <danbri_dna> their purpose, I'm suggesting, is to represent ambiguous data.

16:45:48 <sh1mmer> ah

16:45:56 <danbri_dna> can you live with that?

16:46:03 <bengee> maybe we could need another term_status value to better represent a term's lifecycle state, e.g. unstable->testing->stable (=success after testing) vs. unstable->testing->not_cool/discouraged/.. (= didn't really work *after* testing)

16:46:14 <sh1mmer> I can, but it needs to be very clearly marked

16:46:17 <JimH> I think the Portuguese use a similar system to Spain, hence so does most of South America

16:46:24 <sh1mmer> otherwise its too tempting to western culture

16:46:28 <danbri_dna> i'd like to move on to the _ vs Camelcase etc thing

16:46:39 <danbri_dna> Yes, I agree it needs careful flagging

16:46:41 <sh1mmer> JimH I don't see any reason you couldn't have two family names

16:46:49 * sh1mmer looks at spec

16:47:13 <danbri_dna> basically the spec needs a 'naming' section, to cover the issue, not just a bit of text on each naming property

16:47:28 <danbri_dna> I'll take an action to add one, then each definition can point to that

16:47:40 <danbri_dna> loggerf, pointer?

16:47:40 <danbri_dna> See http://www.ilrt.bris.ac.uk/discovery/chatlogs/foaf/2004-09-15#T16-47-40

16:48:24 <sh1mmer> Ready for another issue?

16:48:26 <danbri_dna> action: danbri update FOAF spec to add a main section on naming, to provide a unified treatment of first/last/given/family addressing Misha's concern re first/last http://rdfweb.org/pipermail/rdfweb-dev/2003-August/011756.html

16:48:37 <danbri_dna> not yet :)

16:48:45 <danbri_dna> so bengee suggested another term_status...

16:49:08 <danbri_dna> I think we'll come back to that topic as the W3C SW Best Practices WG taskforce on vocab management gets underway

16:49:29 <danbri_dna> ...if you've got your thoughts written up on that, could draw their attention to it, bengee

16:49:31 <sh1mmer> yeah sounds good

16:49:37 <danbri_dna> the VM TF is only just starting

16:49:44 <sh1mmer> both the suggestion and the forum to deal with it

16:50:21 * bengee really looking 4ward to the VM TF

16:50:35 <danbri_dna> I should also mention that various W3C colleagues have asked me to think about possibilities for bringing some FOAF work closer into W3C. There are a bunch of things that could mean, and which deserve a separate meetting w/ more advance materials (agenda etc).

16:51:20 <danbri_dna> ...a SW BP TF for FOAF stuff, as a channel from FOAF world into W3C (eg. "We want a better naming vocab pls", "here's how we digitally signed our RDF") might be one possibility. Another might be a WG in the trust/reputation space.

16:51:49 <danbri_dna> OK so sticking w/ naming, I'd like to talk about givenname vs family_name

16:51:58 <danbri_dna> ...having those two spelt differently seems wrong

16:52:06 <danbri_dna> ...do we know how widely used they are?

16:52:11 <danbri_dna> anyone got stats handy?

16:52:22 <sh1mmer> crschmidt maybe a good person

16:52:37 <crschmidt> I don't have a lot of stats, one second

16:52:53 <danbri_dna> thx

16:52:58 <danbri_dna> libby, how are your stats?

16:53:08 <libby> old.... :/

16:53:24 <libby> hang on...

16:53:50 <danbri_dna> so, stuff in the namespace at the moment: foaf:nick foaf:name foaf:firstName foaf:givenname foaf:surname foaf:family_name

16:54:03 <libby> [as an asside, qandeers ardo is writing a crawler based on an html crawler. should be good :) ]

16:54:09 <danbri_dna> firstName and surname "go together" (though the camelcase is annoying)

16:54:10 <libby> anders ardo, rtaher

16:54:21 <danbri_dna> [anders doing *rdf* crawler?]

16:54:29 * danbri_dna has flashbacks to the DESIRE project :)

16:54:32 <crschmidt> Yeah, creating this data is not easy, I'll have to get back to you on stats

16:54:41 <danbri_dna> if you could that'd be great

16:54:55 <danbri_dna> can i record an action on you, chris? will chase it up for a next time

16:55:02 <crschmidt> Absolutely

16:55:07 <libby> it's the same code asthe desire project, rewritten :)

16:55:10 * CaptSolo does not find camelCase annoying

16:55:28 <danbri_dna> action: crschmidt to circulate crawler stats on foaf:nick foaf:name foaf:firstName foaf:givenname foaf:surname foaf:family_name

16:55:34 <libby> itas annoying cos firstName and surname are different

16:55:39 <libby> i.e. not surName

16:55:42 <danbri_dna> (yep, its just re consistency)

16:56:02 <danbri_dna> (well w/ surname its ok, but the later naming vocab doesn't capitalize)

16:56:15 <danbri_dna> So we're nearly at the top of the hour

16:56:28 <libby> re stats, this is what I have for foaf properties: http://swordfish.rdfweb.org/discovery/2004/02/stats/foafstats.jsp?schema=http://xmlns.com/foaf/0.1/

16:56:32 <libby> but is pretty old

16:56:34 <danbri_dna> Would anyone be interested in another meetup, same time next week?

16:56:43 <sh1mmer> Could we make it 15 mins later?

16:56:47 * sh1mmer has to get home from work

16:56:51 <libby> [[

16:56:52 <libby> 105 http://xmlns.com/foaf/0.1/givenname

16:56:52 <libby> 97 http://xmlns.com/foaf/0.1/lastName

16:56:53 <libby> ]]

16:57:00 <danbri_dna> I'd like to get some progress on the naming vocab and having a meeting to report to helps ;)

16:57:14 <libby> [[

16:57:14 <libby> 80 http://xmlns.com/foaf/0.1/family_name

16:57:15 <libby> ]]

16:57:26 <sh1mmer> danbri_dna we can also whip you if you think it'd help

16:57:29 <danbri_dna> I prefer it on the hour, ideally sh1m. but there are logs...

16:57:35 <libby> [[

16:57:35 <libby> 80 http://xmlns.com/foaf/0.1/family_name

16:57:36 <libby> ]]

16:57:38 <sh1mmer> danbri_dna fair enough :)

16:57:39 <libby> oops

16:57:45 <danbri_dna> or 1/2 hour

16:57:47 <libby> [[

16:57:48 <libby> 23 http://xmlns.com/foaf/0.1/firstname

16:57:48 <libby> ]]

16:57:59 <libby> [[

16:58:01 <libby> 16 http://xmlns.com/foaf/0.1/familyname

16:58:01 <libby> ]]

16:58:02 <danbri_dna> ok those stats are so low I'd happily update all those rdf files by hand ;)

16:58:03 <libby> sigh.....

16:58:22 <libby> perhaps we could set up a little perl script that does it for people

16:58:33 <libby> maybe after we've made a bunch of decicisions

16:58:34 <sh1mmer> Well my grand foafstore project should be underway soon, plenty of stats then.

16:58:41 <danbri_dna> cool

16:58:49 <danbri_dna> .utc

16:58:59 <crschmidt> .t Z

16:59:02 <phenny> 2004-09-15T16:59:00Z

16:59:13 <danbri_dna> libby, can you make one of those urls for the time that get used on www-rdf-calendar?

16:59:18 <libby> I wouldn;t rely on mine; what the really only do is show the confusion around the issue

16:59:22 <danbri_dna> 5.30pm BST next week today?

16:59:37 <libby> yeah probly

16:59:48 <sh1mmer> surely

16:59:51 <danbri_dna> the confusion stemmed from our old 'only put it in the namespace after ppl have experimented with it' approach, but we never had a 'heres how you spell it' holding ground

16:59:55 <CaptSolo> libby - lastName not surname?

17:00:57 <CaptSolo> there was also plan to have IRCmeet for RDF desktop stuff

17:01:30 <CaptSolo> but that can also be week after next - no rush

17:01:59 <libby> there's allsorts of crap in there

17:02:03 <danbri_dna> ok. if folks want a chunk of agenda time, for next week, send mail to rdfweb-dev

17:02:07 <bengee> [[

17:02:07 <bengee> from swoogle:

17:02:13 <bengee> nick => populated by 12648 files as 933708 instances

17:02:21 <bengee> name => populated by 10135 files as 59994 instances

17:02:27 <bengee> firstName => populated by 1968 files as 3768 instances

17:02:32 <bengee> givenname => populated by 366 files as 501 instances

17:02:38 <bengee> family_name => populated by 376 files as 443 instances

17:02:44 <bengee> surname => populated by 1896 files as 3418 instances

17:02:45 <bengee> ]]

17:02:49 <crschmidt> [1 crschmidt@peanut ~]$ cat givenname | wc -l

17:02:49 <crschmidt> 51

17:02:57 <crschmidt> That's from julie's store

17:03:08 <libby> cool bengee, where'd you get that frpm?

17:03:10 <crschmidt> But swoogle data is probably much better :)

17:03:12 <sh1mmer> hey DanC_BRS

17:03:28 <bengee> swoogle.org

17:03:36 * DanC_BRS waves, gathers there's something organized going on here, looks around for an agenda or whatever

17:04:10 <danbri_dna> organised, hmm that's overly generous

17:04:22 <danbri_dna> I suggested a meetup time of now, during foaf galway

17:04:35 <sh1mmer> DanC_BRS winding up now.

17:04:38 <danbri_dna> but didn't do much/any followup prep beyond arriving back here in and not falling asleep

17:04:48 <danbri_dna> i'll send out an announce re next week same slot

17:05:03 <danbri_dna> couple actions to chase up, be good to get into a bit more of a ritual

17:05:06 <CaptSolo> hi DanC

17:05:36 <crschmidt> @@ Data from Julie on name/givenname/surname/etc. etc. etc.

17:05:52 <danbri_dna> so resolved, meet again 2004-09-22 17.30 BST for 1hr in #foaf?

17:06:06 <CaptSolo> agreed

17:06:39 <CaptSolo> (if most of people can make it|)

17:07:06 * bengee almost (tm) finished his term browser and a tool for multilingual documentation/term annotation. would like to discuss possible documentation items (e.g. doc, use case, example, issue, etc) next week

17:07:06 <danbri_dna> well, after next week we should prolly pick a different antisocial time

17:07:23 <danbri_dna> bengee, sounds good

17:07:53 <danbri_dna> can you circulate some b/g material on it before (>24hrs) the meeting?

17:07:54 <bengee> the FOAFCommunityProcess page is great. I'm trying to use it as a requirement doc

17:08:08 <DanC_BRS>http://www.timeanddate.com/worldclock/fixedtime.html?day=22&month=9&year=2004&hour=17&min=30&sec=0&p1=136

17:08:09 <danbri_dna> a world writable requirements doc ;)

17:08:15 <crschmidt> http://crschmidt.net/foaf/names is output of Running RDQL query: SELECT ?p ?pr ?o where (?p ?pr ?o) and ?pr =~ /name/

17:08:22 <DanC_BRS> Chicago *Wed 11:30 AM

17:08:34 <bengee> heh

17:09:03 <danbri_dna> thanks for the date url danc

17:09:30 * DanC_BRS learned a lot reading FoafCommunityProcess

17:09:34 <crschmidt> oops, that doesn't include firstName etc.

17:09:41 <danbri_dna> rdql, interesting

17:09:57 <danbri_dna> now if only the query language had count() and distinct() etc in it,... ;)

17:10:45 <danbri_dna> hmm it lets you regex against the ?pr, interesting. that's news to me.

17:11:17 <danbri_dna> the pics/filtering/labelling folks will find that useful, even though it feels a bit un-semwebby

17:11:42 <DanC_BRS> I'd rather and ?pr log:uri ?prstr and ?prstr =~ /name/

17:11:44 <crschmidt> okay, updated with second query

17:12:02 <crschmidt> Web interface is at http://crschmidt.net/rdfpython/python.cgi

17:12:07 <crschmidt> same data store as julie

17:12:08 * danbri_dna notes belatedly that we're ADJOURNED; thanks all for stopping by

17:12:18 <sh1mmer> heh

17:12:19 <danbri_dna> (...can of course carry on talking)

17:12:28 <libby> we'll let you ;)

17:12:30 <danbri_dna> +1 re pstr design

17:12:34 <sh1mmer> i have something that is related but more ontologicl

17:12:51 <crschmidt> please don't try and do anything that would break things, non-robust

17:12:54 <danbri_dna> though I was also wishing I could use OWL to populate a class with members whose URIs match a regex, and OWL can't do that.

17:13:03 <sh1mmer> we had a discussion a week or so back about i18n and literals as identifiers

17:13:30 <sh1mmer> we need something to exclude mbox_sha1sum from i18n

17:14:02 <DanC_BRS> loggerf, pointer?

17:14:02 <DanC_BRS> See http://www.ilrt.bris.ac.uk/discovery/chatlogs/foaf/2004-09-15#T17-14-02

17:14:04 <crschmidt> Or to identify that it is not a language-depenadnt string

17:14:08 <sh1mmer> right

17:14:11 <danbri_dna> exclude mbox_sha1sum from i18n? I don't understand, sh1m

17:14:21 <crschmidt> danbri_dna: sha1sum1

17:14:24 <danbri_dna> what is a language-dependent string?

17:14:25 <crschmidt> er,

17:14:32 <inkel> I've run with the same problem last week

17:14:37 <crschmidt> "sha1sum"@en is different than "sha1sum"@de

17:14:49 <danbri_dna> xml:lang is somewhat promiscuous

17:15:02 <sh1mmer> So <foaf:name lang="en">Tom Croucher</foaf:name> != <foaf:name lang="fr">Tom Croucher</foaf:name>

17:15:09 <sh1mmer> er xml:lang

17:15:12 <crschmidt> There should be a way to identify in the schema, "This property does not depend on the language it is in"

17:15:12 <sh1mmer> heh

17:15:14 * sh1mmer tired

17:15:24 <danbri_dna> one way would be to datatype it, cos xml:lang doesn't get inside them, but then we'd need datatype URIs everywhere

17:16:00 <danbri_dna> inkel, can you expand on your version of the problem?

17:16:23 <inkel> It was the same problem

17:16:44 <danbri_dna> this seems related to the desire to i18n-ise markup like <Person foaf:nick="danbri" foaf:gender="male"/> without requiring all controlled-vocab properties to take URIs as arguments

17:16:55 <inkel> I was playing with rdql and I've found that "sha1sum" != "sha1sum"@es

17:17:14 * DanC_BRS scrolls up... reads introductions... nifty

17:17:24 <danbri_dna> oh, test-case level headache. cool.

17:17:38 <sh1mmer> danbri_dna but we were saying that if you declair the lang of document then it shouldn't apply to the mbox_sha1sum because otherwise you couldn't identify the same person if two documents used different langs

17:17:40 * danbri_dna wonders how these could be recorded

17:17:50 <inkel> solution: I was using xml:lang as an rdf:RDF attribute, so every literal were in the same language

17:17:55 <sh1mmer> danbri_dna thinking in schemas

17:18:00 * bengee uses an annotation property (sys:langIndependent) as a work-around

17:18:13 <sh1mmer> inkel well some things shouldn't be equivelent

17:18:14 <inkel> so I remove xml:lang from rdf:RDF and use it wherever I need it

17:18:20 <danbri_dna> inkel, sh1m, could at least one of your record this in http://rdfweb.org/topic/IssueTracker if it ain't already there?

17:18:25 <sh1mmer> inkel it's only special cases

17:18:32 <inkel> Yes, off course

17:18:44 <sh1mmer> danbri_dna can do.

17:18:49 <danbri_dna> can xml:lang be turned off? eg. xml:lang="" ?

17:18:51 <danbri_dna> ta

17:19:00 <sh1mmer> um... that I don't know

17:19:07 <sh1mmer> I just worry about all the existing data out there

17:19:22 <DanC_BRS> yes, xml:lang="" turns it off. it might be new in xml1.1

17:19:46 <sh1mmer> I also think it would be useful generally to be able to declair props as exlcuded from i18n equivelency

17:19:51 <DanC_BRS> the future's longer than the past, sh1mmer

17:19:51 <danbri_dna> bengee, you've run into this as well. hmm Something Must Be Done. Noting your experience in the wiki is a start.

17:19:58 <sh1mmer> danbri_dna which group should that go to?

17:20:06 <sh1mmer> DanC_BRS agreed.

17:20:23 <sh1mmer> DanC_BRS but it amuses me to use the argument dan used on me against him :P

17:20:27 <inkel> From XML 1.0 spec: "In particular, the empty value of xml:lang is used on an element B to override a specification of xml:lang on an enclosing element"

17:20:39 <danbri_dna> I think there's a lot more identity reasoning stuff we could do than is licensed by OWL semantics, and identity in face of similar but non-identical literals (and URIs) is definitely one

17:21:56 * DanC_BRS noodles on feedback loops... what will get people to fix/change their data... returns to foaf-auth/single-sign-on idea-space...

17:22:11 <sh1mmer> danbri_dna well if emails are resources, and mbox_sha1sum is a transformation of an email what about making it a resource

17:22:18 * sh1mmer wonders if that make _any_ sense

17:22:19 <DanC_BRS> my intuition says foaf files can be part of a powerful solution to the blog-comment-spam problem.

17:22:32 <libby> yeah, that would rule

17:22:34 <danbri_dna> OK so we could say in mbox_sha1sum spec that (a) the canonical form of the literal should be without language tagging, to aid data merging using apps, APIs and QLs that don't understand that (b) for any ?x and ?y where ?x has a mboxsha1sum of ?a and ?z has an mboxsha1sum of ?b and ?a and ?b differ only in their language tag, ?x = ?y ...

17:23:02 <sh1mmer> DanC_BRS that would be nice.

17:23:08 <DanC_BRS> you can _almost_ write that as an N3 rule, danbri...

17:23:17 <DanC_BRS> ... we just need a built-in to relate "foo"@en to "foo"

17:23:34 <danbri_dna> Yep, N3 can't inspect literal substructure...

17:23:59 <DanC_BRS> well, no more or less than it can peek into URIs with log:uri magic

17:24:03 <danbri_dna> it hadn't occured to me that ppl would actually be running into probs 'cos RDQL/Jena does the literal identity stuff right

17:24:33 <sh1mmer> danbri_dna it just seems good housekeeping to define it explicitly.

17:24:40 <danbri_dna> ...so yeah I guess foaf:mbox_sha1sum underspecified w.r.t. its intended purpose of aiding data merging

17:25:05 <DanC_BRS> have you guys seen the javacript password generator? I wonder if we could mix that with foaf somehow...

17:25:24 <sh1mmer> DanC_BRS don't follow...

17:25:25 <danbri_dna> rings a bell

17:25:43 <danbri_dna> oh, i didn't write up my single signon idea

17:26:25 <danbri_dna> which is basically to use a pgp thingie on desktop and have sites send account info pgp encrypted to an IMAP-accessible store, so clients could retrieve all passwds via IMAP and populate browser wallet locally

17:26:31 <DanC_BRS> Password generator bookmarklet http://angel.net/~nic/passwdlet.html

17:26:33 <danbri_dna> still needs a certain amount of pgp deployment

17:27:44 <DanC_BRS> that bookmarklet basically implements digest auth in a bookmarklet, except without the nonce. sigh... but...

17:27:53 <DanC_BRS> ... I wonder about mixing it with foaf somehow...

17:27:54 <sh1mmer> danbri_dna did you read the um... remembers name... um... the dick fella

17:28:05 <sh1mmer> danbri_dna the writeup from that session

17:28:10 <danbri_dna> sx1p

17:28:15 <sh1mmer> right

17:28:18 <sh1mmer> sxips

17:28:23 <danbri_dna> nope, i got a workshop report to write tho!

17:28:37 <sh1mmer> danbri_dna that handled single signin

17:28:43 <DanC_BRS> so when I submit a comment, I push a bookmarklet and it somehow provides credentials that I'm the owner of the foaf file.

17:29:38 <danbri_dna> there were some experiments in that area around...

17:29:44 <sh1mmer> interesting

17:29:56 <danbri_dna> sx1p has a small highly trusted company in the loop, making a little money off everyone

17:29:58 <sh1mmer> i have a lot of stuff i _really_ need to write up for foaf trust

17:30:05 <DanC_BRS> let's see... if the blog comment page had a nonce/challenge, and the bookmarklet had the secret key corresponding to a public key in the foaf file, it could sign the nonce, and the blog comment host could verify it. that's secure, but pretty heavy.

17:30:37 <danbri_dna> are you prepared to assume browser cookies?

17:30:51 <DanC_BRS> at one point, netscape supported signed form submission. does anybody know if that survived the mozilla rewrite?

17:31:13 <DanC_BRS> I think a cookie-based solution would be acceptable, but what I just cooked up doesn't use cookies.

17:32:23 <DanC_BRS> and then there's the really quick-n-dirty solution: the blog comment form has a challenge/nonce, and you prove that you own the foaf file by writing the challenge into the foaf file. messy, but simple.

17:32:36 <sh1mmer> right im off for my dinner

17:32:39 <sh1mmer> ttfn

17:32:40 <sh1mmer> back later

17:32:50 * sh1mmer try and write up trust issues tonight

17:32:54 <DanC_BRS> hmm... neither of these solutions take advantage of the social network, advogato-style.

17:33:37 <DanC_BRS> hmm... what about something that proves to the blog comment host that "either I'm who I say I am or these 20 people are lying"

17:33:59 * danbri_dna drops a brand new 200g disk 20 cms onto floor

17:34:10 <danbri_dna> it landed with a clunk but doesn't rattle

17:34:17 * danbri_dna pretends nothing happened

17:34:20 <DanC_BRS> get some sleep!

17:34:40 * danbri_dna gets offline for a bit anyway

17:34:56 <sh1mmer> DanC_BRS ask me tonight and ill take you through what I have.

17:35:00 <sh1mmer> It could be used for that.

17:35:10 * sh1mmer off

19:42:21 <pjz> danbri_dna: oh, re:single signon, I independently came up with something similar, except that you automate it with an agent and a password. The password is to *your agent* though.

19:44:51 * pjz means to implement this at some point to allow tiered access to foaf data.

20:52:53 <mcroydon> mcroydon is now known as Netminder

22:09:15 <edd> coroks