Friend of a Friend (FOAF) IRC Chat Logs for 2003-06-24

00:07:43 <lilo> [Global Notice] Hi all. Current status: the main rotation is currently pointed to 127.0.0.1, and one of the MR servers is down, two have been shut down since the problem started

00:08:13 <lilo> [Global Notice] We're currently trying to get a bit ahead on the klines, but it's slow going

00:08:51 <lilo> [Global Notice] It's just the usual "summer in the US" nonsense....apologies for the inconvenience, and we'll keep things going as best we can. Thanks.

00:51:33 <lilo> [Global Notice] So, with the network pointing to 127.0.0.1, and us churning out klines, maybe this is a good time to point out that we need a new protocol. :)

03:35:19 <bitsko> foaf:mbox needs to be 'mailto:a@b.c' not 'a@b.c', correct?

03:38:09 <ericP> i believe so

03:38:34 <ericP> are you foafing yourself up?

03:40:21 <bitsko> even better, I've nearly completed a python utility for accepting a foaf uri as the sole field for a comments entry (or other web form), verifying its signature, and summarizing its contents to the caller

03:40:41 <bitsko> I'm testing it as a blosxom plugin

03:40:54 <bitsko> or rather, in the blosxom writeback plugin

03:41:23 <ericP> cool!

03:42:51 * ericP needs to sign http://www.w3.org/People/Eric/ericP-foaf.rdf

03:43:19 <ericP> and need to find a picture of myself, rather than a picture of someone with a picture of someone that may be some part of me

03:43:57 <ericP> how do you sign your foaf?

03:44:39 <bitsko> http://usefulinc.com/foaf/signingFoafFiles

03:45:11 <bitsko> also, add a <rdf:seeAlso rdf:resource="" /> to "your" foaf:Person resource

03:45:27 <ericP> tx

03:45:40 <bitsko> the latter is for my tool to find you ;)

03:47:06 * ericP gets some gear together

03:47:14 <ericP> i'm working on a friends mac

03:47:59 <ericP> got fink and X working with .fvwm but really struggling with the lack of mouse buttons

03:48:14 <bitsko> hehehe, many foafs include a thumbnail photo of the user, which one can include next to their comments!

03:48:44 <bitsko> I used X on a mac for a long time. got used to control and/or command ;)

03:49:36 <ericP> yeah, i poked around image.google.com for a thumbnail of myself but it's tough looking for a name with quote chars in it

03:50:02 <ericP> you used OSX? with X?

03:50:11 <bitsko> did you try quoting around it? google's pretty good about that

03:50:54 <bitsko> no, Tenon MachX

03:50:54 <bitsko> pre-OSX

03:51:05 <bitsko> I have touched my mac in months :(

03:51:30 <bitsko> woot!! success! houston, we are FOAF enabled. I repeat, FOAF enabled!

03:51:50 <ericP> becoming conceared with re-entry...

03:51:59 <ericP> s/conceared/concearned/

04:00:19 <ericP> what's the difference between img and depiction?

04:01:03 <ericP> seeAlso ref added

04:05:17 * ericP heads oot

04:48:07 <bitsko>http://bitsko.slc.ut.us/blog/new-blog.writeback

04:48:31 <bitsko> apparently no chump ;)

04:49:05 <bitsko> that's ok. when people wake up tomorrow, probably before I make the official posting, there's a FOAF-enabled weblog comments

04:50:54 <bitsko> the "Homepage" field, with the FOAF-icon to the right, can take the URI of a FOAF instance. it'll download the foaf, download the wot:assurance signature identified within, verify the signature, then use the contents of the FOAF instance to fill in all the remaining data (plus some, hehe)

04:51:42 <bitsko> to use it, your foaf must be signed per http://usefulinc.com/foaf/signingFoafFiles (nothing new for most folks using foafbot)

04:53:08 <bitsko> but *also*, the foaf:Person this FOAF instance describes *must* include an rdf:seeAlso with the same URI entered in the form field

04:53:19 <bitsko> this can be done with <rdf:seeAlso rdf:resource="" />

04:55:17 <bitsko> error reporting isn't working, so right now it's pass/fail. if it works, the page will come back with yours (and others) comments, if not, it'll come back with an "empty" page (no comments) but also no indication of error :(

04:58:41 <bitsko> future developments may include: using more of the FOAF info, like thumbnail; using challenge/response to authenticate the foaf owner (anybody can post a FOAF url); finishing the simple "World" RDF-nodes class; clean up.

05:06:00 <nym> hi

06:36:05 <nymble> woo

06:36:15 <nymble> contacted the guy behind mtfoaf

06:36:22 <nymble> he's going to do a rewrite with me i think

06:37:15 <eikeon> Nice

06:37:27 <nymble> yeah

06:37:33 <nymble> i'm very excited

06:37:42 <nymble> we can do some nice tweaks

06:37:56 <nymble> i plan to publish a movable type package with this

06:39:19 <eikeon> What lang are movable type packages written in?

06:43:12 * eikeon is calling it a night... later.

08:37:42 <sbp`-> sbp`- is now known as sbp`

10:10:57 <danbri> <bitsko> even better, I've nearly completed a python utility for accepting a foaf uri as the sole field for a comments entry (or other web form), verifying its signature, and summarizing its contents to the caller

10:11:00 <danbri> this sounds good :)

10:11:43 <danbri> I _think_ the sixapart utility linked from http://www.sixapart.com/log/2003/01/fun_with_foaf.shtml ie http://www.sixapart.com/log/mt-foaf.txt doe similar in Perl

10:12:17 <danbri> It would be good if people didn't need to know their FOAF URL, ie. if it could take foaf:homepage (or other uniquely identifying field?) as its argument...

10:12:46 <danbri> ...and sniff the homepage for a link-rel to the real foaf file

10:13:03 <JibberJim> or for an email address to sha1ify

10:13:17 <darobin> FOAF Discovery

10:25:24 <zool> mornin

10:25:40 <mattb> heya zool

10:29:44 <zool> hi engel, gregor

11:46:07 <danbri> http://www.kieranshaw.co.uk/projects/proposal.php [[Research proposal - Summary

11:46:07 <danbri> "Using FOAF/RDF data and improved information publication about relationships between people and their projects, courses and research to enable a better learning and co-operation environment at the University of Warwick"]]

11:46:16 * danbri wonders how kieran is getting on

11:46:58 * JibberJim was wondering that yesterday

11:47:19 <JibberJim> Would the MSc have started yet though?

11:47:32 <danbri> I think it is 'by research' rathe than taught, if i recal

11:49:38 * danbri drops him a quick email

12:09:55 <JibberJim> zool?

12:40:48 <bitsko> danbri: FOAF discovery +1

12:42:45 <bitsko> in a way, though, people would still need to "know" that they're home page had that information. entering only the one field means the user knowing "there's this foaf thing that lets me just do this"

12:43:46 <danbri> yup. i think either should work...

12:43:52 <danbri> ppl more likely to remember a homepage tho

12:45:02 <JibberJim> how are you identifying the person in the foaf doc, is the one you're interested in?

12:46:17 <bitsko> danbri: ah, yes, I recall reading that MT page.

12:46:49 <bitsko> JibberJim: the foaf:Person associated with the FOAF URI must have a rdf:seeAlso with the same FOAF URI

12:47:14 <bitsko> whicha can be done using <rdf:seeAlso rdf:resource="" />

12:47:16 <JibberJim> hmm, that would be a little odd though - have you considered using dc:creator?

12:47:33 <bitsko> how so?

12:48:25 <JibberJim> well if the rdf doc says the dc:creator is Person with the name "Fred Bloggs", then you can take the Fred Bloggs person as it being the "foaf file for"

12:48:40 <JibberJim> (not that I ever like the idea of people having a foaf file)

12:49:21 <danbri> http://rdfweb.org/mt/foaflog/archives/000031.html (wrote up the Spring FOAF stuff)

12:49:26 * danbri wonders if mortenf is around

12:50:12 <bitsko> ala:

12:50:13 <bitsko> <rdf:Description rdf:about="">

12:50:13 <bitsko> <wot:assurance rdf:resource="foaf.rdf.asc" />

12:50:13 <bitsko> <dc:creator>

12:50:13 <bitsko> <foaf:Person>

12:50:13 <bitsko> <foaf:mbox_sha1sum>bb9fe68ffa844717c1e8752d42ca589ba2044719</foaf:mbox_sha1sum>

12:50:15 <bitsko> </foaf:Person>

12:50:17 <bitsko> </dc:creator>

12:50:19 <bitsko> </rdf:Description>

12:51:48 <bitsko> iow, a signed FOAF file must have an rdf:Description rdf:about="" to "sign" the "current instance", to which we can add a dc:creator that references the foaf:Person that signed this file

12:52:46 <JibberJim> yeah, but I don't think we have a way of associating a file yet with a person.

12:52:59 <bitsko> when comparing RDF nodes for "equality" in this way, without a subject URI, every property must match (logical AND), correct?

12:53:25 <bitsko> see also http://usefulinc.com/foaf/signingFoafFiles

12:53:34 <JibberJim> yes, that doesn't solve it.

12:53:56 <JibberJim> I could sign something about someone else - it wouldn't make it a foaf doc about me.

12:54:35 <bitsko> nix, authentication and signing are two seperate issues

12:54:55 <JibberJim> my point's not about authentication, it's about this idea that people have foaf files.

12:55:14 <bitsko> k

12:55:43 <JibberJim> the signing doesn't solve this, it only solves who said what.

12:55:56 <bitsko> that, too, may be a seperate issue

12:56:28 <bitsko> lets start at the beginning

12:56:32 <danbri> there were some long threads on this on rdfweb-dev a few months back, but i never boiled them down into an answer...

12:56:38 * danbri thinks foaf:made is way to go

12:57:41 * JibberJim doesn't think foaf:made solves it.

12:59:05 <bitsko> what I need is a way to bootstrap getting information about a particular foaf:Person. logically, the "best" way to do that, assuming for just a moment that it's the "real" person providing the info, is to accept a foaf:Person rdf:Description foaf:mbox[_sha1sum] that I then scan the world and retrieve the entire rdf:Description

12:59:46 <JibberJim> yes, scanning the world is fine, then you get all the info about them.

12:59:58 <JibberJim> but scanning the world is too expensive in your use I think.

13:00:10 <bitsko> given the URI of a foaf instance that also includes the partial foaf:Person would save me scanning the world

13:00:59 <JibberJim> as long as you've got the person, yes that would be fine - then though you need to get two pieces of information from the person before you can do anything - email address, and uri

13:02:07 <bitsko> if I understand your point, it's that any particular instance of rdf:RDF may include rdf:Descriptions of lots of various foaf:Persons, so it's bad design to assume a particular foaf instance "belongs to" or "describes" a single individual, correct?

13:02:36 <JibberJim> yes, that's my point, and said so much better than I ever did!

13:04:02 <bitsko> k. I'm not sure how to get around that, without resorting to bookmarklets or smart agents that can provide a *lot* more pasteability than a single URI. particularly because we haven't touched on authentication, yet

13:05:37 <JibberJim> I'm wondering if there is a way of doing it from within the RDF though.

13:05:39 <bitsko> or maybe I'm being too lazy (for me or the users?), and maybe we *should* be pasting in rdf:Descriptions of foaf:Persons

13:06:16 <bitsko> and *only* require the user to enter a URI of a foaf instance? or are you thinking something else?

13:07:19 <JibberJim> yes, just the URI would be the best way.

13:07:34 <JibberJim> The option may be to have a crawled db, and extract the data about the person from that.

13:10:04 <bitsko> yes, but how do you pick that particular person, given just a URI?

13:10:46 <JibberJim> there'd need to be something in the foaf: maybe a new unambigous property foaf:foafDoc="..."

13:11:04 <bitsko> or <rdf:seeAlso rdf:resource="" />?

13:13:16 <JibberJim> nah, because rdf:seeAlso already has semantics which aren't enough for us.

13:15:05 <bitsko> how so? and wouldn't rdf:seeAlso in a partial foaf:Person be the link to the more full description as well?

13:15:30 <JibberJim> yes, but it could also link to another document which talks about hundreds of people.

13:16:03 <JibberJim> for example the W3 may maintain a document about all their staff, danbri might then say rdf:seeAlso that document so he doesn't need to maintain his office phonenumber etc.

13:16:17 <bitsko> true, but in this case I need only match the triple of rdf:seeAlso <URI> of the <URI> *I* already have in my possession :)

13:17:00 <bitsko> yes. we back abuot 10 msgs, <JibberJim> I'm wondering if there is a way of doing it from within the RDF though.

13:17:19 <JibberJim> yes, but you're forcing us to use some magic syntax in the RDF which doesn't make sense outside of your application.

13:17:55 <JibberJim> it's a worthless triple, which you're giving extra importance.

13:18:26 <bitsko> I may have an idea, though: we do need authentication, and performing a challenge/response check against each foaf:Person, while a little costly, would only give one positive match.

13:21:28 <bitsko> the way that "other" applications would do something similar would also be to have two pieces of information: the URI of an RDF instance known to contain the data you need, and an identifier (resource URI or partial rdf:Description)

13:22:06 <bitsko> if other applications wanted to "only" use one piece of info (a URI), they'd be facing the same issue (and potential solution)

13:23:32 <bitsko> "known to contain" can include rdf:seeAlso of further information to include, I wouldn't mind seeing a paper on the ramifications of that, though

13:23:32 <JibberJim> maybe we should try out your solution on the mailing list, I don't like it as it seems to require a triple that has no reason to exist, and you're giving it a meaning that it doesn't itself carry.

13:24:33 <bitsko> k. I will post an announcement to the list and include the potential solutions

13:25:11 <JibberJim> cool

13:25:38 <bitsko> note that the final answer may be URI+auth, because auth is necessary regardless (otherwise you can paste anyone's foaf URI)

13:25:58 <JibberJim> anyone can type in other peoples info though...

13:26:19 <bitsko> maybe I should just continue to drive toward that solution and keep these hacks as an interim

13:26:43 <bitsko> "auth" means authenticating that "you are who you say you are"

13:27:17 <bitsko> unfortunately, that *will* require a client side tool :(

13:27:45 <JibberJim> Yes, but I was meaning if you're doing this as a quick way of filling in forms for weblogs etc. then auth doesn't exist in the current methods, does it need to in the new?

13:27:45 * bitsko wonders if pgp/gpg has some magic already designed for that, hmmm

13:27:58 <JibberJim> yes, keep implementing!

13:28:30 <danbri> http://rdfweb.org/pipermail/rdfweb-dev/2003-June/011267.html "[rdfweb-dev] XSLTs for FOAF, Spring v1.3.1 and plans for FOAF spec improvements"

13:28:35 <danbri> any thoughts re the xslt thing?

13:28:39 <bitsko> yes, the comment forgeries are increasing

13:28:59 <bitsko> re. do we need auth

13:29:51 <bitsko> with FOAF+auth, a signed comment can include a browser-like indicator of "secure" or "verified"

13:31:30 <bitsko> heh, also increasing are the number of instances where people are missing (usually by hitting the back button) various optional fields of info. reducing the required fields for FOAF-users to just two (FOAF+challenge-response), that may solve that problem too ;)

13:31:53 <JibberJim> how are you going to do the challenge/response?

13:33:03 <bitsko> my proposal for UI is for the FOAF-icon to be on the "Homepage" field, with a N-digit challenge next to it. one copies the challenge to a client-side tool and pastes the response into the form's name field

13:33:40 <bitsko> N ~= 10 alphanumeric, so it's double-clickable

13:34:31 <bitsko> you can see the foaf-icon portion, without the challenge, at http://bitsko.slc.ut.us/blog/new-blog.writeback

13:36:22 <bitsko> the tool that checks the foaf also can create the challenge

13:38:15 <danbri> re challenge/response, i had some loose ideas on using FOAF to store crypted passwords, http://rdfweb.org/mt/foaflog/archives/000005.html

13:38:42 <danbri> would the Javascript sha1 thing fit in a bookmarklet?

13:39:15 <JibberJim> only a remote bookmarklet.

13:39:16 <danbri> My foaf file coudl say: danbri---hashedWebLogCommentPassword-->"sdfsdfsetawerdhbe"

13:39:44 <danbri> and weblog sites could test me (by clientside js) to see if i knew it. hmm no that don't really work does it.

13:39:49 <JibberJim> no.

13:40:02 <danbri> but server side no work either, cos i don't trust them with my password.

13:40:10 <JibberJim> exactly...

13:40:18 <danbri> or maybe i do, as its only my weblog commenting password

13:40:27 <danbri> nah

13:43:48 <JibberJim> would you trust a validation service with it?

13:44:32 <JibberJim> so you send it to a 3rd party for validation - meaning you can comment on all weblogs whilst only trusting 1 person.

13:45:00 * danbri ponders http://tbray.org/ongoing/When/200x/2003/06/23/SamsPie

13:45:05 <deusx_away> deusx_away is now known as deusx

13:47:18 <bitsko> +1 on SamsPie

13:47:39 <bitsko> er, Sam's Pie, not to tbray's article in whole ;)

13:47:41 <danbri> its tricky to know where to take things that should be lightweight, etc.

13:48:05 <danbri> or what to model ones process, docs on etc if you just do things as a bunch of internet collaborators

13:48:16 <danbri> (rss-dev; or closer to home, FOAF, ...)

13:48:21 <bitsko> I think Sam has good balance there. I think a first possibility is that there will be a "core" of elements and a set of standard extensions, all in namespaces.

13:48:48 <danbri> With RSS, the killer question is scope. WTF is RSS *for*? Who are the stakeholders?

13:48:55 <bitsko> I actually kinda like the SAX model, but we don't have a driver that everyone could get behind

13:48:58 <danbri> That imho is what will drag down any attempt at a formal WG

13:49:02 * danbri nods

13:49:13 <danbri> Sam seems energetic and respected

13:49:16 * danbri doesn't know him

13:49:33 <bitsko> he also apparently is being funded full-time to work on it, as tbray mentions

13:49:49 <danbri> Rael was too, before he got chewed up by the RSS-DEV WG experience and RSS wars... I wouldn't wish chairing RSS on anyone right now

13:50:36 <bitsko> Sam's got severl fewer battles to fight and far less "ownership" in RSS thanthe rss-dev authors ever did

13:51:52 <bitsko> Sam's very open minded and tries everything. he favors SOAP, in a RESTful way, but doesn't push it

13:54:26 <bitsko> re. challenge/response, the benefit there is that there's no passwords anywhere but the client machine. standard public-key stuff

13:55:01 <bitsko> are you familiar with the steps?

13:57:52 <danbri> yes, though its not fresh in my mind (clearly!)

13:58:26 * danbri adds Spring article to http://rdfweb.org/topic/FoafProject and also a 'dating' section

13:59:57 <zool> re, briefly

14:00:07 <mattb> briefre zool

14:01:18 <danbri> bitsko, re announcing your tool, has sbp released his code yet?

14:02:36 <bitsko> no! but I may just refer to a "parser that provides triples". I started down that road a couple of weeks ago, but forget the command-line utility that generates triples. in rdflib or similar

14:02:59 <danbri> 'rapper' is dave's one, part of raptor (redland's parser)

14:03:03 <bitsko> that's it

14:03:07 <sbp`> whoops, hi. sorry. I'm only in for a few minutes--what's up?

14:03:09 <danbri> by triples, you mean N-Triples

14:03:32 <danbri> bitsko's using your parser, made nice thing, can you ship it so we can have a bundled version of his thing?

14:03:41 <danbri> <- precis

14:03:45 <danbri> hi btw :)

14:03:50 <bitsko> sbp`: I'll be posting my FoafCheck here shortly. wondering your status, or if I should just say I need "a parser that provides triples", ala rapper

14:03:58 <sbp`> heh. yes, of course. he can just go ahead and release it

14:04:03 * bitsko what danbri said

14:04:07 <danbri> any particular license?

14:04:15 <sbp`> nope, just include it in your distro. GPL 2, share and enjoy, as usual

14:04:22 <bitsko> file says GPL2, and I'm good with that

14:04:29 * zool thinking about small ntriples->rdfxml serialiser in perl

14:04:59 <bitsko> sbp`: I did want to chat about the triples representation. re-parsing them seems more difficult maybe than it should be

14:05:03 * JibberJim wonders about a FOAF licence - a licence to all your foaf's to do what they want with it, but anyone else has different conditions.

14:05:20 <zool> heh

14:05:24 <bitsko> JibberJim: Creative Commons

14:05:36 <JibberJim> so you can BSD it to your FOAF, but GPL it to the wider world.

14:05:36 <sbp`> yup. my apologies, but I'm extraordinarily busy today--could you possibly drop me an email, and I'll get on it tonight? or you can ping me on IRC later, perhaps

14:05:46 <dajobe> zool: I've got something like that you could probably hack

14:05:47 <danbri> cute but unimplementable; we don't what a FOAF is...

14:05:51 <bitsko> sbp`: yup, no problem. may even include a patch...

14:06:10 <sbp`> heh, great! thank you. sorry for ducking out so quickly

14:06:16 <dajobe> zool start with http://cvs.ilrt.org/cvsweb/redland/librdf/perl/serialize.pl and chop out the bits you don't need

14:06:16 <sbp`> ttfn

14:06:22 <zool> dajobe, sounds good... i have almost-that kicking around, but would like to see comparisons

14:06:48 <dajobe> note lots of that file is commented out, isn't used

14:06:58 <bitsko> sbp`: do you have a URL for the parser yet?

14:07:09 <zool> thx dajobe

14:27:30 <sbp`> bitsko: I'll be using http://infomesh.net/2003/rdfparser/ I've put a few words up there for now saying it'll be released later on

14:29:16 <bitsko> great, thx!

14:32:43 <dajobe> what's the diff between yours and rdflib?

14:34:25 <bitsko> it's a 10k standalone Py parser that produces triples

14:34:51 <dajobe> including an XML parser?

14:35:04 <bitsko> no, using Py's built-in xml parser

14:35:11 <dajobe> ok, not standalone then

14:35:25 <bitsko> ??

14:37:21 <bitsko> I'm not sure I understand how to read that. it doesn't implement its own urllib, either

14:38:02 * eikeon looks to see how big rdflib is... as it sits on top of Py's stdlib as well.

14:38:08 <dajobe> so what's the difference between sean's and eikeons?

14:38:28 <eikeon> One file vs. several? And a bit narrow focus?

14:38:35 <dajobe> I suggest you jointly test against each other :)

14:38:43 <dajobe> (I use arp to check my parser)

14:40:50 <bitsko> I'll do that.

14:49:38 * JibberJim can't fix foafnaut...

15:10:06 <ephidrina> ephidrina is now known as iphedra

15:10:13 <iphedra> iphedra is now known as ephidrina

15:42:39 <JibberJim> Hi iwai

15:46:50 <danbri> foafbot, picture of Jim Ley?

15:46:53 <foafbot> Picture at http://swordfish.rdfweb.org/photos/2002/06/24/002250.JPG -- according to Anon179

15:46:53 <foafbot> Caption: Jim, Dave, Nadia in the Watershed pub.

15:46:53 <foafbot> Picture at http://swordfish.rdfweb.org/photos/2002/06/24/002256.JPG -- according to Anon178

15:46:53 <foafbot> Caption: Pub Scene

15:46:53 <foafbot> Picture at http://jibbering.com/dinner/Jbeer.jpg -- according to Jim Ley

15:46:53 <foafbot> Caption: Jim, in a DJ, drinking Beer with lights behind him.

15:46:58 <danbri> workie :)

15:57:54 * darobin infers that Jim Ley can only be found in pubs

15:59:24 <JibberJim> the last one was in a castle.

16:52:13 * danbri tarts up rdfweb.org a little more

16:52:55 * JibberJim wants to give up on foafnaut totally.

16:53:15 <danbri> aww

16:53:23 <danbri> after version ?.0...

16:53:33 <danbri> what's driving you nuts most?

16:53:39 <JibberJim> all of it.

16:53:45 <danbri> ASV headaches?

16:53:53 <JibberJim> Nah Batik ones today.

16:54:07 <JibberJim> Norm Walsh emailed "I don't have ASV so was looking in Batik..."

16:54:47 <danbri> Fancy kicking off the foafnaut-dev list with a status report of what more or less works, key urls etc?

16:54:49 <JibberJim> but the lines problem in template version of ASV has also had 2 hours of no improvements on it.

16:55:02 <danbri> :(

16:55:12 <libby> aw poor jim

16:55:26 <libby> do something else for a bit

16:55:28 <JibberJim> batik does the nice fonts too though!

17:03:29 <bitsko>http://bitsko.slc.ut.us/blog/2003/06/24/foaf-check

17:11:42 <mortenf> hey there.

17:11:46 <JibberJim> hey

17:12:17 <mortenf> anything we can do re foafnaut, besides buying your beer?

17:13:04 <JibberJim> Agree to it being buried!

17:13:15 <mortenf> negative.

17:13:33 <mortenf> isn't svg open coming up?

17:13:37 <danbri> sorry jim, we can't do that.

17:14:01 <danbri> beer deaden the pain.

17:14:16 <danbri> maybe a lovely new ASV will be out soon and it'll all be easier...

17:14:22 <danbri> and have embedded smil videos...

17:14:57 <JibberJim> no, because we can't just do it on one viewer, got to support other UA's

17:15:40 <danbri> I don't mean use proprietary hacks, but to have it really slick on ASV would motivate others to implement SVG...

17:15:47 <danbri> it's imho one of the coolest svg apps i've seen

17:16:01 <danbri> but then i thought javascript prolog was cool; tastes vary ;)

17:16:35 <mortenf> ken, fwiw i use *[local-name()='Person' or foaf:mbox or @foaf:mbox or foaf:mbox_sha1sum or @foaf:mbox_sha1sum][1] in the FoaF Explorer for identifying "the" person in a file.

17:17:12 <danbri> how do you know it's 'the' person of the file, its owner or whatever, not just someone else also mentioned there?

17:17:30 * danbri can feel a bugzilla entry looming

17:17:37 <mortenf> i don't...

17:17:47 <danbri> you pick the first?

17:17:53 <mortenf> yep.

17:17:54 <danbri> that comes back from the match?

17:18:18 <mortenf> a resource that's describing a person.

17:19:06 <mortenf> hmm, foaf-check doesn't seem to like me (or my foaf)...

17:32:13 <bitsko> mortenf: see the discussion starting http://ilrt.org/discovery/chatlogs/foaf/2003-06-24.html#T12-45-02

17:32:43 <bitsko> plus "lets start at the beginning" at http://ilrt.org/discovery/chatlogs/foaf/2003-06-24.html#T12-56-28

17:33:02 <mortenf> yeah, i read that.

17:34:09 <bitsko> ah. ok, in that context, I don't see how you're saying FoaF Exp identifies "the" person

17:34:11 <mortenf> one more thing: you could, like most people have already done, add a dc:creator to the document resource, pointing to the person in the file.

17:34:24 <mortenf> the "'s are important :)

17:34:43 <mortenf> all i'm saying is: that's how *I* guess :)

17:34:47 <bitsko> ah!

17:35:12 <bitsko> the anti-example is when one instance has multiple fully described foaf:Persons

17:35:23 <mortenf> with the creator, you have a valid statement, and won't need the seeAlso.

17:35:30 <mortenf> yep, that's a problem.

17:36:45 <mortenf> that's why xslt is actually useful here, for document order (although it won't help for a real group of people).

17:48:38 <golbeck_> golbeck_ is now known as golbeck

17:51:45 * JibberJim just realised that svgopen has some weird tracks. foafnaut is in the "visualisations" track, but "visualising DB management with SVG" is in the Semantics and Metadata track.

18:22:49 <danbri> eikeon, yr foaf webviewer is screenshottd in http://anoh.s10.xrea.com/03-06.php#d2003-06-20c1

18:24:27 <mortenf> picture, thousand words, go figure...

18:25:00 <danbri> i only undertsand the pictures

18:25:10 <danbri> http://antipop.zapto.org/wiki/?The%20%27friend%20of%20a%20friend%27%20project%20-%20FOAF is filling up

19:12:52 * danbri reorganises http://rdfweb.org/topic/FoafProject somewhat

19:50:50 <danbri> evening edd!

19:50:58 <eikeon> danbri: Neat. I love seeing the characters. Almost make all the fighing with unicode seem worth it.

19:50:59 <edd> hi

19:51:10 <edd> just reading some (rather dense) stuff on WS-Trust

19:51:21 <edd> these web service guys have really gone deep into security, trust etc.

19:51:26 <edd> i wonder what we could plunder

19:51:28 <danbri> WS being Web Serv... ah yes

19:51:49 <danbri> are any of them trusting real $$$ to non-closed-world web services?

19:52:05 <edd> it's at the spec stage largely

19:52:15 <edd> WS-Trust seems almost a kerberos-a-like

19:52:25 <danbri> I'm all for re-use. Was very happy to find GPG met 80%+ of my rdf signing/encryption needs

19:52:29 <edd> but this stuff is so heavy i'm getting a headache :)

19:53:14 <edd> talking about trust brokers, essentially

19:53:28 <edd> so agents can communicate without being able to understand all the trust mechanisms in the world

19:54:07 <edd> i suppose it's a bit anti-web though

19:55:19 <deusx> deusx is now known as deusx_away

19:56:03 <edd> i do suspect that some of WS-* is a bunch of consultants making sure their industry survives into the next generation

19:56:09 <danbri> if it creates a useful market, maybe it'll get traction

19:56:24 * danbri was thinking 'hmm, corba brokers' there for a second

19:56:57 <edd> corba works for me, as part of gnome.

19:57:31 <edd> back to ws-trust. intermediate trusted brokers doesn't sound so far away from the semweb dream

19:58:15 * danbri nods

19:58:21 <danbri> did you see Ken's thing earlier?

19:58:33 * danbri has been updating http://rdfweb.org/topic/FoafProject some more

19:58:38 <edd> this is the first time i'm on today

20:00:15 <deusx_away> deusx_away is now known as deusx

20:04:32 <deusx> deusx is now known as deusx_away

20:08:40 <danbri> edd, re ken's doc see http://rdfweb.org/topic/FoafIdentityAssurance

20:13:30 <deusx_away> deusx_away is now known as deusx

20:16:11 <edd> danbri: inneresting

20:17:06 <danbri> I'm trying to write foaf new stuff up in weblog as it happens; backlogs never get dealt with

20:17:17 <edd> that's the truth

20:17:43 <danbri> ...backlogs get wiki entry if lucky, i figure a wiki entry (link or sentence) quicker than weblog, since there is more of a sense of 'i can fix it later'

20:17:56 <danbri> with weblog, i don't like to go back and edit entries (except fix 404s etc maybe)

20:18:19 <dajobe> you could have gone with blosxom, you can then use wiki syntax or whatever you like

20:18:40 <dajobe> tho I guess there must be a MT plugin for wiki now

20:18:55 <danbri> could've, and might do elsewhere

20:19:00 <danbri> i have a soft spot for MT tho

20:19:17 <dajobe> well, I'm nearly 1400 entries into using it...

20:19:39 <dajobe> the point was to reduce your syntaxes by 1

20:20:04 <edd> i'm playing with using zope

20:20:14 <edd> a weblog is merely a way of indexing the content

20:20:20 <edd> it needn't be stored in the way it's presented

20:23:32 * edd progressively finds that while he left foafbot gathering dust, much has happened in the foafworld

20:24:52 <danbri> yeah, i was suprised just now (rewriting http://rdfweb.org/topic/FoafProject) to realise that your article for developerWorks was just a year ago

20:25:22 <danbri> so, i fixed up mailing lists a bit, rss feeds, wiki front page, rdfweb.org front page

20:25:32 <danbri> next stop the foaf spec, and/or foaf homepage. prolly spec.

20:29:28 <mortenf> danbri, any chance the various (growing set of) chat IDs could go in a separate vocab?

20:31:12 <danbri> I'm worried that purism re keeping it general (versus naming companies products services), which i'm otherwise inclined towards, could have cost for users...

20:31:31 <danbri> remembering the property name is hard enough, remembering which of a half-dozen extension vocabs it lives in is another

20:32:00 <mortenf> i was thinkin a single chat-ID vocab...

20:32:02 * danbri thinks there is a way for non-foaf namespaces to also name chat account servers, and have the schema contain info that allows us to figure the homepage of the acconut server

20:32:24 <danbri> problem is that chat IDs and general website logins (eg. your Amazon account) are basically the same problem.

20:32:35 <mortenf> perhaps the foaf ns could have a super-property that the others could do subPropertyOf off?

20:32:43 <danbri> yes, exactly

20:32:59 * danbri thinks the subProperty defintion could also mention serviceHomepage and suchlike

20:33:15 <mortenf> yeah, with owl.

20:34:01 <danbri> i hadn't thought to use owl in particualr, but maybe that could add something, make the assumptions explicit etc

20:34:46 * mortenf just uses irc, no clue to what'd be interesting for other systems...

20:35:17 <danbri> http://antipop.zapto.org/foaf.rdf has <msnChatID>kentaro_tan</msnChatID>

20:35:46 <danbri> ok, taking a break. back later.

20:36:12 <mortenf> c ya

20:38:35 <danbri> css for http://rdfweb.org/ is still screwy though... in Konquoror anyway, articles all squished up.

20:38:46 * danbri still doesn't really get CSS, never took the time to learn enough

20:39:13 <mortenf> i've stopped looking at css in konq - it seems way off.

20:39:31 * danbri stops blaming self, heads offline

21:29:18 <danbri> I've updated the FoafNaut wiki entry somewhat, see http://rdfweb.org/topic/FoafNaut but it is still pretty sparse

21:29:31 <danbri> In particular, the screenshot is ancient. Pre-Liz. Contributions welcome!

21:33:05 <mortenf> heh, just noticed your '98 homepage :)

21:34:01 * JibberJim remembers the version of the screenshot with fondness

21:34:21 <danbri> simple times, before foafnaut took over your life? ;)

21:35:12 <danbri> homepage -- yeah, most of foaf idea was there except (a) for seeAlso linking, (b) actually using it in the real web, vs entity-escaped HTML whimsy

21:39:30 * danbri checks archives; seeAlso got added ~Jan 1999 to RDFS

21:40:49 <danbri> so anyone got some more foafnaut screenshot urls?

21:43:14 <JibberJim> nope

21:57:06 <danbri> anyone know of an RDF property for representing rfid's?

22:08:56 * edd hears a pin drop

22:09:00 * danbri updates (by obsoleting) http://rdfweb.org/dev.html

22:10:13 <danbri> rfid: well i'll make one then. need to be able to talk about physical objects... (buyable, sellable physical objects)

22:11:50 <mortenf> hmm, foaf:weblog is for "..., a group,..." and an owl:iFP ?

22:12:40 <danbri> a weblog is the weblog of something (we don't say what)

22:13:20 <mortenf> it's not clear, that no two people share a weblog.

22:14:14 <danbri> there are shared weblogs; those weblogs can't be modelled as the foaf:weblog of both of them. so yeah, the definition could do with clarifying.

22:14:21 <danbri> i'd put foaf:weblog in the 'testing' category

22:14:34 <mortenf> but typepad...

22:16:19 <danbri> the OWL semantics of foaf:weblog won't break typepad's implementation.

22:16:58 <danbri> ie we can tweak the def'n and the world mostly survives (-> testing)

22:17:03 <mortenf> true, but then let's remove the iFP for now?

22:17:07 <danbri> versus do that and cause chaos

22:17:44 <mortenf> quite so.

22:17:47 <danbri> well, i saw a 'work-weblog' / 'home-weblog' distinction somewwhere that was tempting, so I think you're prolly right, it can be commented out

22:18:15 <danbri> straw poll: any objections to foaf:weblog no longer being advertised as an owl:InverseFunctionalProperty ?

22:18:16 <mortenf> btw, still quite a few daml+oil references on the ns page, in case it's next on your list :)

22:18:29 * danbri left the d+oil in the markup on purpose

22:18:36 <mortenf> oh.

22:18:53 <danbri> all those daml fundees might find it useful

22:20:31 <mortenf> well, i guess the same "problem" holds for homepage (except there's an Organization there that makes it easier to explain - any conclusions on Group?).

22:21:08 <danbri> companies have homepages, groups have homepages, that's important to allow.

22:21:50 <mortenf> yeah, and i guess that's why i guess it's ok for weblog as well, basically same thing.

22:22:01 <danbri> the rdfs:domain should make clear that we don't assume people are the only thing w/ homepages

22:22:09 <danbri> this will be useful for FoafCorp, in time

22:22:14 <mortenf> yep.

22:22:21 <danbri> hmm, did we just swap opinions re foaf:weblog? ;)

22:22:28 <mortenf> :)

22:22:38 * danbri DCCs mortenf his old opinion back

22:22:47 <danbri> wanna send mail on this?

22:22:52 <mortenf> never could've this on a weblog or wiki...

22:22:55 <mortenf> sure.

22:23:11 <danbri> I was set to comment out the assertion, too: danbri@fireball:~/cvs/cvs.danbri.org/xmlns.com/htdocs/foaf/0.1$

22:23:25 <danbri> it does need documenting though

22:23:38 <mortenf> yeah.

22:24:55 <mortenf> so, did i just volunteer to send a mail to the list re this?

22:25:12 <danbri> like how do i say that i write the weblog at rdfweb.org? me -project->_1--weblog->http://rdfweb.org/

22:25:32 <danbri> if you've got time, i'm all typed out, but want to save some for hacking on spec

22:25:39 <mortenf> yep, that's the way to do it.

22:26:04 <mortenf> ok, i'll send a probe, and you can write about it on the weblog tomorrow :)

22:26:36 <danbri> heh ok

22:26:56 <danbri> thanks re mail

22:27:07 <mortenf> or another day, esp. audience-wise...

22:28:25 * danbri will be trying to batch up summaries of tech discussion, for those with half-an-eye on foaf

22:40:12 <deusx> deusx is now known as deusx_away

22:40:52 <mortenf> mail sent.

22:42:27 <danbri> thx!:)

22:42:49 <Morbus> ben trott just uploaded xml::foaf 0.2

22:42:49 <mortenf> a couple of hopefully non-controversial ns suggestions: drop linkedWith, theme, topic_interest - or provide better explanations.

22:42:55 <Morbus> trying to find an announcement of it.

22:43:14 * danbri nods, sympathises w/ vocab thing

22:43:39 <danbri> topic_interest... was a reason for that... hum, but escapes me. 'interest' is screwed up, gets used for uris and strings, though mostly for uris.

22:43:48 <danbri> all could go without too much damage, i think

22:43:58 <danbri> xml:foaf .2, excellent

22:44:11 <mortenf> perhaps topic_interest was for the actual topic instead of the page about the topic?

22:44:21 <mortenf> or something like that?

22:44:28 <Morbus> danbri: doesn't look like its mirrored on cpan yet.,

22:44:36 <Morbus> but the perlbot over in #perl just annouced it, so shortly.

22:44:39 * mortenf can't find it

22:44:41 <Morbus> no notes on any of his blogs.

22:45:01 <danbri> all in good time :)

23:07:22 <bitsko> wow, coincidence! I just posted a note on their 0.1 blog entry page

23:14:45 <mortenf> nn.