Semantic Web Interest Group IRC Chat Logs for 2008-03-10

00:38:20 <bear> bear is now known as bear_afk

02:15:26 <ramsey_> ramsey_ is now known as ramsey

03:23:10 <bear_afk> bear_afk is now known as bear

07:03:54 <mhausenblas> morning SWorld

07:04:38 * mhausenblas looking at LeeF' survey and discussion on modeling stats in RDF

07:04:40 <mhausenblas>http://www.thefigtrees.net/lee/blog/2008/03/modeling_statistics_in_rdf_a_s.html

07:04:42 <dc_swig> A: http://www.thefigtrees.net/lee/blog/2008/03/modeling_statistics_in_rdf_a_s.html from mhausenblas

07:04:57 <mhausenblas> A:| Modeling Statistics in RDF - A Survey and Discussion

07:04:59 <dc_swig> Titled item A.

07:05:10 <mhausenblas> I certainly agree with him ;)

07:12:46 <mhausenblas> dajobe around?

07:13:15 * mhausenblas asked Manu regarding license of http://rdfa.digitalbazaar.com/librdfa/

07:13:35 * mhausenblas now knowns better: it's LGPL v3

07:13:37 <mhausenblas> :)

08:20:32 <Shepard`> Shepard` is now known as Shepard

08:31:49 <scor> fgiasson: ping

08:45:48 <EtnaRosso> morning all

09:25:34 <TipTop> anyone has a bit of knowledge about the Sesame2 API?

09:25:46 <TipTop> i am going from beta4 to official release

09:25:55 <TipTop> and some classes have just disappeared

09:26:03 <TipTop> from the API ;(

09:27:36 <jeen> TipTop, which ones are those?

09:27:52 <jeen> there has been some renaming I think, but I don't think anything has disappeared....

09:30:16 <TipTop>http://pastebin.com/d16b0bb94

09:30:17 <dc_swig> B: http://pastebin.com/d16b0bb94 from TipTop

09:30:32 <TipTop> SailConfig is not found by my IDE

09:31:00 <TipTop> nor repConfig.addSailConfig

09:31:45 <jeen> hmmm. the configuration stuff changed quite drastically, it's completely RDF-based now.

09:32:25 <jeen> so you might be right that SailConfig is gone.

09:33:53 <TipTop> SailConfig has become RepositoryImplConfig ?

09:34:37 <TipTop> or something like that?

09:35:47 <jeen> not quite. there's a org.openrdf.sail.config package which contains a SailFactory and a SailImplConfig class.

09:36:13 <jeen> but I don't think you actually need to touch it, if your goal is simply to create a repository object for querying

09:36:36 <TipTop> i must port some code from beta4 to 2.0

09:37:01 <TipTop> which contains several initialization methods

09:37:11 <TipTop> for example the one that i pastebin-ed

09:37:16 <TipTop> (which links to a database)

09:37:57 <jeen> that wouldn't work anyway I guess, AFAIK Sesame 2 does not have a relational DB backend yet (and neither did beta-4 so that wouldn't have worked then either)

09:38:12 <TipTop> ok let's forget that one then

09:38:31 <jeen> but creating, say, a native store Sail is very easy:

09:38:52 <TipTop>http://pastebin.com/d647a0ea2

09:38:53 <dc_swig> C: http://pastebin.com/d647a0ea2 from TipTop

09:39:08 <TipTop> my old code to create a native store Sail

09:39:26 <jeen> TipTop, have a look at http://www.openrdf.org/doc/sesame2/users/ch08.html#d0e679

09:39:36 <jeen> contains a code example for creating a native store

09:40:31 <jeen> basically you don't use config objects anymore, you simply create the NativeStore object with appropriate parameters and wrap it in a SailRepository object.

09:41:07 <TipTop> that sounds like my answer :)

09:41:14 <jeen> ok glad to help :)

09:41:40 <TipTop> if you go to tenerife in june, i will happily offer you a few beers :)

09:42:27 <jeen> not sure I'll be there, but cheers.

09:51:15 <TipTop> jeen: btw, have you had a look at the store benchmark available here: http://www4.wiwiss.fu-berlin.de/benchmarks-200801/

09:51:17 <phenny> TypeError: object does not support item assignment (file "/var/www/inamidst.com/htdocs/phenny/modules/head.py", line 147, in noteuri)

09:54:21 <nslater> naughty phenny

09:55:23 <sbp> (fixed)

09:56:28 <mahound> hello

09:58:25 <sbp> hey mahound

10:28:27 <BenO> Anyone here familar with OAI-ORE? I've got a query about derefencing (object) URIs

10:32:25 * mhausenblas no OAI-ORE wiz, but very interested

10:32:42 * mhausenblas what is the matter BenO?

10:34:09 <BenO> It related to the old issue of whether a URI should resolve to the record page for an item or to its 'node'

10:34:35 <mhausenblas> ah. so looking at http://www.openarchives.org/ore/documents/CompoundObjects-200705.html ?

10:35:02 <mhausenblas> or another source?

10:35:10 <BenO> Yes, so should an objects URI resolve to a handy HTML page for it, or to a serialisation of its resource map

10:36:19 <nslater> hey BenO

10:36:25 <BenO> Hi

10:39:52 <mhausenblas> BenO, are you aware of http://efoundations.typepad.com/efoundations/2008/01/following-your.html ?

10:42:54 <BenO> At the moment I am sorting out the url routes for the new site - see http://archive.sers.ox.ac.uk:5000/information/urls for dosumentation on it

10:43:18 <BenO> (I had read that, and it and other articles like it are why I'm cautious)

10:44:33 <BenO> In essence, all the (compound) items in the archive have URIs and I am providing a mechanism to simply append the URI onto a root url to resolve it

10:45:32 * mhausenblas still very interested to discuss these issues (now gotta run to a meeting). BIAB (or maybe follow-up via mail ? :)

10:54:54 <jeen> TipTop, sorry, was away. Yes, I've seen that. bit of a shame they used a beta release (but I'm sure that was all that was available at the time).

10:58:43 <TipTop> jeen: you think performances have increased between beta and official release?

10:59:11 <TipTop> i was wondering which indexes virtuoso has that make it so performant

11:12:11 <iv_an_ru__> TipTop, we're using bitmap indexes in all cases when the last field of the key is IRI ID, we pack all indexes field by field (so we don't waste disk cache on storing repeating values), we try to keep logically sequential disk pages in right sequence in database file so we sometimes can read perform disk I/O by long extents, not by individual pages.

11:24:17 <TipTop> iv_an_ru__: virtuoso is coded in C or C++, i presume

11:25:53 <iv_an_ru__> Pure C. There were too many mismatches between different C++ compilers in 1995 when it's been started.

11:26:10 <danieljohnlewis> Does anyone know if Horus still exists?

11:26:24 <danieljohnlewis> (i.e. the one that used to be here: http://www.wiwiss.fu-berlin.de/suhl/bizer/rdfapi/tutorial/horus/ )

11:26:37 <kwijibo> is it still 404ing?

11:26:48 <kwijibo> yep

11:26:51 <danieljohnlewis> yep

11:27:08 <kwijibo> dunno - all I know is it's been down for ages and ages now

11:27:56 <danieljohnlewis> oh well, thats off the list which I am currently making then

11:31:00 <danieljohnlewis> the jFresnel site doesn't seem to be working either: http://jfresnel.gforge.inria.fr/

11:31:19 <kwijibo> oat has a fresnel engine doesn't it?

11:32:12 <danieljohnlewis> it does indeed :-)

11:33:23 <kwijibo> is there a tutorial or anything for it?

11:35:32 <danieljohnlewis> found something: http://demo.openlinksw.com/DAV/JS/oat/docs/OATDOCfresnel.html

11:35:49 <danieljohnlewis> :)

11:36:22 <danieljohnlewis> and because the OpenLink RDF Browser uses OAT, you get fresnel through RDF Browser too

11:37:56 <TipTop> OAT looks pretty cool

11:38:35 <danieljohnlewis> TipTop: it is quite cool: Semantic Web Tool + AJAX + Database Connectivity + Eye Candy

11:39:04 <danieljohnlewis>http://oat.openlinksw.com/

11:39:05 <dc_swig> D: http://oat.openlinksw.com/ from danieljohnlewis

11:39:34 <danieljohnlewis> D: OpenLink AJAX ToolKit = Semantic Web Tool + AJAX + Database Connectivity + Eye Candy

11:39:35 <TipTop> this is basically something i hacked for myself

11:39:36 <dc_swig> Added comment D1.

11:39:49 <danieljohnlewis> D: capable of running Fresnel

11:39:50 <dc_swig> Added comment D2.

11:39:55 <danieljohnlewis>http://demo.openlinksw.com/DAV/JS/oat/docs/OATDOCfresnel.html

11:39:56 <dc_swig> E: http://demo.openlinksw.com/DAV/JS/oat/docs/OATDOCfresnel.html from danieljohnlewis

11:40:14 <TipTop> i coded the visual query builder in java, in the end :)

11:40:17 <danieljohnlewis> Documentation for running Fresnel through OpenLink Ajax Toolkit (OAT)

11:40:23 <danieljohnlewis> E: Documentation for running Fresnel through OpenLink Ajax Toolkit (OAT)

11:40:25 <dc_swig> Added comment E1.

11:53:49 <TipTop>http://demo.openlinksw.com/DAV/home/demo/Public/Reports/MapMashups/employee_sales_by_ship_country_pivot_google_all.xml

11:53:50 <dc_swig> F: http://demo.openlinksw.com/DAV/home/demo/Public/Reports/MapMashups/employee_sales_by_ship_country_pivot_google_all.xml from TipTop

11:53:53 <TipTop> this example fails

11:57:03 <danieljohnlewis> TipTop: we are having a few problems with the demos at the moment, I believe someone is looking into that

12:13:37 <bengee> heya phenny

12:13:38 <phenny> bengee: 08 Mar 20:58Z <TedThibodeauJr> tell bengee if you colorize your text, foafbot doesn't hear you. See http://chatlogs.planetrdf.com/foaf/2008-03-08#T20-58-15

12:13:41 <phenny> bengee: 09 Mar 22:03Z <TedThibodeauJr> tell bengee it seems foafbot also can't understand trailing whitespace. See http://chatlogs.planetrdf.com/foaf/2008-03-09#T17-35-03

13:34:58 <yvesr>http://blog.dbtune.org/post/2008/03/10/Finding-a-new-flat-using-Exhibit

13:35:00 <dc_swig> G: http://blog.dbtune.org/post/2008/03/10/Finding-a-new-flat-using-Exhibit from yvesr

13:35:12 <yvesr> G:| Building an Exhibit out of Gum Tree

13:35:14 <dc_swig> Titled item G.

13:36:16 <danieljohnlewis> ooh I need to find a new flat

13:40:14 <yvesr> danieljohnlewis: i hope this thing can be useful :)

13:40:30 <TipTop> it looks useful for sure!

13:40:50 <yvesr> G: Exactly 100 lines of code

13:40:52 <dc_swig> Added comment G1.

13:41:59 <TipTop> does it require some python stuff?

13:46:21 <yvesr> TipTop: yes, universal feed parser, simplejson and geopy

13:46:28 <yvesr> and rdflib, of course

13:46:42 * yvesr should write that down, actually

13:46:49 <TipTop> could it be done by the client using something like OAT

13:49:06 <yvesr> hmm

13:49:14 <yvesr> actually, just the scraping is written in python

13:49:34 <yvesr> then, you can produce an exhibit json file

13:49:43 <danieljohnlewis> theres nothing stopping you from feeding it into OAT

13:49:52 <yvesr> yes, of course

13:49:58 <yvesr> especially as you can output rdf also

13:50:09 <danieljohnlewis> perfect :-)

13:50:55 <TipTop> my idea was to reduce the server part to the minimum

13:51:12 <TipTop> plain rdf serving, or may be sparql endpoint

13:51:18 <TipTop> no business logic

13:51:50 <kwijibo> yvesr: cool, I didn't know about that blog

13:52:06 <TipTop> it is a geoRSS, right?

13:52:32 <yvesr> TipTop: what is a georss?

13:52:40 <yvesr> the gum tree rss is clearly not geo at all :)

13:53:42 <TipTop> how do you extract geolocation then?

13:53:45 <TipTop> by script?

13:54:18 <TipTop>http://en.wikipedia.org/wiki/GeoRSS

13:54:19 <dc_swig> H: http://en.wikipedia.org/wiki/GeoRSS from TipTop

13:56:06 <yvesr> TipTop: the gum tree rss provides a link to the entries, i scrap them to extract a location label, and i geocode using google map api

14:11:01 <SinDoc> SinDoc is now known as SinDoc|Lunch

14:20:33 <kwijibo> yvesr: is a web service in the works?

14:38:15 <mahound> yvesr, nice :)

14:49:05 <SinDoc|Lunch> SinDoc|Lunch is now known as SinDoc

14:50:23 <KiYanWang> KiYanWang is now known as mmmmmRob__

14:56:03 <mmmmmRob__> mmmmmRob__ is now known as KiYanWang

15:00:08 <yvesr> kwijibo: yes, should be - http://moustaki.org/flatscrap/

15:00:29 <yvesr> (displaying last 40 ads for north of the river)

16:24:08 <DanC> .g ping

16:24:09 <phenny> DanC: 09 Mar 19:47Z <tommorris> tell DanC that #sxsw exists and he might want to hang out there

16:24:10 <phenny> DanC: http://www.pinggolf.com/

16:24:26 <DanC> yeah, but it doesn't have logger, chumpbot, etc.

16:30:43 <DanC> how many have an openid account? (~everybody ; full room; ~500)

16:30:54 <DanC> how many have used it in the last 5 days? (lots)

16:31:03 <danbri> are you scribing from sxsw?

16:31:08 <DanC>http://2008.sxsw.com/interactive/programming/panels_schedule/?action=show&id=IAP060295

16:31:09 <dc_swig> I: http://2008.sxsw.com/interactive/programming/panels_schedule/?action=show&id=IAP060295 from DanC

16:31:16 * danbri just joined to join #sxsw for that

16:31:30 <danbri> thanks danc :)

16:31:42 <DanC> I:|A Critical Look At OpenID

16:31:43 <dc_swig> Titled item I.

16:32:05 <e_s_p> Hmm

16:32:23 <DanC> I:panelists from Wikia, AOL, Six Apart, Google

16:32:24 <dc_swig> Added comment I1.

16:32:28 <e_s_p> Haw!

16:32:30 <e_s_p> Wikia!

16:33:34 <DanC> I:Simon Willison gets a laugh for saying Microsoft proposed to solve the password problem by having everybody have an account with them, but I don't hear anybody bristle at the idea that everybody should have a twitter account

16:33:35 <dc_swig> Added comment I2.

16:33:53 <e_s_p> I find it annoying that they haven't rolled out OpenID.

16:33:58 <e_s_p> Wikia, that is.

16:34:12 <DanC> I:"An OpenID is a URL" yay!

16:34:13 <dc_swig> Added comment I3.

16:34:40 <e_s_p> Well, not 100% true

16:34:41 <nslater> For small values of URL anyway, at least while OpenID continues to ignore HTTP semantics. :)

16:34:52 <e_s_p> Is an XRI an URL?

16:34:56 <e_s_p> I don't think so.

16:34:57 * danbri sweeps IRI vs URI vs URL under the lumpy carpet, happily

16:35:03 <danbri> XRI is just a URI in bad taste

16:35:11 <e_s_p> *rimshot*

16:35:19 <DanC> yeah... I wonder about openid and http semantics, but I haven't worked thru the details; have you, nslater ?

16:35:24 <Arnia> danbri: what is the bad taste? Grapefruit?

16:35:33 <e_s_p> Wha...?

16:35:36 <e_s_p> HTTP semantics?

16:35:39 <nslater> DanC: I have a link to a thread I started with the OpenID folk, one second...

16:35:44 <e_s_p> C'mon!

16:35:55 <danbri> bad taste is intermind leftovers

16:35:55 <nslater> DanC: http://openid.net/pipermail/general/2008-March/thread.html

16:36:03 <nslater> DanC: Problems with OpenID and TAG httpRange-14

16:37:04 <DanC> hmm... httpRange-14; I wasn't expecting that; I was more worried about 3xx responses in place of 4xx "credentials needed"

16:37:10 <GabeW> hi nslater, danbri, DanC, and others

16:37:22 <GabeW> danbri's tweet brought me here

16:37:27 <nslater> DanC: yeah, I really mislabled the subject, my concern is with 302, 303 and 307

16:37:36 <DanC> er... "I am arguing that this is broken"... but I don't see the argument

16:37:38 <danbri> hi gabeW

16:37:45 * danbri was looking for the argument either

16:37:48 <nslater> DanC: yes, the mailing list software munged my mail

16:37:59 * danbri shouldn't grumble casually about XRIs either, it wasn't quite fair...

16:38:14 <GabeW> everyone does, why shouldn't you?

16:38:33 * DanC is happy that XRIs aren't mentioned

16:38:35 <GabeW> but seriously, I think openid 2.0 is kinda baked

16:38:42 <GabeW> like has been for a while

16:38:50 <ab> half baked? or just kinda

16:38:52 <e_s_p> nslater: given that OpenID already has a delegation mechanism, what is the purpose of using 3xx responses?

16:38:52 <GabeW> haha

16:39:06 <danbri> i'll grumble about everything intermind (or any renamings of that company) touches feeling tainted by the p3p patent disgrace

16:39:10 <GabeW> ah

16:39:12 <danbri> i've never reviewed XRIs in detail

16:39:21 <GabeW> danbri: i understand your sour stomach

16:39:29 <nslater> My argument boils down to the fact that OpenID canonicalises 302, 303 and 307 redirects as if the final URI is a replacement for the original one.

16:39:40 <e_s_p> Yes, I see that

16:39:47 <GabeW> nslater: I wasn't following that thread

16:39:47 <DanC> I:ah... that's what openid 2.0 gives you... a "sign in with my yahoo openid" button

16:39:48 <dc_swig> Added comment I4.

16:40:14 <e_s_p> Wow

16:40:20 <e_s_p> This is actually quite topical for me

16:40:34 <DanC> nslater, have you elaborated the argument with an example? (and have you seen the AWWSW rules for testing such things in a black-and-white fashion?)

16:40:39 <e_s_p> I'm debugging the MediaWiki OpenID extension for OpenID 2.0 right now

16:40:40 <GabeW> danbri: well, you know my involvement, both in the XRI stuff (IPR and spcs) and the openid IPR work... so if you ever have questions, you definitely know how to get in touch

16:40:56 <GabeW> ;)

16:41:04 <danbri> well we seem to be in touch :)

16:41:13 <GabeW> true

16:41:15 <danbri> so how would you summarise the enduring charm of those specs?

16:41:27 <GabeW> which ones?

16:41:27 <danbri> i think somehow technologies sometimes 'fit' a personality or brain

16:41:35 <danbri> like CSS float and XSLT don't fit mine, nor LISP

16:41:35 <GabeW> hehe

16:41:35 <nslater> DanC: yes, I took my OpenID <http://bytesexual.org/> as an example which 303s to another URI. Most OpenID applications then use the final URI as my OpenID, which I am claiming is broken. I'm a little over my head with all this, so I wouldn't know how to take it futher.

16:41:40 <GabeW> danbri: really?

16:41:40 <danbri> but RDF is sorta obvious fit

16:41:47 <GabeW> lisp? I thought lisp was the universal language

16:41:49 <danbri> and Ruby, for eg

16:41:54 <danbri> heheh

16:41:57 <GabeW> oh I'm with ya there, on Ruby

16:41:59 <DanC> I:SW: email is a single-sign-on mechanism; that "forgot your password?" callback is single-sign-on, but with a really sucky UI. you can think of OpenID as single-sign-on with a UI that doesn't suck. [much audience appreciation/chuckle]

16:42:00 <dc_swig> Added comment I5.

16:42:10 <danbri> 'those', ... i'm a bit vague as I don't know the field, but "XRI etc"?

16:42:35 <DanC> nslater, can you make the argument without using the word "broken"? i.e. try to derive P and not P?

16:42:44 <GabeW> well

16:43:08 <GabeW> danbri: nothing has really changed significantly in a long time

16:43:22 <nslater> DanC: no, I am sorry but you lost me there. :)

16:43:31 <GabeW> well, long time in internet time

16:44:12 <DanC> "P and not P" for example "fred is a dog and fred is not a dog"

16:44:33 <DanC> or "sqrt(2) is rational and sqrt(2) is not rational"

16:44:37 <e_s_p> nslater: Maybe the way you want to say it is that the proper way to re-request a resource after non-permanent 3xx redirects is to request the resource from the same URL

16:45:00 <nslater> DanC: Instead of using broken, I have mostly pointed out that this violates RFC 2616 § 10.3.4

16:45:02 <danbri> I don't believe openid requires the openid URI to directly be considered the URI of the agent that owns it

16:45:13 <danbri> so i can use the page http://danbri.org/ as my openid

16:45:21 <e_s_p> Yes, of course

16:45:24 <nslater> e_s_p: surely quoting RFC 2616 § 10.3.4 is sufficient to explain the problem?

16:45:38 <danbri> if nslater has a uri for something that is suposed to be the 'thing itself', ... maybe that is fine, but shouldn't be used as the openid

16:45:46 <DanC> er... how about quoting by value (i.e. excerpting) as well as by reference?

16:45:57 <nslater> well, yes I have done that too :P

16:46:04 * e_s_p goes to look up quoted section, somewhat annoyed.

16:46:28 <nslater> 10.3.4 303 See Other

16:46:32 <nslater> The new URI is not a substitute reference for the originally requested resource.

16:46:34 <DanC> annoyed at me? sorry, but my attention is divided between #swig and this in-person panel

16:46:52 <nslater> annoyed at me, I think :)

16:46:53 <e_s_p> Probably what your correspondents thought, too.

16:47:00 <e_s_p> "The new URI is not a substitute reference for the originally requested resource."

16:47:26 <nslater> e_s_p: I was careful to quote but the URI of the RFC, the section and excerpt the text.

16:47:32 <nslater> *both

16:47:39 <DanC> I'm worried about caches getting confused between the protected photo and the authentication questions around it

16:48:12 <nslater> danbri: why don't you think that a URI for the thing it's self should be used as the OpenID?

16:48:29 <e_s_p> DanC: annoyed to have to look things up.

16:48:55 <e_s_p> Being lazy and not having memorized RFC 2616 verbatim.

16:49:27 <e_s_p> nslater: what was the counter-argument on this point? I think you're right here.

16:49:28 <DanC> I:panel discusses educating the masses and how to educate people about having URLs that represent them. A: at a restaurant here in Austin, if you ask random people what their myspace url is, most of them can tell you

16:49:29 <dc_swig> Added comment I6.

16:49:57 <e_s_p> Probably worth noting is that OpenID 2.0 doesn't require that you know the URL of your own account!

16:50:08 <DanC> I:Q: is there any central record keeping involved in OpenID? A: no, which makes it really hard to quantify adoption

16:50:09 <dc_swig> Added comment I7.

16:50:11 <e_s_p> If you just know that you have an account on Yahoo, that should be enough.

16:50:17 <DanC> yup

16:50:23 <e_s_p> Actually, that's not entirely true

16:50:35 <e_s_p> I think the OpenID Directory folks are trying to collect some statistics

16:50:36 <TipTop> what's the benefit of having an openID account?

16:50:39 <nslater> e_s_p: well, the thread got to the point where they basically suggested that I pay money to board or trust or something in order to get attention for the bug, which I find, interesting

16:50:43 <TipTop> (sorry for the FAQ)

16:51:15 <e_s_p> TipTop: I'm not sure there's an actual answer to that from the user perspective

16:51:16 <DanC> I:hmm... the title of the panel is "a critical look at OpenID" which suggests some diversity of positions, but this seems to be an openid love-fest

16:51:17 <dc_swig> Added comment I8.

16:51:39 <Arnia> Is OpenID using HTTP but not following the RFC in all matters generally considered a bug?

16:52:07 * Arnia has his position, but is interested to hear others' comments

16:52:21 <nslater> Well, I guess it depends who you ask. Some of the community members were certain that OpenID could reasonably sit on top of HTTP while at once breaking it's semantics.

16:52:38 <Arnia> Better than rehashing the entire thread in #swig I think :)

16:52:42 <e_s_p> Can we agree that this isn't a grievous breaking of semantics?

16:53:01 <ab_> Arnia: if you edit a spec called HTML5 and explicitly oppose webarch in numerous ways, its not considered a bug anywyas

16:53:10 <nslater> I think it is, though "grievous" is a pretty emotional word to use.

16:53:10 <e_s_p> It's a fairly niche situation -- most people will use OP-provided OpenID URLs.

16:53:14 <Arnia> e_s_p: I think it is though... or could potentially be. Depends how people use OpenID I guess

16:53:16 <DanC> if there were a clear argument that openid breaks http, I'd take it to the TAG, nslater . but I don't see it, yet.

16:53:45 <e_s_p> "breaks" is really strong.

16:53:46 <ab_> does <http://mysite/me.rdf#i> work as an openID URI

16:53:54 <e_s_p> ab_: no

16:54:06 <DanC> doubt it; ab; openid depends on an HTML representation

16:54:10 <TipTop> is openID a UUID for web objects?

16:54:12 <nslater> DanC: okay, perhaps not breaks, but OpenID does clearly violate some of the explicit semantics of HTTP which I think is an issue.

16:54:13 <DanC> .g openid grddl connolly

16:54:14 <phenny> DanC: http://dig.csail.mit.edu/2007/id/talk19

16:54:22 <e_s_p> nslater: "some"?

16:54:29 <e_s_p> Or this particular one?

16:54:29 <nslater> e_s_p: 302, 303 and 307 redirects

16:54:31 <Arnia> ab_: that was a little snarky (although I'm so stressed right now I should probably apologise for unwarranted snark myself)

16:54:37 <nslater> e_s_p: I count three :p

16:55:21 <DanC> yes, well, so the issue is documented, nslater ; given the benefit of openid, that sort of thing is pretty easy to sweep under the rug. I'm looking for arguments that are strong enough that I'd be willing to get people to change the way their software works

16:55:25 <e_s_p> That's mincing things pretty small.

16:56:18 <DanC> these slides http://dig.csail.mit.edu/2007/id/talk19 are my noodling on mixing openid and RDF via GRDDL

16:56:49 <Arnia> nslater: didn't you have problems with OpenID implementations getting confused about who you are? Or something like that

16:56:49 <danbri> nslater, because it seems to go against the way openid is being deployed

16:56:57 <Arnia> That might be a different issue though

16:56:58 <danbri> where the document view of the thing is important

16:57:01 <danbri> eg. for autodiscovery

16:57:12 <danbri> well i prefer LINK REL to the fancy new stuff, maybe thats my problem

16:57:46 <nslater> Well, when I input my OpenID in various sites, the OpenID software minces my OpenID identity.

16:57:47 <DanC> LINK REL is simpler than lots of other stuff, but lots of people just want to maintain one file

16:58:17 <nslater> So, on one site I am <http://bytesexual.org/about/>, on another I am <http://bytesexual.org/> and on one I was </about/>!

16:58:26 <nslater> So, that's a pretty sorry state of afairs.

16:59:09 <danbri> do you ever find your url blocked by helpful 'family friendly' filters?

16:59:26 <DanC> I:Yahoo is an OpenID provider as of January. Yahoo is *not* an OpenID relying party. [:-/]

16:59:27 <dc_swig> Added comment I9.

16:59:30 <nslater> the only problem I ever had was freenode's spam filter when registering a channel :p

17:00:01 <DanC> I:"yahoo only supports openid 2.0 and will never support openid 1.0"

17:00:02 <dc_swig> Added comment I10.

17:00:11 <DanC> I:... says Levitt

17:00:12 <dc_swig> Added comment I11.

17:00:37 <DanC> he referes to developer.yahoo.com/opendi/loginbuttons.html

17:00:55 <DanC> some relying parties : twitterfeed.com plax.com fatdrive.tv [darn; too fast]

17:01:06 <GabeW> plaxo.com ?

17:01:07 <e_s_p> AOL is also a provider and not a relying party

17:01:13 <e_s_p> wikitravel.org

17:01:15 <e_s_p> Is one

17:01:17 <DanC> drupal 6 with patch node/216101

17:01:20 <e_s_p> hampr.com

17:01:23 <DanC> logger, chump I:

17:01:23 <DanC> I:See [http://chatlogs.planetrdf.com/swig/2008-03-10#T17-01-23|discussion]

17:01:24 <dc_swig> Added comment I12.

17:01:29 <e_s_p> vinismo.com, kei.ki

17:01:37 <DanC> wordpress 2.2+ with disco patch

17:01:51 <DanC> I:see yahoo.com/openid

17:01:53 <dc_swig> Added comment I13.

17:02:05 <e_s_p> The problem is that on a Social Web, everybody wants to have lots of fat identities they own

17:02:38 <e_s_p> Also, it's HARD

17:02:49 <e_s_p> Implementing the provider side is easy

17:03:04 <e_s_p> Implementing the RP side is a pain in the keester

17:03:16 <GabeW> yes

17:03:18 <e_s_p> Having done both I can speak from experience

17:03:48 <DanC> I:DC: what about relying parties? wikipdeia? Bergman: say they won't until they solve their own sso problem ... because wikimedia has separate user dbs for... collisions...

17:03:49 <dc_swig> Added comment I14.

17:04:15 <e_s_p> What about Wikia?

17:04:28 <DanC> I:Williams: basecamp and [another]... niche... [darn; this is going too fast]

17:04:29 <dc_swig> Added comment I15.

17:05:03 <e_s_p> MediaWiki!

17:05:06 <e_s_p> B-)

17:05:12 <DanC> some nike push... goruneasy.com

17:05:17 <DanC> or rebok

17:05:26 <DanC> and example of guerilla dev deployment

17:05:31 <DanC> s/and/an/

17:06:45 <DanC> I:Q: why is yahoo not a relying party? how can we expect it to jump out of the niche if the big players won't do it? A: google did it on blogger Q: but that's still in the niche. Levitt: I can't speak to why yahoo didn't do it, but it's a bigger engineering effort

17:06:47 <dc_swig> Added comment I16.

17:08:40 <DanC> williams just gave a very nice deployment story

17:08:55 <DanC> I think this is being recorded. check it out.

17:09:18 <Arnia> Wonder how late I can go to bed and still get up in time for a 5.50am train from Durham to London.

17:09:25 <Arnia> Although I suppose I can sleep on the train.

17:10:15 <e_s_p> "recorded": usually the recordings don't come out for a few months after the conference.

17:10:28 <nslater> Yes, I can't find any links on the URI DanC pasted.

17:16:54 <e_s_p> Arnia: 5:49am is the answer

17:17:33 <Arnia> e_s_p: hah. At least the train has free wireless if I get bored :p

17:17:58 <Arnia> e_s_p: should probably try and rest for the mad dash across London when I arrive though

17:25:33 <DanC> hmm... an openid camp in the KC area ... that would be cool

17:26:10 <DanC> I:Q: what's blocking adoption? Recordon: let's poll... result shows usability gets many more hands than security, technology

17:26:12 <dc_swig> Added comment I17.

17:32:11 <DanC> I:Q: will openid help solve blog comment spam? A: yes, with distributed whitelisting via tiwtter friends lists and the like

17:32:13 <dc_swig> Added comment I18.

17:32:19 <DanC> </panel>

17:32:32 <DanC> battery down to 31%

18:23:38 <danbri>http://www.w3.org/2005/Incubator/urw3/

18:23:39 <dc_swig> J: http://www.w3.org/2005/Incubator/urw3/ from danbri

18:23:55 <danbri> J:|W3C XG on Uncertainty Reasoning

18:23:57 <dc_swig> Titled item J.

18:24:01 <danbri> J:Anyone following this?

18:24:02 <dc_swig> Added comment J1.

18:24:10 * Arnia ponders

18:25:20 <danbri> J:I was just chatting with [http://www.deri.ie/about/team/member/hak%20lae_kim/ Hak Lae Kim] from DERI Galway about his work on the [http://scot-project.org/ SKOT project] ...

18:25:21 <dc_swig> Added comment J2.

18:25:53 <danbri> J2:I was just chatting with [http://www.deri.ie/about/team/member/hak%20lae_kim/|Hak Lae Kim] from DERI Galway about his work on the [http://scot-project.org/|SKOT project] ...

18:25:54 <dc_swig> Replaced comment J2.

18:26:37 <Arnia> Yet more to read :)

18:27:08 <danbri> J:The current Wordpress exporter generates co-occurancy stats for tags (see [http://scot-project.org/scot/scot.rdf|example scot.rdf]), but I was thinking this should have conditional probabilities, ie. let us know, for some repository, how strong a predictor 'bristol' is for 'work', or vice-versa. Wondered if the XG had anything general that could be used.

18:27:09 <dc_swig> Added comment J3.

18:29:08 <Arnia> I'd also want a sample size measure such as confidence

18:29:20 <Arnia> So that I know how much evidence went into the conclusion

18:52:05 <danja__>http://blogs.talis.com/n2/archives/39

18:52:06 <dc_swig> K: http://blogs.talis.com/n2/archives/39 from danja__

18:52:27 <danja__> K?!

18:52:33 <danja__> K:| Noodling with Atom/RDF

18:52:34 <dc_swig> Titled item K.

18:53:30 <danja__> K: using a Planet Venus cache as source data, pushing through xslt into a triplestore

18:53:32 <dc_swig> Added comment K1.

18:55:31 <danja__> K: [SPARQL endpoint|http://hyperdata.org/sparql/demo/sparql-editor.html] (select "twitcrit reviews" from the drop-down)

18:55:33 <dc_swig> Added comment K2.

18:56:16 <mahound> dc_swig:list

18:56:17 <dc_swig> Not understood: list

19:02:53 <danbri> Arnia, yeah exactly

19:03:17 <Arnia> danbri: take it you've seen NAL then?

19:03:20 <danbri> ... the SKOT markup currently has some basic info, ... but we were discussing whether its best treated outside of the domain vocab

19:03:23 <danbri> .g NAL

19:03:29 <phenny> danbri: http://www.nal.usda.gov/

19:03:38 <Arnia> .g Non-Axiomatic Reasoning System

19:03:39 <phenny> Arnia: http://www.cogsci.indiana.edu/farg/peiwang/papers.html

19:03:48 <danbri> I have now, if you mean NASA's Agricultural ah.. .wait ;)

19:04:08 <danbri> nope, not familiar with that site nor acronym

19:04:22 <Arnia> The name 'non-axiomatic' refers to the fact that derived statements aren't treated as undeniable truths

19:04:23 <danbri> but I'm generally from that camp, in cogsci

19:04:32 <danbri> neural nets and whatnot

19:04:33 <Arnia> Uses higher-order uncertainty

19:04:55 <danbri> hmm ok well thats still pretty propositionalist, so i guess NARS is another camp :)

19:05:09 <danbri> is it deployed to real tools yet?

19:05:26 <Arnia> Yes. There is an open-source implementation I'm a contributor to; open-nars

19:05:29 <Arnia> .g open-nars

19:05:30 <phenny> Arnia: http://code.google.com/p/open-nars/

19:19:28 <danbri> ah cool, i'll nose around

19:19:38 <danbri> have you tried wiring it to rdf/sparql?

19:22:14 <Arnia> danbri: that's on my to-do list

19:22:23 <danbri> :)

19:22:49 <Arnia> danbri: I'm a bit swamped right now with getting a funding proposal together, but I've made my first steps by defining a new (nicer) serialisation inspired by N3 called LOAN

19:23:04 <Arnia> (Non-Axiomatic Ontology Language)

19:23:05 <LeeF_> Any recommendations for fun SPARQL'able public datasets for doing some SPARQL-by-example slides? foaf, dbpedia, and govtrack leap to mind, but thought i'd solicit more (GRDDL'able or RDFa'd data would be great too)

19:23:33 <Arnia> The first application I'm going to try is actually an experiment on del.icio.us tags

19:24:01 <Arnia> (seeing how we might integrate 'expert' and 'end-user' knowledge in a way which deals with disagreement)

19:24:24 <yvesr> LeeF_: you can try the dbtune datasets - http://dbtune.org/ quite a lot of music data in there

19:35:40 <ab_> heh ajaxian defenDing case-sensitive HTML els

19:36:11 <ab_> at least it should be consistent. path part of URI is case sensitive but the host part isnt?

19:36:38 <ab_> do any CMS/etc use case-insensitive URIs? i know mediaWiki goes out of its way to f' you on case

19:36:54 <ab_> 'this article must begin with a period and a leading letter, please follow this link (way to not use 303)

19:37:09 <LeeF_> yvesr, thanks, will do

19:37:26 <ab_> .title http://lunchfood.tumblr.com/post/28448128

19:37:27 <phenny> ab_: - Oh, you're at SXSW? Fuck off.

19:49:23 * bear waves to the nick that represents Tom Morris

20:02:26 <mahound> dc_swig:view

20:02:32 <dc_swig> G: Building an Exhibit out of Gum Tree (http://blog.dbtune.org/post/2008/03/10/Finding-a-new-flat-using-Exhibit)

20:02:32 <dc_swig> H: http://en.wikipedia.org/wiki/GeoRSS

20:02:32 <dc_swig> I: A Critical Look At OpenID (http://2008.sxsw.com/interactive/programming/panels_schedule/?action=show&id=IAP060295)

20:02:33 <dc_swig> J: W3C XG on Uncertainty Reasoning (http://www.w3.org/2005/Incubator/urw3/)

20:02:33 <dc_swig> K: Noodling with Atom/RDF (http://blogs.talis.com/n2/archives/39)

21:05:53 <Shepard`> Shepard` is now known as Shepard

21:39:53 <LeeF_> dajobe: does http://librdf.org/query/ understand (some dialect of) RDFa, by any chance?

21:47:54 <DanC> it groks GRDDL, LeeF_

21:48:03 <DanC> so if you have XSLT for RDFa, you're good to go

21:48:57 <LeeF_> Right, DanC, I saw that. Was just curious if any of the publicly deployed SPARQL endpoints do RDFa w/o GRDDL. Also don't know the state of anyone's RDFa XSLT these days

21:49:09 <danbri> is it feasible for such an XSLT to be automatically discoverable via GRDDL rules?

21:49:19 <danbri> the 'go rummage around the namespace' thingy...

21:49:51 <danbri> LeeF, I think ARC does RDFa

21:50:00 <DanC> which namespace, danbri ?

21:50:13 <LeeF_> As I understand it (though DanC will surely correct me if I'm wrong), the RDFa page needs to either point to the GRDDL HTML metadata profile and then point via <link rel="transformation" ..> to the XSLT...

21:50:19 * DanC should be heading to ballroom A for portable social networks...

21:50:21 <danbri> thats what I wasn't sure about

21:50:37 <danbri> (but portable social networks should come to you, surely?)

21:50:43 <LeeF_> ...or point to an RDFa metadata profile that itself can be GRDDL'ed into RDF which points (via grddl:profileTransformation ?) to the XSLT

21:50:45 <DanC> badump-bump... psh

21:51:00 <danbri> wish i was there, but i've used up my travel vouchers for a while

21:51:10 <LeeF_> danbri: thanks, I'll check our ARC

21:51:39 <danbri>http://arc.semsol.org/docs/v2/extractors

21:51:40 <dc_swig> L: http://arc.semsol.org/docs/v2/extractors from danbri

21:52:03 <danbri> L:|ARC2 Extractors, including RDFa

21:52:03 <dc_swig> Titled item L.

21:53:00 <LeeF_> Am I asking too much if I wonder if there are any ARC SPARQL endpoints out there configured to extract RDFa and willing to load arbitrary Web content for queries? :)

21:53:52 <kidehen> LeeF_: just replied to ericp re. http://demo.openlinksw.com/rdfbrowser/?uri=http%3A%2F%2Fwww.w3.org%2F2008%2FTalks%2F0311-swGathering%2F

21:54:07 <ericP> tx kidehen

21:54:37 <kidehen> LeeF_: grant SPARQL_UPDATE privilege to user SPARQL for sponging (rdfization) to be enabled etc.. It's off by default

21:54:45 <LeeF_> kidehen: does it work with http://demo.openlinksw.com/sparql ?

21:54:56 <kidehen> Leef_: yes

21:54:59 <LeeF_> thanks

21:55:20 <kidehen> LeeF_: n

21:55:35 <kidehen> LeeF_: np, I meant :-)

21:55:48 <ericP> kingsly, i haven't seen your mail yet. should my query have worked at http://demo.openlinksw.com/sparql ?

21:56:00 <LeeF_> my test worked, ericP

21:56:09 <ericP> same stuff?

21:56:17 <LeeF_> against the inchi RDFa

21:56:24 <kidehen> ericP: yes, but the default option may not be sponging

21:56:35 <kidehen> ericP: look at the drop down options

21:56:36 <LeeF_> no idea if the RDFa in the slides themselves is right or not

21:56:36 <ericP> sponging?

21:56:50 <kidehen> ericP: our lingo for RDFization on the fly

21:56:57 <ericP> roger

21:56:59 <LeeF_> i ran mine with "Retrieve all missing remote RDF data that might be useful"

21:57:09 <ericP> so perhaps there's no well-formed RDFa in the slide

21:57:10 <ericP> s

21:57:13 <LeeF_> right

21:57:16 <LeeF_> quite likely? :)

21:57:18 <kidehen> ericP: so we grab RDF, but if we don't get RDF our RDFization cartridges kick in

21:58:57 <kidehen> both: popping out (driving home) etc.

22:24:58 <danbri>http://sundaygang.com/dave/marcCanterInterview.mp3

22:25:00 <dc_swig> M: http://sundaygang.com/dave/marcCanterInterview.mp3 from danbri

22:25:16 <danbri> M:Proof that Marc Canter and Dave Winer aren't the same person

22:25:17 <dc_swig> Added comment M1.

22:25:36 <Arnia> danbri: do we have voiceprint identification?

22:25:37 <phenny> Arnia: 21:23Z <sbp> tell Arnia good luck for tomorrow!

22:36:34 <Arnia> phenny, tell sbp thanks and that I'll try and report on the event whilst I'm there (or failing that, on the train home)

22:36:34 <phenny> Arnia: I'll pass that on when sbp is around.

22:37:05 <dajobe> LeeF_: I ripped out RDFa via xslt since I couldn't deal with cceil license

22:58:56 <LeeF_> dajobe: ah, thanks. pity.

23:00:24 <ab_> my next projects modules are going to be named after characters

23:00:27 <ab_> zealot, heretic, etc

23:00:30 <ab_> WTFPL of course

23:00:57 <dajobe> another WTF moment

23:17:32 <ericP> kidehen, is there some way to get http://demo.openlinksw.com/sparql to re-read a source?

23:34:16 <tommorris>http://overtheair.org/blog/?page_id=20

23:34:17 <dc_swig> N: http://overtheair.org/blog/?page_id=20 from tommorris

23:35:11 <tommorris> N:|Over the Air (London, 4th and 5th April)

23:35:12 <dc_swig> Titled item N.

23:35:50 <tommorris> N: Mobile developer event taking place over 48 hours ("over night hacking" included)

23:35:51 <dc_swig> Added comment N1.

23:36:10 <tommorris> N: Perhaps of interest to SemWebbies and HTMLies too

23:36:11 <dc_swig> Added comment N2.

23:36:25 <tommorris> N: SIGNUP IS LIVE NOW DAMMIT, TAKE OFF EVERY ZIG etc.

23:36:27 <dc_swig> Added comment N3.