Semantic Web Interest Group IRC Chat Logs for 2005-04-19

06:12:17 <lilo> [Global Notice] Hi all. We're experiencing lag in Europe and will need to rehub. Mention this as needed to EU users as they come back.

06:31:37 <lilo> [Global Notice] Hi all. Recapping, we experienced European hubbing problems a few minutes ago, and we've reconfigured our hubbing to resolve the issue. Thanks for your patience and understanding, and thank you for using freenode!

09:12:19 <sh1mmer> good morning worl

09:12:20 <sh1mmer> d

09:23:14 <libby> heya sergio!

09:31:23 <myriapod> hi libby!

09:31:55 <libby> how are you?

09:32:10 <danbri> hi sergio

09:32:38 <myriapod> fine! i'll fly to dublin tomorrow

09:32:42 <myriapod> hi danbri

09:32:56 <libby> oh nice :)

09:33:21 <myriapod> i'm trying to follow the gargonza hacking

09:33:59 <myriapod> alberto completeted that php5 stuff i started there

09:34:15 <myriapod> perhaps i could help with the php4 backporting if it's needed

09:36:55 <libby> sounds like a plan...I dunno whether people use php4 or 5 or what...

09:38:17 <myriapod> i think 5 is cleaner but morten asked for php4

09:40:58 <libby> ah right

09:41:08 <libby> sounds like it would be useful then :)

09:42:00 <myriapod> anyway... if there's any php or perl stuff to do feel free to ask

09:42:25 <libby> thanks!

09:45:56 <myriapod> oh you uploaded more pics, checking them out :)

09:46:27 <libby> yeah, not that good

09:46:34 <libby> I bought a new camera though, having whinged about my dead one the whole time

09:46:43 <myriapod> :)

09:46:52 <sh1mmer> libby that reminds me, the search on swordfish/photos don't work :(

09:47:20 <libby> ah yeah

09:47:21 * danbri has a new camera too; time to revisit my ageing thumbnail scripts

09:47:35 <libby> my camera is better though

09:47:39 <sh1mmer> libby I was trying to find the photos of when you came up to newcastle

09:47:49 <sh1mmer> libby whats the new one?

09:48:27 <libby> http://www.amazon.co.uk/exec/obidos/ASIN/B0002XNPPM/qid=1113904087/sr=1-1/ref=sr_1_0_1/026-8470586-2731639

09:48:32 <myriapod> i need to buy one too, what model did you two chose?

09:48:51 <libby> if you have firefox, best photos interface is now http://swordfish.rdfweb.org/discovery/2004/12/icalphotos/2005ical.xml

09:48:55 <libby> (for mine)

09:49:23 <libby> I got a smaller version of my old one sergio - danbri got a similar version with a huuuge screen and 5 MP

09:49:43 <sh1mmer> yeah the exilim are nice

09:49:46 <danbri> s/similar/better/

09:49:56 * danbri ducks

09:49:59 <myriapod> danbri: :D

09:50:02 <libby> humph

09:50:09 <libby> pixels aren;t everything you know

09:50:19 <sh1mmer> its what you do them that counts

09:50:21 <sh1mmer> ha!

09:50:31 * libby has a tiny camera and is fine with that

09:50:33 <libby> :)

09:50:37 <danbri> :)

09:50:39 <libby> eeexactly!

09:50:51 <danbri> the best thing is the huge screen, not the megapixelage

09:50:56 <sh1mmer> yeah, mine is tiny. Maybe I should be worried that it zooms inside

09:51:12 <sh1mmer> danbri thats true, I really want one of those ickle Sonys which are 90% screen

09:51:45 <libby> there were some big cameras at sparqlthingy: http://swordfish.rdfweb.org/photos/2005/04/12/2005-04-12-Pages/Image39.html

09:52:28 <sh1mmer> my gf is a pro photographer and I shudder everytime she talks about the one she wants

09:52:39 <danbri> what's the best rdf vocab around for talking about versions of images? thumbnails (shrunk) vs lossy/cropped squares etc., as in http://rdfweb.org/people/danbri/media/2005/04/unsorted/gl-rd-shops-20040418/sq-index

09:52:51 <sh1mmer> its about £6k, the 'realistic' one she wants is still about £1,200

09:53:02 <libby> morten's done some stuff danbri...a bit

09:53:10 <libby> eek...

09:55:37 <libby> danbri, moren has something like orginal, thumbnail, something else

09:55:59 <danbri> I think a transformation event log would be useful

09:56:12 <danbri> with md5 and/or sha hash of content

09:57:53 <CaptSolo> sh1mmer: about the one she wants?

09:57:56 <CaptSolo> cameras?

09:58:18 <sh1mmer> yes?

09:58:21 <CaptSolo> 'realistic' being something approx Canon EOS 20D

09:58:28 <sh1mmer> yeah

09:58:33 <sh1mmer> it's one of the canon ones

09:58:52 <sh1mmer> she really liked my Dad's (non digital) EOS, it has autofocus which tracks your eye

09:58:58 <CaptSolo> the one she wants being EOS 1Ds Mark II

09:59:10 <sh1mmer> you just look in the view finder at what you want and it focuses on it

09:59:22 <CaptSolo> oh, isn't it same as mine? EOS 5 (whatever is the US name)

09:59:45 <CaptSolo> i don't really use that feature though

09:59:59 <CaptSolo> not even sure that will work when i wear glasses

10:00:07 <sh1mmer> CaptSolo not sure, I have an EOS 500 myself but I don't really keep up with the two of them (gf and father)

10:00:29 <CaptSolo> sh1mmer: what does your gf currently use?

10:00:38 <sh1mmer> some horrible old Nikon

10:00:45 <sh1mmer> no autofocus

10:00:46 <CaptSolo> (though good images are not about equipment)

10:00:57 <sh1mmer> she says she isn't going to change it until she gets a nice digital

10:01:15 <sh1mmer> although she sometimes borrows mine if she needs the autofocus for wildlife etc

10:01:27 <CaptSolo> :) manual camera is probably creating a different thinking about photos

10:01:44 <CaptSolo> i.e., you rarely think about where to focus if you have autofocus

10:01:56 <CaptSolo> [end of offtopic]

10:02:25 <CaptSolo> [almost] and what is yours?

10:02:35 <sh1mmer> my digital?

10:02:38 <sh1mmer> konica x31

10:02:59 <sh1mmer> it's about the best value camera for what you get

13:31:55 <libby> does anyone know how to add models to a joseki installation?

13:32:34 * libby can't find out how to do it at all :/

13:36:33 <AndyS> Libby - DanC said you had a question

13:36:56 <AndyS> <libby> does anyone know how to add models to a joseki installation?

13:37:26 <libby> oh, hello andys :)

13:37:46 <AndyS> Hi there

13:37:48 <libby> yeah, I just want to post some photo rdf stuff to an installation i have

13:37:55 <libby> of joseki

13:38:07 <AndyS> To add a new model you have to edit the config file (etc/joseki.n3) - you can't (yet!) add a new model from outside unless ...

13:38:18 <AndyS> ... you write a custom handler.

13:38:30 <libby> ah gotya

13:38:39 <AndyS> To add data into an existing model, you could enable the AddProcessor.

13:38:54 <libby> ah, ok, that would do for my purposes

13:38:55 <AndyS> Major caveat on security. Makes wikis look tame.

13:39:01 <libby> hehe

13:39:18 <AndyS> If you use Tomcat, you can probably uses its security features.

13:39:26 <danbri> the day our RDF stores full up with porn ads and chinese spam, is the day we stop worrying about SW adoption

13:39:46 <AndyS> It's junk blank nodes that are a really pain!

13:39:51 <libby> heh

13:39:59 <libby> enabling the addprocessor - in config?

13:40:02 <danbri> although weblog trackbacks are written in RDF, and being abused already

13:40:36 <crschmidt> anyone know if I can add two profile elements to a <head> element and use the features of both?

13:40:48 <AndyS> and DanBri - becareful what you wish for!

13:40:56 <danbri> space separated uris I think

13:41:11 <AndyS> Libby - yea - in the config file.

13:41:59 <AndyS> If you add to a model, it might need to be a database-backed one - otherwise the changes are only in-memory.

13:42:07 <libby> righty

13:42:22 <libby> ok, I have joseki:hasOperation joseki:BindingAdd

13:42:44 <AndyS> Spot on.

13:42:46 <libby> is there a way to post stuff to a url to add things?

13:42:54 <libby> or am I barking up the wrong tree...?

13:43:17 <AndyS> POST http://host/model?op=add

13:43:26 <libby> ah

13:43:31 <libby> many thanks andys, appreciate it

13:44:09 <crschmidt> danbri: thanks

13:44:24 * crschmidt just added some GRRDL to http://crschmidt.net/blog/

13:45:10 <AndyS> No probs: I'd point at the doc on SF but SF is rather slow at the moment.

13:45:30 * DanC checks crschmidt /blog/ with http://www.w3.org/2003/11/rdf-in-xhtml-demo

13:45:34 <libby> ah, must have been looking in the wrong place for docs

13:45:38 <crschmidt> worked here

13:45:51 <libby> (thanks for getting andys for me danc :)

13:45:59 <DanC> cheers

13:46:20 <DanC> crschmidt, I see lat/long and title

13:46:25 <crschmidt> DanC: yep

13:46:41 <crschmidt> DanC: you were asking for SPARQL in julie, it's not in the IRC yet, but SPARQL interface at http://crschmidt.net/julie/sparql , if you want to have a play

13:49:26 <DanC> hmm... fun... seems to use (s p o) syntax...

13:49:55 <crschmidt> yeah, it's using the most recent draft

13:50:18 <crschmidt> nothing's been published with the turtle syntax yet

13:51:35 <AndyS> Draft should change very soon now :-)

13:52:51 <crschmidt> Yeah, so I've heard, but until it does, there's not going to be Redland support for it, and I follow the latest Redland release.

13:53:16 * DanC added an entry to http://esw.w3.org/topic/DawgShows

13:53:42 <DanC> "supports most of latest [WWW] TR syntax " is a risky thing to write :)

13:53:51 <AndyS> Hmm - thought Dave has stuff working

13:54:05 <dajobe> some queries don't crash

13:54:08 <DanC> working, yes, but not released, I don't think

13:54:31 <AndyS> Esp when refs the undated TR URL!

13:54:47 <crschmidt> heh, heh

13:54:55 <crschmidt> right, major difference between working and released

13:55:08 <AndyS> Release != CVS update ???

13:55:16 <crschmidt> no! :p

13:55:37 <crschmidt> my last experience in CVS tarballs resulted in repeated crashes in the query engine (which ire ported as bugs, and dajobe promptly fixed)

13:55:48 <crschmidt> but after that, I swore I'd stick to releases

13:56:07 <dajobe> rasqal's particularly changing a lot

13:56:17 <crschmidt> which only crash when I expect them to :)

13:56:34 <AndyS> I make sure I get green lines before check in. But not always possible I admit.

13:56:37 <crschmidt> More seriously, in the last 6 months, I haven't had any issues with the Redland releases other than my own build-stupidity

13:56:49 <AndyS> Junit - a mixed blessing

13:56:53 <DanC> "get green lines"?

13:57:23 <AndyS> Junit draws a nice green line on all tests pass - a big red line when there is one or more failure.

13:57:45 <DanC> ah... "The familiar JUnit green line turns red if any of the tests fail" -- http://sys-con.com/story/?storyid=48887&DE=1&eclipse=on

13:58:03 <AndyS> So here "red lines" does not mean over revving the engine

14:00:28 <AndyS> Been testing Jena 2.2 for the last day (insert comments on about Macs and IANA here)

14:07:05 * AndyS goes off to DAWG telecon

14:07:18 <libby> cheers andys!

14:09:48 <libby> darn, I really need "order by..."

14:15:00 <jeen_> jeen_ is now known as jeen__

15:06:35 <libby_> libby_ is now known as libby

16:23:10 <CaptSolo> dajobe - are you here?

16:23:19 <dajobe> hi

16:23:32 <CaptSolo> i wonder if bloggers.rdf is being generated as described in http://journal.dajobe.org/journal/archives/2004_05.html#001703

16:23:52 <dajobe> yes

16:24:32 <CaptSolo> it's a nice description, but i don't see any span class= as written in the article

16:25:06 <CaptSolo> so i wondered how your semantic weblogs page becomes a bloggers.rdf

16:25:29 <dajobe> look harder

16:25:32 <dajobe> nothing has changed

16:25:35 <CaptSolo> ok :)

16:26:24 <CaptSolo> ah, true

16:27:13 <CaptSolo> span class= is only used for the metadata that are not available from a href= tags

16:27:26 <CaptSolo> group or person names

16:28:24 <CaptSolo> thanks

16:29:22 <CaptSolo> btw - your description of how things work is the clearest explanation i've seen (though it's focused on getting the foaf file mainly)

16:29:37 <danbri> is it grddl'd?

16:30:19 * crschmidt points danbri at the link above, which says "I expect this process possibly could be GRDDLed." :)

16:30:29 * danbri was just reading it

16:30:35 <danbri> no need now, thx ;)

16:30:37 <CaptSolo> i've hard people askinghow to get the software runing planet rdf after they've read the explanation on planet rdf page

16:30:47 <CaptSolo> (obviously they should've found it themselves by then)

16:31:12 <danbri> i had some trouble w/ the planet aggregator, it got confused on revisiting sites. but i couldn't make a nice test case for failure.

16:31:16 <danbri> might try again...

16:31:29 <crschmidt> keep in mind that planetrdf doesn't use planet

16:32:04 <CaptSolo> btw planet-php aggregator is quite nice

16:32:12 <danbri> well, whatever the planetrdf code is called

16:32:14 <danbri> chumpologica

16:33:04 <DanC> the sidekick's web browser does something funky where lots of blogs (including plantrdf) get really skinny... one word per line

16:33:26 <dajobe> the data's all in the index.rdf

16:33:32 <dajobe> in fact we just xslt it to the web content

16:33:37 <DanC> ndw worked around it on his site when he got a sidekick, but I never did figure out what the issue is

18:19:11 <crschmidt> dajobe: Redland GRRDL Parser hasn't been released yet, right? There's no release of Raptor with it?

18:19:36 <dajobe> right

18:19:50 <crschmidt> good, just checking

19:45:52 <marc_> redland quewstion: has anyone tried to add a timestamp to every statement yet?

19:46:18 <danbri> cheaper to associate a timestamp with a redland context, i think

19:46:25 <danbri> though i'm guessing

19:47:06 <marc_> yeah

19:47:09 <marc_> nice thing about Class::RDF

19:47:17 <marc_> every node/statement had a created timestamp

19:48:46 <danbri> i could see that being handy

19:48:52 <marc_> yeah very handy

19:49:01 <danbri> maybe a utility function or two could clone that functionality?

19:49:04 <marc_> gunna hack it in on a per object basis for now

19:49:05 <marc_> nod

19:49:13 <marc_> but not at the per-node, per-statemnet level

19:49:32 <marc_> ->last_modified ->created for object

19:52:01 <marc_> freakin 100 degrees out

19:52:19 <crschmidt> I was going to say that that's the kind of thing redland contexts are perfect for

19:52:34 <marc_> yeah so just plop it in the context?

19:52:42 <marc_> i was actually planning on *using* the context tho

19:52:57 <marc_> to say where the source data came frm

19:53:41 <crschmidt> m.append(statement, "%s,%s" % (sourceuri, timestamp))

19:53:51 <marc_> nod

19:53:55 <crschmidt> in other words, make the context be a string with both

19:53:58 <marc_> yeah

19:53:59 <marc_> good call

19:54:00 <crschmidt> kind of hackish

19:54:08 <crschmidt> but better than nothin

19:54:13 <marc_> that covers sstatements

19:54:14 <crschmidt> might be better to put timestamp first

19:54:24 <crschmidt> then you can just split on ,[limit1]

19:54:43 <crschmidt> since you know there's never going to be a comma in a timestamp, but there might be in a uri

19:54:46 <marc_> good call

19:57:05 * crschmidt is trying to work out how to create a redland Statement with NULLs in PHP

19:58:52 <crschmidt> it seems that Python can just use a None as a node in librdf_new_statement_from_nodes, but PHP NULLs don't seem to work

22:48:35 <crschmidt> Anyone using Redland+PHP?

22:48:35 * DanC hunts for XRI docs...

22:48:35 <DanC> found http://www.oasis-open.org/apps/group_public/download.php/11857/xri-intro-V2.0-wd-04.pdf ...

22:48:35 <DanC> which bears the "Location" http://docs.oasis-open.org/xri/xri/V2.0 which, by supreme irony, is 404

22:48:35 <crschmidt> ^witw crschmidt

22:48:35 <julie> 45 days, 9 hours, 5 minutes, 46 seconds ago, crschmidt was at long: -71.4369, lat: 42.9813

22:48:35 <crschmidt> guess I need to build in an updating interface

22:48:35 <DanC> hmm... ""

22:48:35 <DanC> er... "The purpose of this TC is to define an URI-compatible identifier scheme and resolution protocol for abstract identifiers, i.e., identifiers that are location-, application-, and transport-independent, and thus can be shared across any number of domains and directories."

22:48:35 <DanC> hmm... xri-intro says "Independent of a specific domain, authority, or owner of a resource at a particular point in time."

22:54:32 <DanC> [[

22:54:34 <DanC> A much better solution would be to assign the resource “govdoc.pdf” an identifier that never needs to change or be reassigned. This can be accomplished using a fully persistent XRI such as the following:

22:54:37 <DanC> xri://@!9990!A58F!1C3D/!2495

22:54:37 <DanC> ]]

22:54:49 * crschmidt wonders how such URIs are determined

22:55:10 <crschmidt> seems kind of non-human readable, easy to typo, that kind of stuff

22:55:27 <DanC> phpht. it can also be accomplished by using http://9990_A58F_1C3D.xri.org/!2495

22:58:15 <DanC> [[ Once we have reached the public XRI authority @example.org*agency*department, it can switch to internal delegation ]]

22:58:24 <DanC> sounds like the apache InternaRedirect mechanism.

23:01:13 <DanC> [[ 3.5 Metadata Identification ]] sounds an awful lot like URCs, which have pretty much turned into RDF

23:02:21 <DanC> [[ Because there will always be new and better ways to identify resources, the “X” in “XRI” reflects the same design principle as it does in “XML”: extensibility-by-design.]] chuckle.