Semantic Web Interest Group IRC Chat Logs for 2003-02-18

01:21:46 <soccos|away> soccos|away is now known as soccos

02:20:29 <burtonator> danbri: you around?

02:22:01 <burtonator> anyone know what was used to layout this?

02:22:04 <burtonator>http://rdfweb.org/foaf/examples/edd-foaf.jpeg

02:22:05 <dc_rdfig> A: http://rdfweb.org/foaf/examples/edd-foaf.jpeg from burtonator

02:22:09 <burtonator> what graphic program

02:22:09 <burtonator> ?

02:22:14 <burtonator> A:|FOAF layout example

02:22:14 <dc_rdfig> Titled item A.

02:22:28 <burtonator> I am pretty sure it isn't isaviz

02:24:35 <burtonator> looks like foafcorp

02:25:19 * AaronSw downloads

02:25:49 <AaronSw> oh, that's damien's OS X Java RDF editor

02:26:24 <burtonator> ok... URL?

02:26:28 <AaronSw> .google RDFAuthor

02:26:29 <datum> RDFAuthor: http://rdfweb.org/people/damian/RDFAuthor/

02:26:50 <burtonator> sweet

02:27:07 * burtonator needs a portable and high performance graph layout package

02:27:20 <burtonator> I don't want to write one either!

02:27:40 <Mutiny> ew

02:27:52 <Mutiny> make those arrows hit the edge of the node not the center, good lord

02:28:39 <burtonator> yeah... agreed

02:28:44 <burtonator> does it not do automatic layout?

02:29:02 <AaronSw> it has a layout button

02:29:12 <AaronSw> if you click it a lot things eventually work out

02:29:24 <AaronSw> OmniGraffle's layout engine is even nicer

02:30:04 <Mutiny> ahh omnigraffle

02:30:36 <burtonator> I find it strange that there isn't an OSS portable and high performance graph layout tool...

02:31:57 <burtonator> of course they will be hella old anyway

02:32:05 <burtonator> maybe I should write one that does SVG output

02:33:12 * burtonator builds graphviz from source

02:34:00 <burtonator> the fact that Mozilla doesn't have SVG is really bad

02:34:26 <Mutiny> i think there are a few

02:34:34 <Mutiny> nothing with layout engines as nice as graphviz though

02:35:32 <Mutiny> gef.tigris.org is nice

02:36:33 <burtonator> the layout engine in graphviz is dot

02:36:38 <burtonator> it is very simple

02:36:43 <burtonator> maybe I will port it to java

02:36:57 <burtonator> can't be that hard

02:38:08 <burtonator> death to swing!

02:38:53 <burtonator> this is essentially what isaviz does

02:39:03 <burtonator> uses dot and then gets the paths and then builds that into an SVG

03:13:10 * burtonator compiles RDFAuthor for Java

03:15:34 <burtonator> it seems to work but when I import an RDF file it doesn't draw it

03:18:50 <burtonator> AaronSw: was your RDF graph in OmniGraffle automatically generated or did you have to edit it by hand

03:19:00 <burtonator> ... this one

03:19:03 <burtonator> ... http://blogspace.com/rdf/omnigraffle.gif

03:19:55 <AaronSw> I edited it by hand but it can import dot files

03:20:33 <AaronSw> and you can use rdf2dot

03:21:38 <burtonator> ok...

03:21:41 <burtonator> I will do that when I get home

03:21:45 <burtonator> ... but my OSX box sux :(

03:30:13 <burtonator> why are people using using "web" XML prefixes for RDF data now?

03:31:45 <burtonator> AaronSw: rdf2dot requires a digraph RDF schema?

03:31:47 <burtonator> that is dumb

03:32:14 <AaronSw> what do you mean?

03:32:32 <AaronSw> oh, directed graph

03:32:37 <burtonator> yeah

03:32:41 <AaronSw> hm, rdf2dot is the wrong thing. i'm trying to remember the name

03:32:50 <burtonator> I don't see why it can't take raw RDF input

03:33:37 <burtonator> if you could remember it I could appreciate it... I don't want to spend an hour writing my own XSL

03:34:34 <AaronSw> well, http://www.w3.org/RDF/Validator/ will work

03:35:00 <burtonator> ok... I am worried I will DoS them

03:35:08 <burtonator> it took 30 minutes for isaviz to render this ;)

03:35:24 <AaronSw> source is at http://dev.w3.org/cvsweb/java/classes/org/w3c/rdf/examples/ARPServlet.java

03:37:01 <burtonator> thanks... I forgot that the validator provided this option

03:42:56 <soccos> soccos is now known as soc|away

03:43:28 <burtonator> AaronSw: you get any of that snow?

03:43:44 <AaronSw> no, everyone else took ours

03:44:52 <burtonator> yuk... this servlet isn't threadsafe!

03:51:52 <soc|away> soc|away is now known as soccos

03:51:54 * burtonator says screwit and writes his own rdf2dot.xsl that works with his RSS output

03:52:58 <AaronSw> heh

03:56:11 <soccos> soccos is now known as soccos|away

05:25:02 <burtonator> oh great... google and blogger on rss-dev

08:41:05 <minddog> minddog is now known as minddogZZ

08:51:19 <Jibbler|sleep> Jibbler|sleep is now known as Jibbler

09:24:52 <arnarl> hi

10:15:59 <golbeck> golbeck is now known as golbeck_zzz

11:21:14 <dmiles> wb

11:21:44 <zool> re dm

11:32:13 <dajobe>http://www.semanticwebjournal.org/

11:32:14 <dc_rdfig> B: http://www.semanticwebjournal.org/ from dajobe

11:32:21 <dajobe> B:|Journal on Web Semantics

11:32:21 <dc_rdfig> Titled item B.

11:32:27 <dajobe> B:*not* Semantic Web

11:32:27 <dc_rdfig> Added comment B1.

11:32:43 <dajobe> B:has an online version

11:32:43 <dc_rdfig> Added comment B2.

11:33:10 <dajobe> B:|Elsevier Journal on Web Semantics

11:33:11 <dc_rdfig> Titled item B.

11:33:29 <dajobe> B2:has an online version for 2003 only; then locked down?

11:33:29 <dc_rdfig> Replaced comment B2.

11:37:27 <dajobe> B:(petty) needs a bit of a spell check

11:37:27 <dc_rdfig> Added comment B3.

11:58:46 <minddogZZ> minddogZZ is now known as minddog

12:07:39 <dajobe>http://www.wiwiss.fu-berlin.de/suhl/bizer/d2rmap/D2Rmap.htm

12:07:39 <dc_rdfig> C: http://www.wiwiss.fu-berlin.de/suhl/bizer/d2rmap/D2Rmap.htm from dajobe

12:07:47 <dajobe> C:|2R MAP - Database to RDF Mapping Language and Processor

12:07:47 <dc_rdfig> Titled item C.

12:08:26 <dajobe> C:with an LGPLed implemenation of the processor in Java with Jena

12:08:26 <dc_rdfig> Added comment C1.

12:08:58 <dajobe> C:FROM access/mysql TO RDF/XML or N3 or N-Triples

12:08:58 <dc_rdfig> Added comment C2.

12:24:31 <dajobe>http://kaon.semanticweb.org/alphaworld/reverse/view

12:24:31 <dc_rdfig> D: http://kaon.semanticweb.org/alphaworld/reverse/view from dajobe

12:24:47 <dajobe> D:|KAON REVERSE GUI app for mapping RDF and relational databases (early prototype)

12:24:48 <dc_rdfig> Titled item D.

12:25:24 <dajobe> D:based on/working with other apps such as [http://sourceforge.net/projects/hmafra|hMAFRA]

12:25:24 <dc_rdfig> Added comment D1.

12:35:54 <Wack> hi

12:57:55 <danbri> hi Wack

13:00:21 * danbri admires http://www.xs4all.nl/~wackin/angelite-editor-recursive.png

13:00:53 <Jibbler>http://www.visualthesaurus.com/index.jsp

13:00:54 <dc_rdfig> E: http://www.visualthesaurus.com/index.jsp from Jibbler

13:02:05 <Wack> hehe

13:02:20 <Wack> it lacks the toolbar though :]

13:02:29 <Wack> add, edit, remove etc

13:02:47 <danbri> very nifty to be able to look inside rdf datasources, for debugging etc at least

13:02:51 <Wack> but I just found out that when I open an rdf file, all changes are flushed automatically

13:03:17 <Wack> so I'm working around/on that

13:05:58 <danbri> there is some moz method for flushing changes back to the on-disk version...

13:06:14 <danbri> i wrote a note to a mozilla-rdf recently looking for just this...

13:06:23 <danbri> ...wanted to reverse engineer an SChema for the moz rdf data formats

13:07:41 <Wack>http://lxr.mozilla.org/seamonkey/source/rdf/base/idl/nsIRDFRemoteDataSource.idl

13:07:41 <dc_rdfig> F: http://lxr.mozilla.org/seamonkey/source/rdf/base/idl/nsIRDFRemoteDataSource.idl from Wack

13:07:47 <Wack> FlushTo() etc

13:08:04 <danbri> where's the best url for mozilla-rdf archives, nowadays?

13:09:02 <Wack> ah great, the in-memory-datasource can be purged to disk

13:09:12 <Wack> mozilla-rdf mailinglist archives?

13:09:23 <danbri> yes (and newsgroup, gated)

13:09:31 <danbri> hmm i can't find the msg I was sure I'd sent out! oh well...

13:09:58 <danbri> basically, people want an rdf vocab for bookmarks and related stuff, and know mozilla has relevant work, but don't know the schema

13:10:18 <danbri> would it be possible to use your code to dump out a list of classes and properties from a bookmarks db?

13:10:29 <danbri> or i guess, just a 'save as RDF' for some bookmark data would work

13:10:47 <danbri>http://www.geocrawler.com/archives/3/125/2003/

13:10:48 <dc_rdfig> G: http://www.geocrawler.com/archives/3/125/2003/ from danbri

13:10:57 <danbri> G:|mozilla-rdf archives, at geocrawler

13:10:57 <dc_rdfig> Titled item G.

13:11:53 <Wack>http://groups.google.com/groups?&group=netscape.public.mozilla.rdf

13:11:54 <dc_rdfig> H: http://groups.google.com/groups?&group=netscape.public.mozilla.rdf from Wack

13:12:14 <Wack> H:|mozilla-rdf archives, at google

13:12:14 <dc_rdfig> Titled item H.

13:12:15 <Wack> :]

13:12:30 <Wack> ah nice

13:12:49 <danbri> :)

13:13:03 <danbri> pity mozilla is still using rdf:instanceOf instead of rdf:type

13:13:16 <danbri>http://www.xs4all.nl/~wackin/angelite-editor-recursive.png

13:13:25 <dc_rdfig> I: http://www.xs4all.nl/~wackin/angelite-editor-recursive.png from danbri

13:13:25 <danbri> can you describe that for the webloggy thing?

13:13:37 <danbri> (output is at http://rdfig.xmlhack.com if you'd not seen this...)

13:13:52 <Wack> yeah, I saw it: 14:12 <Wack> ah nice

13:14:07 * danbri remembers the url is in the /topic :)

13:15:26 <Wack> I: Screenshot of a work-in-progress sitemap editor using a Mozilla XBL rdf editor/viewer widget

13:15:27 <dc_rdfig> Added comment I1.

13:16:21 * danbri wonders is Wack has seen http://brownsauce.sourceforge.net/ (re generic rdf browser designs)

13:17:34 <Wack> I: It works by creating a crude reified model of the input model, crude because of the mozilla rdf-tree that displays it and my not-so-complete understanding of it :]

13:17:34 <dc_rdfig> Added comment I2.

13:18:18 <Wack> danbri: no, I havent, looking now :]

13:18:19 <danbri> I:Why did you need to reify?

13:18:19 <dc_rdfig> Added comment I3.

13:18:36 <danbri> I:Does it use Moz's XUL templates?

13:18:36 <dc_rdfig> Added comment I4.

13:18:39 <danbri> questions questions!

13:20:10 <Wack> yes, I create a seperate datasource that holds the information you see in the screenshot, based on the input datasource

13:20:24 <Wack> I try to find the resources that are subject of, but not object of statements and go walk the model from there

13:21:04 <Wack> all statements with the same subject end up in a rdf sequence

13:21:19 * danbri nods

13:22:47 <Wack> and each statement, represented as an anonymous resource, (with rdf:object, rdf:subject etc statements about it, that's why I called it 'reified') has a statementsAboutObject predicate that targets another sequence

13:22:58 <danbri> do you have any more info about the sitemaps schema you are exploring?

13:23:05 <Wack> this way the rdf xul template can easily nest

13:23:10 <danbri> cool

13:23:22 <danbri> that makes sense

13:23:23 <Wack> as long as you have an initial root to give it

13:24:07 <danbri> does Mozilla really have a notion of 'anonymous resource'? ie if i create a resource without specifying a uri, and it makes up some ID itself, can I tell later that this was a mozilla-generated id?

13:24:28 <danbri> (sorry to grill you, am just trying to catch up on status of mozilla rdf stuff)

13:24:42 <Wack> it tracks cyclic references by keeping a history of which resources are already explored

13:27:01 <Wack> yes, the nsIRDFService interface has ::GetAnonymousResource() and bool ::IsAnonymousResource(in nsIRDFResource aResource) methods

13:27:21 <danbri> oh sweet, moz isn't as out of date as i'd feared! :)

13:27:45 <danbri>http://ilrt.org/discovery/2000/08/bized-meta/

13:27:45 <dc_rdfig> J: http://ilrt.org/discovery/2000/08/bized-meta/ from danbri

13:27:54 <danbri> J:|Biz/ed RDF Metadata Testbed

13:27:54 <dc_rdfig> Titled item J.

13:28:15 <danbri> J:"At the time of writing, there is no consensus around the best form for an RDF sitemaps vocabulary. For historical interest, here is a screenshot of an early RDF sitemap implementation: the first release of Mozilla (Netscape 5)."

13:28:15 <dc_rdfig> Added comment J1.

13:28:34 <danbri> J:"That format was based on the (over) use of nc:child arcs to indicate structure. As such, the sitemap format didn't preserve document ordering of items described sequentially. Other subequent variants have experimented with the use of rdf:Seq."

13:28:34 <dc_rdfig> Added comment J2.

13:29:09 <danbri> J:Preserving document order is in my experience quite critical to having a sitemap file format that users can understand. In practice with RDF, this means using rdf:Seq.

13:29:09 <dc_rdfig> Added comment J3.

13:29:19 <Wack> the sitemap is for a webserver framework we're working on

13:29:32 <danbri> J:Just as we do in RSS 1.0 feeds.... even though there will always be lots of other interesting ways to (re)order the data.

13:29:33 <dc_rdfig> Added comment J4.

13:30:00 <Wack> and we use sequences as well

13:30:12 <danbri> Wack, sounds interesting... whenever you've got more info on it, it'd be great to drop an URL in here, or to www-rdf-interest@w3.org

13:30:32 <Wack> I: I create a seperate datasource that holds the information you see in the screenshot, based on the input datasource

13:30:33 <dc_rdfig> Added comment I5.

13:30:36 <danbri> ...RDF sitemaps are so long overdue, its about time folk agreed on a basic format for them

13:30:36 <Wack> I: I try to find the resources that are subject of, but not object of statements and go walk the model from there, all statements with the same subject end up in a rdf sequence

13:30:36 <dc_rdfig> Added comment I6.

13:30:50 <Wack> I: Each statement, represented as an anonymous resource, (with rdf:object, rdf:subject etc statements about it, that's why I called it 'reified') has a statementsAboutObject predicate that targets another sequence, this way the rdf xul template can easily nest

13:30:50 <dc_rdfig> Added comment I7.

13:30:59 <danbri> logger_1, pointer?

13:30:59 <danbri> See http://ilrt.org/discovery/chatlogs/rdfig/2003-02-18#T13-30-59

13:31:26 <danbri> I:See [http://ilrt.org/discovery/chatlogs/rdfig/2003-02-18#T13-30-59|#rdfig ig discussion of 2003-02-18] for context.

13:31:26 <dc_rdfig> Added comment I8.

13:32:48 <Wack> hehe, you guys have quite an irc setup here :]

13:33:32 <danbri> I:The other thing Mozilla did in their v5 sitemaps implementation was split large site descriptions across multiple files. eg [http://ilrt.org/discovery/2000/08/bized-meta/webcat.rdf|webcat.rdf] and [http://ilrt.org/discovery/2000/08/bized-meta/bized.rdf|bized.rdf] in my 1998-era example data.

13:33:32 <dc_rdfig> Added comment I9.

13:33:41 <danbri> yeah, the irc tools are nifty :)

13:34:09 <danbri> I:The moz 5 linking was done like this: <child rdf:href="http://www.ilrt.bris.ac.uk/discovery/demos/rdf/bized.rdf#bized" name="listings by subject"/>

13:34:09 <dc_rdfig> Added comment I10.

13:35:14 <danbri> I:...but that left things unclear; client tools couldn't be sure when the target was HTML, when it was parsable as RDF. In [http://xmlns.com/foaf/0.1/|FOAF] we use rdfs:seeAlso instead, as a strong hint that the referenced document may be readable as more RDF data.

13:35:14 <dc_rdfig> Added comment I11.

13:35:33 <danbri> Wack, do you have a design for linking amongst overlapping sitemap documents?

13:36:59 <Wack> no, we hardly have any designs, we're only in the early stages of testing the framework itself

13:37:28 <Wack> a 'page' in a 'sitemap' in our system is usually an anonymous node

13:37:53 <Wack> so if we have 2 sitemaps in a model that share pages, we simply target the same nodes

13:38:05 <Wack> that only works if the sitemaps are in the same model ofcourse

13:38:40 <Wack> but we intend to have 1 model for sites that span over multiple sitemaps

13:38:44 <danbri> yes... otherwise you need algorithms to merge together anonymous nodes for cases where you can infer that the two nodes are representations of the same thing-in-the-world

13:38:46 <mariyo> Thanks a lot lars marius. I really tried. i even ended up with some "german" during the import, maybe that's why japanese eudora complained. I had some problems with oracle on a japanese machine today. we really need to work with these international issues :)

13:39:03 <danbri> ...which is interesting...

13:39:20 * danbri waves to mariyo, wonders if that was meant for #topicmaps -- hi anyway!

13:39:59 <danbri> <Wack> but we intend to have 1 model for sites that span over multiple sitemaps

13:40:05 <danbri> for large sites, i think that's critical...

13:40:06 <mariyo> hi danbri, yes we are moving our mail lists. i am really tired was reading your discussion with interest.

13:40:38 * danbri was just thinking about topicmaps, as the old old mozilla sitemaps stuff uses <Topic> rdf class to represent structure...

13:40:44 <danbri> eg http://ilrt.org/discovery/2000/08/bized-meta/webcat.rdf

13:41:02 <mariyo> so maybe this is a good coincidence :)

13:41:02 * danbri wonders what timezone you're in...

13:41:05 <danbri> :)

13:41:57 <mariyo> it's now 10:46 and my kids should be in bed !

13:42:19 <mariyo> pm that is

13:42:26 * danbri guessed!

13:43:07 * danbri wonders what to do to help move rdf sitemaps discussion along... its never really taken off in a big way, but shouldn't be too hard

13:43:44 <danbri> I think mozilla expertise is critical, and many rdf folk are mozilla sympathisers but find it hard to contribute... so Wack's tools are pretty handy / useful in getting that moving again

14:43:11 <pixel_2> pixel_2 is now known as pixel

14:43:19 <zoyd> what converts the xml generated by dailychump into html ?

14:44:25 <dajobe> xslt

14:45:10 <zoyd> xslt support in apache ?

14:45:44 <dajobe> I think so

14:46:08 <dajobe> the chump sources are available, follow the links

14:48:48 <zoyd> when a .xsl URI is accessed i'm not getting it styled, just the xml, does that mean the apache setup doesn't have xsl enabled ?

14:49:14 <DanCon> I wouldn't expect the .xsl to get styled.

14:49:55 <zoyd> then ?

14:50:52 <zoyd_> sorry.

14:51:10 <zoyd_> zoyd_ is now known as zoyd

14:51:16 <DanCon> there's a script that runs the .xsl over the .xml to make .html whenever the .xml changes, I think, zoyd.

14:53:35 <zoyd> ok.

15:02:17 <dajobe> actually I think it is done server-side by mod_xslt, mod_sablotron or saxon. I'm not clear on the details

15:09:27 <zoyd> aah no mod_xslt support from my web hosting service.

15:10:34 <danbri> i'd like to see a non-mod_xslt howto for chump setup... eg. with chump triggering commandline rebuild of pages...

15:10:47 <danbri> i might have a go, sometime (not this week...)

15:25:12 <jhendler> B:

15:25:12 <dc_rdfig>http://www.semanticwebjournal.org/

15:25:13 <dc_rdfig> Elsevier Journal on Web Semantics

15:25:14 <dc_rdfig> (1:dajobe) *not* Semantic Web

15:25:15 <dc_rdfig> (2:dajobe) has an online version for 2003 only; then locked down?

15:25:16 <dc_rdfig> (3:dajobe) (petty) needs a bit of a spell check

15:25:45 <jhendler> B: There will be an active online version - not just for 2003.

15:25:46 <dc_rdfig> Added comment B4.

15:26:21 <jhendler> B: in addition, will have part of the web site for submission of demos and content -- with short articles appearing in the print journal discussing these

15:26:22 <dc_rdfig> Added comment B5.

15:26:47 <jhendler> B: metadata for the journal is expected to be made available as well

15:26:47 <dc_rdfig> Added comment B6.

15:27:17 <jhendler> B: Stefan Decker, the Editor in Chief for the online part of the journal, would welcome your suggestions for the online journal

15:27:18 <dc_rdfig> Added comment B7.

15:28:06 <danbri>http://www.cfmentor.com/~faser/xulex/amazon_browser/index.html

15:28:06 <dc_rdfig> K: http://www.cfmentor.com/~faser/xulex/amazon_browser/index.html from danbri

15:28:19 <danbri> K:|An Amazon browser, using Mozilla XUL

15:28:19 <dc_rdfig> Titled item K.

15:28:29 <danbri> K:I've not studied the innards. Does it use RDF, SOAP, etc?

15:28:29 <dc_rdfig> Added comment K1.

15:28:33 <jhendler> B: see also the [http://www.elsevier.com/homepage/sac/050/semweb_cfp.html | Elsevier home page] for more info

15:28:34 <dc_rdfig> Added comment B8.

15:28:36 <danbri> K:Perhaps... both?

15:28:37 <dc_rdfig> Added comment K2.

15:29:28 <danbri> K:See [http://www.cfmentor.com/~faser/xulex/amazon_browser/amazonbrowser.xul|XUL source code] for details.

15:29:28 <dc_rdfig> Added comment K3.

15:29:41 <jhendler> B8: see also [http://www.elsevier.com/homepage/sac/050/semweb_cfp.html] for more info

15:29:41 <dc_rdfig> Replaced comment B8.

15:29:48 <danbri> K:Runs in Mozilla (my 1.3a version, anyways)

15:29:49 <dc_rdfig> Added comment K4.

15:32:43 <danbri> K:See also [http://www.cfmentor.com/~faser/xulex/amazon_browser/Amazon.js|Amazon.js] for the code that queries the amazon web service.

15:32:43 <dc_rdfig> Added comment K5.

16:05:27 <Wack> Wack is now known as _Wack

16:08:27 <_Wack> ha! I can now actually edit literals and (named) resource uri's in my editor :[

16:08:32 <_Wack> er :]

16:08:41 <_Wack> that is, of existing statements

16:15:32 <zoyd> is there any vocab to describe a city/town ?

16:20:42 <DanCon> can anybody else GET http://www.advogato.org/article/364.html ?

16:27:14 <_Wack> Waiting for...

16:27:20 <_Wack> looks dead :]

16:27:32 <zoyd> slashdotted perhaps.

16:28:10 * danbri can't GET it

16:28:39 * sandro can't either

16:29:13 <DanCon> thx for confirming the problem

16:30:49 <dajobe> it's working now

16:31:08 <dajobe> plus google cache had it too

16:32:55 * danbri reads http://www.advogato.org/article/364.html notes the Ogg-to-W3C suggestion...

16:35:13 <_Wack> _Wack is now known as Wack

17:01:49 <AaronSw> Patrick?

17:04:06 <PStickler> Hi Aaron. Am lurking from my mobile phone...

17:04:27 <AaronSw> Custom software or will my next phone have an IRC client?

17:05:41 <PStickler> If you choose to install one...

17:06:35 <AaronSw> neat. I know DanC uses an AIM2IRC bridge I wrote, and his phone's AIM client.

17:06:58 <PStickler> The problem is typing. OK to follow thrreads but not to say much

17:07:22 <PStickler> Which is great for loudmouths like me ;-)

17:07:27 <sbp`> you need a plugin-able keyboard...

17:07:29 <SethR> if we got context rooms, there is less to type

17:08:43 <PStickler> I'd love a palm like kbd with bluetooth conn, but all this is some what off topic, eh?

17:09:04 <sbp`> don't worry, it's a quiet day here on #rdfig... :-)

17:10:09 <AaronSw> IRC channels have interesting waves of total silence followed by intense discussion

17:11:28 <sbp`> [total silence (broken by me typing "[total silence (broken by me typing "[to... dangit

17:11:30 <SethR> hey that's almost a quotable

17:11:39 <PStickler> I don't think I would survive much intense discussion via my mobile phone

17:11:58 * Jibbler played with visualizing irc channels once

17:12:11 <SethR> if your phone could talk and listen, you coud very well with it, me thinks

17:12:40 <SethR> typeing doesn replace talking very well

17:12:41 <sbp`> Jim (assuming that's you): in SVG/Javascript? anything on'th Web?

17:13:06 <PStickler> Text to speech on a small device is no prob, but the other way... not yet

17:13:06 <Jibbler> i'm Paul.... why does everyone think I'm jim? did i pick a bad nick? :-/

17:13:40 <SethR> so Patric, why doesnt nokea make there phones talk ??????

17:14:12 <sbp`> Paul: my apologies, but yep. yep, you did :-) Jim Ley usually uses Jibber or JibberJim

17:14:24 <Jibbler> bah :)

17:14:34 <Jibbler> sorry about that

17:14:50 <sbp`> no problem. I should've done a WHOIS first

17:15:10 <Jibbler> it's my real nickname with some people, hence why i use it here (i normally use "Paul" but it's taken on this server already)

17:15:15 <Jibbler> .google jibble

17:15:15 <datum> jibble: http://www.jibble.org/

17:15:37 <SethR> why do people avoid chumping?

17:15:49 <Jibbler> i'm not familiar with chumping

17:15:53 <Jibbler> tbh

17:15:58 <PStickler> It's tough to get the 95+ pct needed for a solid product, but maybe sooner than you expect

17:16:01 <SethR>http://www.jibble.org/

17:16:02 <dc_rdfig> L: http://www.jibble.org/ from SethR

17:16:48 <SethR> L:| Homepage of Paul Mutton

17:16:48 <dc_rdfig> Titled item L.

17:17:34 <SethR> speech synthesis from text is fairly well off the shelf ... blind people use it all the time

17:18:04 <sbp`> L:haha: "7080000000 jibbles per day"--[http://www.jibble.org/jibblemeaning.php|What does jibble mean?]

17:18:05 <dc_rdfig> Added comment L1.

17:18:37 <Jibbler> i have to say, that chump thing doesn't look like the easiest thing to use :)

17:18:55 <Jibbler> (and i never was too keen on the user of notices)

17:19:00 <SethR> yep .. it should be far easier .. the syntax is a bit tedious

17:19:18 <Jibbler> tempts me to make my own sometimes :)

17:19:49 <SethR>http://rdfig.xmlhack.com/

17:19:50 <dc_rdfig> M: http://rdfig.xmlhack.com/ from SethR

17:20:04 <Jibbler> yeah, seen that

17:20:13 <SethR> M:|Where things go when they are chumped .. in case you didnt know already.

17:20:13 <dc_rdfig> Titled item M.

17:20:32 <Jibbler> i can imagine that site getting a fair few google hits

17:20:45 <SethR> yep

17:20:52 <sbp`> dude. the word "jibble" is everywhere. you lucky jibbler

17:21:28 * danbri realises jibbler is an irc bot hacker... http://www.jibble.org/pircbot.php

17:21:37 <Jibbler> E:| Visual thesaurus with cool three-dimensional graph navigation

17:21:37 <dc_rdfig> Titled item E.

17:22:28 <Jibbler>http://www.jibble.org/pircbot.php

17:22:28 <dc_rdfig> N: http://www.jibble.org/pircbot.php from Jibbler

17:22:44 <Jibbler> N:| PircBot Java IRC Bot API

17:22:44 <dc_rdfig> Titled item N.

17:22:54 <SethR> me thinks any mention of a uri anywhere in a comment should create a chump and bring the object to the foreground

17:26:47 <Jibbler> is a chump an individual entry on that page, then?

17:28:09 <SethR> i dont know .. dont think we have a detailed ontology for all that yet

17:28:12 <danbri> its more a verb than a noun, i think

17:28:16 <danbri> one chumps a url

17:28:43 <SethR> i think the chump would be the rss item

17:29:12 <Jibbler> .google what does noun mean?

17:29:13 <datum> what does noun mean?: http://www.jibble.org/jibblemeaning.php

17:29:16 <Jibbler> :o)

17:29:51 <SethR> got hoisted on your own patard

17:29:59 * sbp` is extraordinarily jealous

17:30:17 <danbri> interesting!

17:31:42 <Jibbler> blame dajobe.... i read his advice on getting good rankings in search engines ;)

17:32:23 <dajobe> that was pre-google IIRC

17:33:06 <Jibbler> probably

17:35:24 <Jibbler> still, it clearly does the trick

17:35:40 <Jibbler> first ten google results for my name actually point to me ;)

17:36:42 <AaronSw> .googlecount jibble

17:36:43 <datum> jibble: 2,680

17:36:59 <Jibbler> 'tis spreading :)

17:37:06 <Jibbler> .googlecount pircbot

17:37:08 <datum> pircbot: 3,240

17:37:23 <danbri> .googlecount xml

17:37:24 <datum> xml: 16,600,000

17:37:26 <danbri> .googlecount rdf

17:37:27 <datum> rdf: 1,890,000

17:37:32 <danbri> .googlecount rss

17:37:32 <datum> rss: 1,840,000

17:37:39 <danbri> .googlecount dublin core

17:37:41 <datum> dublin core: 415,000

17:37:47 <danbri> .googlecount foaf

17:37:47 <datum> foaf: 36,900

17:37:50 <Jibbler> well, you'd expect that, really

17:38:00 <danbri> .googlecount xslt

17:38:01 <datum> xslt: 904,000

17:38:22 <danbri> (foaf is a pre-existing acronym anyways...)

17:38:50 * danbri was just curious, not drawing any particular conclusions :)

17:39:04 <eikeon>

17:39:05 <eikeon>

17:39:05 <eikeon>

17:39:06 <SethR> .google pircbot

17:39:07 <datum> pircbot: http://www.jibble.org/pircbot.php

17:39:08 <Jibbler> woah :)

17:39:10 <eikeon>

17:39:15 <eikeon> .googlecount XML

17:39:20 <eikeon> Lots-o-whitespace... sorry

17:39:23 <Jibbler> heh

17:39:26 <eikeon> .googlecount XML

17:39:26 <datum> XML: 16,200,000

17:39:26 <dc_rdfig> Label XML not found.

17:39:33 <Jibbler> heh

17:39:45 <sbp`> wow, case sensitive

17:39:53 <eikeon> .googlecount xml

17:39:54 <datum> xml: 16,200,000

17:39:58 <sbp`> hmm

17:40:01 <danbri> .googlecount A:BotInteraction?

17:40:02 <sbp`> <danbri> .googlecount xml

17:40:02 <datum> A:BotInteraction?: 0

17:40:02 <sbp`> <datum> xml: 16,600,000

17:40:03 <dc_rdfig> Added comment A1.

17:40:15 <eikeon> Just lost 200k somewhere :)

17:40:30 <eikeon> Make that 400k

17:40:42 <sbp`> perhaps w3.org went down whilst Googlebot was hammering it

17:40:44 <jang> different bits of the database not being in sync.

17:40:53 <jang> .googlecount xml

17:40:53 <Jibbler> www2.google.com

17:40:53 <datum> xml: 16,200,000

17:41:02 <jang> see?

17:43:32 <SethR> jang, what database?

17:44:02 <danbri> google's server farm

17:44:32 <SethR> does that happen when Google dances ?

17:45:44 <SethR> .googlecount google dance

17:45:45 <datum> google dance: 237,000

17:48:00 <Jibbler> .googlecount mmm pie

17:48:00 <datum> mmm pie: 39,600

17:55:10 <eikeon> Now all we need is a server farm for finding RDF terms... put up a quick hack to count # properties etc of things I have spidered at: http://rdfschema.info/

17:56:23 <AaronSw> neat

17:57:19 <danbri> cool! added it to http://rdfweb.org/rweb/wiki/wiki?FoafVocab

17:57:51 <AaronSw> it's funny there are so many with zero

17:58:14 <danbri> "queries from the current RDF store" is a little vague

17:58:31 <danbri> eikeon, could you add a quick note on what the current store, er, stores!

17:58:49 <danbri> eg. is it an rdfs:seeAlso'ing scutter?

17:59:33 <eikeon> danbri: sure.

18:00:19 <eikeon> Yep... am only following rdfs:seeAlso links I originally added to scutter for FOAF data.

18:00:56 <eikeon> And is generated from the same store as all my other sites.

18:04:15 <eikeon> Looks like I have RDF from about 2700 URLs.

18:04:19 <danbri> cool

18:04:26 <danbri> so does the Semantic Web exist yet? ;)

18:05:48 <eikeon> Has anyone started a schema for sharing scutter info... or mostly sharing lists of rdfs:seeAlso... or like me and not yet sharing with other scutters in either direction.

18:06:11 <mortenf> foaf:Document?

18:06:51 <eikeon> mortenf: That caught my eye to... not sure what it is used for... danbri?

18:07:03 <mortenf> rdf:about=""?

18:09:55 <danbri> things that are documents, such as foaf documents, homepages, ...

18:10:05 <danbri> helpful to be clear that people aren't documents, for eg.

18:11:11 <eikeon> Fair enough.

18:12:00 <eikeon> mortenf: ?

18:12:14 <mortenf> do you need more than that (cf. the rss-dev discussion)?

18:12:44 <danbri> more than...?

18:12:54 <eikeon> Ah... I have not been following the rss-dev discussion.

18:13:16 <mortenf> doctyping: hard in rdf, but sometimes useful anyway - perhaps just a distinction between rdf/non-rdf?

18:13:38 <mortenf> s/hard/non-applicable/

18:13:45 <deus_x> Does there happen to be a schema anyone's worked on to describe HTTP headers?

18:14:43 <danbri> maybe (re subclassing Document), though am in no hurry there

18:15:26 <mortenf> deu; none that i've seen, but there's one for email headers...

18:16:14 <deus_x> mortenf: Hmm. Might have to look at that. So far, I've been pretending to have a schema, munging HTTP header names to camelCaps from hyphenated-words for property names

18:16:46 <mortenf> heh, hyphens work fine though.

18:17:24 <deus_x> For some reason I thought that was a problem. Can't seem to think why though. Probably be much simpler just to use the hyphenated header names. heh

18:18:40 <mortenf> but lowercase for properties seems like a good idea.

18:20:42 <SethR> eikeon, http://rdfschema.info/ is very kewl ... its almost the dictionary we had envisioned at SWAG ... 100 kudos !

18:31:42 <mortenf> deus_x: on http header schema (follow links): http://lists.w3.org/Archives/Public/www-rdf-interest/2002May/0025.html

18:32:24 <mortenf> heh, and re the typing of resources: http://lists.w3.org/Archives/Public/www-archive/2002Mar/att-0054/01-draft-palmer-resrep-type-00.txt

18:36:47 <deus_x> Trying to come up with other metadata to harvest from personal web traffic besides headers. Thinking maybe doing some RSS & FOAF autodiscovery too.

18:37:16 <mortenf> well, of course! :-)

18:50:19 * mortenf -> break

19:11:28 <Jibbler> Jibbler is now known as Jibbler|out

19:25:52 <eikeon> SethR: Thank you. I'll keep working on it when I have a chance. Also, any ideas or help anyone would like to forward along is welcome ;)

20:04:04 <golbeck_zzz> golbeck_zzz is now known as golbeck

20:12:52 <danbri> http://web.mit.edu/simile/

20:13:00 <danbri>http://web.mit.edu/simile/

20:13:00 <dc_rdfig> O: http://web.mit.edu/simile/ from danbri

20:13:09 <danbri> O:|Semantic Interoperability of Metadata and Information in unLike Environments (SIMILE project)

20:13:09 <dc_rdfig> Titled item O.

20:14:05 <Wack> unLike? :]

20:20:21 <Wack> http://haystack.lcs.mit.edu/projects/haystack/screenshots/ hmmm, looks nice :]

20:20:22 <dc_rdfig> P: http://haystack.lcs.mit.edu/projects/haystack/screenshots/ from Wack

20:36:34 <dajobe> if only there was more than screenshots to show for months and months...

20:36:49 <golbeck> golbeck is now known as trustbotbot

20:37:05 <trustbotbot> trustbotbot is now known as golbeck

20:38:06 * zool idly wonders how newsmonster development is progressing, and whether it makes good use yet of reputation/karma

20:39:18 <Morbus> i've got someone building a rating system onto ampheta.

20:39:25 <Morbus> someone's already done click counting, but not analysis

20:39:33 <Morbus> but i've got a random email suggesting that someone is building more.

20:40:51 <zool> (probabilistic development)++

20:41:48 <Morbus> is newsmonster good? i've not tried it yet.

20:42:50 <zool> i havent heard any practical things, and an awful lot of it seems to be coming soon ;)

20:42:57 <zool> intrigued by foaf support though

20:43:35 * zool tickles burtonator

20:44:42 <Jibbler|out> Jibbler|out is now known as Jibbler

20:50:26 <Wack> outs! :]

20:50:45 <Wack> btw, anyone here knows of any c++ rdfxml parsers/serializers?

20:52:08 <dajobe> how about mine

20:53:47 <Wack> redland?

20:53:58 <dajobe> raptor

20:54:12 <Wack> oh yeah, that would have been my second guess ;)

20:54:22 <Wack> (looking at sf project page)

20:54:32 <dajobe> raptor's the parser

20:57:19 * No memory leaks * Fast <- great! :]

20:57:57 <dajobe> I stand by that !

20:58:12 <dajobe> It did ~10% of the dmoz rdf with 0 bytes lost at the end

20:58:34 <dajobe> i've some stats somewhere...

20:59:31 <dajobe> there: http://journal.dajobe.org/journal/archives/2003_01.html#001136

20:59:55 <dajobe> 61K triples/sec, and that's from disk

21:14:32 <Jibbler> rofl at the "NO BUSH" photo http://www.commondreams.org/headlines03/0208-06.htm

21:22:58 <burtonator> zool: it isn't released yet

21:23:15 <burtonator> zool: and by coming soon I mean SOON (as in release early, release often)

21:33:40 <deus_x> Hmm. Need to hack brownsauce to browse an in-memory RDF model.

21:41:59 <Jibbler> so funny

21:41:59 <Jibbler>http://www.embarrassing.net/images/sonyericsson.jpg

21:41:59 <dc_rdfig> Q: http://www.embarrassing.net/images/sonyericsson.jpg from Jibbler

21:42:13 <Jibbler> Q:| Spot the difference

21:42:13 <dc_rdfig> Titled item Q.

21:44:11 <Morbus> hehe

21:59:27 <reagle> the new syntax spec says the attribute value of rdf:ID takes the xml:base declaration

21:59:56 <reagle> is there some definition of which rdf/xml attribute values take xml:base and which don't? (e.g., what are there xml types?)

22:00:57 <dajobe> there is no xml types

22:01:05 <dajobe> we're not doing DTDs or XWD, you know ;)

22:01:08 <dajobe> XSD

22:01:19 <dajobe> the defn is indirect

22:01:33 <dajobe> the rules tell you what to do with the value of each attribute

22:01:44 <dajobe> so for example rdf:ID makes a URI, indirectly

22:01:45 <reagle> IDsymbol =

22:01:45 <reagle> xsd:NMTOKEN

22:01:50 <dajobe> non-normative

22:01:54 <dajobe> anyway, so it makes a uri

22:02:06 <reagle> the rng says its IDsymbole which says (see above)

22:02:07 <dajobe> and the rules for building a URI tell you that you resolve against the in-scope base uri

22:02:17 <dajobe> the in-scope base uri is defined for each element

22:02:46 <dajobe> now, there *are* syntax-restrictions on the legal attribute values

22:02:49 <dajobe> and rdf:ID has some

22:03:00 <dajobe> they match that of XML IDs (in xml namespaces)

22:03:09 <dajobe> which co-incidently is the same as xsd:NMTOKEN, in relaxng

22:03:16 <dajobe> or not co-incidently really, we made it so

22:30:49 * DanCon browses the source of the tz database... amazingly rich in history...

22:31:01 <DanCon> # Part of Kentucky left its clocks alone in 1974.

22:31:01 <DanCon> # This also includes a part of Indiana immediately adjacent to Louisville.

22:31:34 <DanCon> ...

22:31:35 <DanCon> # After prolonged debate, and despite continuing deep differences of opinion,

22:31:35 <DanCon> # Wayne County (central Kentucky) is switching from Central (-0600) to Eastern

22:31:35 <DanCon> # (-0500) time. They won't "fall back" this year. See Sara Shipley,

22:31:35 <DanCon> # The difference an hour makes, Nando Times (2000-08-29 15:33 -0400).

22:32:17 * sbp` read an article about that not so long ago--it caused quite a bit of confusion

22:32:27 * sbp` waves; is only in fleetingly

22:42:02 <DanCon> wow... this is a wonderful study of naming...

22:42:25 <DanCon> I wonder if this stuff is already unpacked in the web, i.e. googleable...

22:43:02 <DanCon> ah.. yes...

22:43:04 <DanCon>http://www.ludd.luth.se/~ams/djgpp/cvs/djgpp/zoneinfo/src/theory

22:43:04 <dc_rdfig> R: http://www.ludd.luth.se/~ams/djgpp/cvs/djgpp/zoneinfo/src/theory from DanCon

22:44:20 <DanCon> R:|Theory, from the Olsen tz/zoneinfo db

22:44:20 <dc_rdfig> Titled item R.

22:44:40 <zool>http://search.cpan.org/author/BLUEFEET/Geo-Distance-0.04/Distance.pm

22:44:41 <dc_rdfig> S: http://search.cpan.org/author/BLUEFEET/Geo-Distance-0.04/Distance.pm from zool

22:44:56 <DanCon> R:a wonderful study in naming, persistence, consensus, etc.

22:44:56 <dc_rdfig> Added comment R1.

22:45:11 <zool> S:*This perl library aims to provide as many tools to make it as simple as possible to calculate distances between geographic points*

22:45:11 <dc_rdfig> Added comment S1.

22:53:22 * mortenf wonders why a US postal code is called a "zip" code?

22:55:23 <zool> [[ ZIP is an acronym for Zone Improvement Plan. ]]

22:55:38 <mortenf> aha (?)

22:55:57 <zool> cf http://www.usps.com/zip4/zipfaq.htm

22:56:09 <zool> i asked jeeves *cough*

22:56:14 <mortenf> thanks. :-)

22:56:39 * zool pleasantly surprised at the actually-not-bad-ness of askjeeves now

22:59:09 <mortenf> dang, one has to pay for a zip file (pun intended).

23:00:39 <zool> the end of monopolistic proprietary geodata is nigh, though perhaps i should feel more ambivalent about that re personal trackability/spammability

23:01:02 <mortenf> I don't know what to think of that either...

23:01:35 <zool> semantic self-defense classes...

23:01:44 <mortenf> heh.

23:03:32 <zool> i suppose it exposes hardware implementation flaws which have to be reimplemented or legislated out of existence... fax spam and autodialers vs. aggressive car alarms and mobile phones with no background noise...

23:03:39 <mortenf> weird, I just got spammed from timezonesforpc.com...

23:04:30 <zool> what an odd service

23:13:01 <DanCon> hmm... I think I need to publish more stuff in order to make these .rdf timezone files really useful... e.g. a list of top-level "areas", and a list of cities within those areas; i.e. the info implicit in the directory structure.

23:14:10 <zool> that sounds lush, DanConn

23:15:50 <DanCon> I'd really like to have lat/long for each "location" in the tzdaya

23:15:54 <DanCon> tz data

23:16:43 <zool> straightforward to extrapolate from the nima.mil files if you have pleeenty of diskspace

23:17:36 <DanCon> hmm... can I substitue bandwidth for disk space? i.e. to they publish their files via http?

23:18:00 <DanCon> wanna show me how to get lat/long for, e.g. America/Yellowknife ?

23:18:08 <zool> only in .zipped tsv format :/

23:18:25 <zool> ah, and they don't publish data for america, just the rest of the world :}

23:20:03 <DanCon> I have name->lat/long rules for u.s. places and for world-wide with pop>100,000

23:20:31 <DanCon> i.e. http://www.w3.org/2000/10/swap/pim/usPlace2LatLong.n3 and

23:21:30 <DanCon> ... http://www.w3.org/2000/04/mem-news/teamToGlobe.n3

23:22:55 <zool_> i've only poked around the UK and spain files at http://www.nima.mil/gns/html/cntry_files.html in any detail but they purport fairly successfully to hold all named settlements

23:26:14 <zool_> i'd happily volunteer to help put something together from the tz files, have a new disk in the post

23:26:23 <danbri> I had a look around this morning, got as far as realising I didn't know how to do the char encoding stuff needed to get the Iraqi data I was looking at into XML

23:26:46 * mortenf is working on the Danish data.

23:26:54 <zool_> hm, i've never looked at a non-roman one :/

23:27:12 <mortenf> danbri: isn't it just utf-8?

23:27:31 <danbri> I don't know about about char stuff to know!

23:28:27 <danbri>http://www.nima.mil/gns/html/cntyfile/iz.zip

23:28:27 <dc_rdfig> T: http://www.nima.mil/gns/html/cntyfile/iz.zip from danbri

23:28:54 <danbri> T:|mimal.mil geo data file (zip of tab sep'd records) for Iraq

23:28:55 <dc_rdfig> Titled item T.

23:29:09 <danbri> T:Settlments, etc.

23:29:09 <dc_rdfig> Added comment T1.

23:29:48 <danbri> T:I have no idea about accuracy or completeness or rules for use, but theres a fair chunk of data there.

23:29:48 <dc_rdfig> Added comment T2.

23:30:18 <mortenf> ""There is no problem in making the geographic names data freely available. A suitable citation note is...""

23:30:30 <danbri> cool

23:30:31 <mortenf> from the gns faq

23:31:58 <DanCon>http://www.w3.org/2003/01/geo/

23:31:59 <dc_rdfig> U: http://www.w3.org/2003/01/geo/ from DanCon

23:32:08 <DanCon> U:|RDFIG Geo vocab workspace

23:32:08 <dc_rdfig> Titled item U.

23:32:35 <danbri> danc, you had some Qs about that which I didn't answer yet

23:32:42 <DanCon> U:some [23Jan comments|http://lists.w3.org/Archives/Public/www-archive/2003Jan/0065.html]

23:32:43 <dc_rdfig> Added comment U1.

23:32:43 <danbri> including 'who is this we'

23:32:52 * danbri nods

23:33:25 <danbri> part answer is that the page was originally hastily ripped off from the rdf cal page, which is more solidly grounded in collaboration

23:33:53 <danbri> at the time the we included i think jimley and zool, who had started to incorp the vocab into their work in place of their own similar namespaces

23:34:10 <danbri> another q: why use a class and properties instead of just properties folk can apply to their own classes?

23:35:14 <danbri> I originally preferred the minimalist view, just a trio of properties... but it seemed that the truth conditions for knowing whether the use of the properties were correct... was easier to specify if they are properties of a known class

23:35:32 <zool_> SpatialThings look good for that certain conceptual level-of-detail; but it's valuable to have the stable reference point for lat/long we were then lacking...

23:35:34 <danbri> i agree that attaching an rdfs:label to a g:Point is a bit of a hack

23:36:11 <zool_> i've never used the geo:Point class

23:36:39 <DanCon> if i can't use these lat/long props directly on people, airports, and meeting rooms, I'm not likely to use them at all.

23:37:02 <danbri> I think I know what <g:Point g:lat="..." g:long="..." g:alt="..." .../> says about the world, and it leaves a gap, need for glue connecting it to the social, yeah airports, people etc.

23:37:16 <danbri> zool, did you use the geo:lat etc properties?

23:37:35 <zool_> don't you have a stable scape - the airport - and a variable scape - the person in the airport context - do you need the same mapping system to describe both?

23:37:36 <danbri> what is it for an airport to have a lat/long/alt, where those specify a precise point?

23:37:42 <mortenf> but shouldn't a location be tied to a point in time (for people)?

23:37:52 <DanCon> in my experience, the glue is tedious and unnecessary.

23:38:25 <danbri> I know what it is for a point to have a lat/long/alt (i think), and (more or less) for an airport to be withinFiveMilesOf some such point

23:38:39 <DanCon> point in time... please, leave those orthogonal. (and please read the cyc stuff on times, places, and contexts.)

23:38:39 <danbri> +1 re tedious, not so sure re necessary

23:38:50 <zool_> but if you opt for the latter option - reference to a chosen centre - dont you have to perpetually readjust / recalculate whenever two systems meet - and also run into curvature problems?

23:39:32 <danbri> I was also concerned about lat/long/alt giving false sense of precision.

23:39:39 <zool_> we could describe a cubic grid and pop the planet into it and that would be wonderfully conceptually clear, but perhaps not so useful...

23:39:45 <danbri> I want to catalogue where some shops are in my street, that I have photos of.

23:40:24 <DanCon> machines don't get "false sense"s, danbri. I get the impression this schema isn't sufficiently tied to any real code/data/experience.

23:41:02 <danbri> I know the geo markup for the street. But I can't be bothered to go gps measure each point in the street, so I want a way of exposing the data I have collected on my Visor/GPS that doesn't represent the shops as all being at the same point.

23:41:07 <zool_> if i have a way of easily publishing and describing my conceptual centre of reference and the scale i'm using, i would happily offer lat/long and local grid, but the appeal of lat/long is as a recognisable standard- a transition thing, rather than a conclusion thing

23:41:21 <DanCon> yes, play with photo data. then tell me whether the indirection still looks necessary.

23:41:47 <danbri> I've tried saying

23:42:06 <DanCon> in what case was it necessary?

23:42:15 <danbri> <Shop><depiction><Photo ...></depiction><lat>...</lat>etc...

23:42:48 <zool_> what's 'indirection' in this context, sorry?

23:42:50 <DanCon> er.. that applies the lat/long properties directly to the shop, no, danbri?

23:42:53 <danbri> ...but the stalling point was that I know this to be a misrepresentation. My GPS readings are of a point within approx 50 M of the shop, not of the centre of legal of architectural gravity of the shop.

23:43:09 <danbri> Yes, it did, hence my inclination towards the 'tedious' indirection

23:43:22 <danbri> s/of/or/

23:44:13 <DanCon> ah; the GPS reading wasn't in the shop at all, but rather from the vantage point of the photograph

23:45:05 <danbri> it was from my previous trip up the street with a Visor+GPS. I didn't take a reading outside each shop, either; and I'm also aware that the accuracy of the GPS gadget is ~10 M.

23:46:00 <DanCon> re legal/arch center, I recommend reading cyc on measurement too. I say I'm 6'2" ; folks understand that to be an imprecise measurement.

23:46:03 <danbri> So I can't bring myself to claim that the Shop (or any other geographic feature I'm describing) has the lat/long/etc, since a lat/long/alt trio indicates a point in space.

23:47:14 <DanCon> ok, so your .rdf file should say the shop is within, say, 1km of the lat/long point you took.

23:47:21 <danbri> Another angle: what do two things that share common values for geo:lat/lang/alt have in common?

23:47:34 <danbri> Yep, exactly... thats the markup style i'm drawn to.

23:47:49 <DanCon> they're... colocated or somesuch, in the cyc vocab... hang on...

23:48:02 <danbri> re the 'have in common' question, the geo vocab i proposed, answer is that they are just the selfsame thing.

23:48:18 <danbri> ...giving solid basis for data merging.

23:48:27 <DanCon> but if you *had* taken the GPS reading inside the shop, surely then you could apply that lat/long right to the shop, no?

23:48:41 <DanCon> selfsame: no, pls don't do that.

23:49:02 <danbri> If we take the scruffier route, the rules for data merging get harder

23:49:39 <danbri> well, it follows from basing descriptions on the class geo:Point. Two things are the same geo:Point if they have a common lat, long and alt.

23:50:02 * DanCon reviews http://www.cyc.com/cycdoc/vocab/spatial-vocab.html

23:50:14 <DanCon> harder: are you sure?

23:50:17 <DanCon> which rules?

23:50:30 <danbri> If you have in a.rdf

23:50:45 <DanCon> yes...

23:50:53 <DanCon> @prefix k: <http://opencyc.sourceforge.net/daml/cyc.daml#>.

23:50:57 <danbri> <g:Point g:lat="1" g:long="1" g:alt="1" s:foo="bar"/>

23:50:59 <danbri> and

23:51:25 <danbri> in b.rdf: <g:Point g:lat="1" g:long="1" g:alt="1" s:nearbyShop="Amazon Coffee"/>

23:52:25 <danbri> and you know what the g: vocab defintion (in prose; OWL doesn't support this) says, then you can conclude that there exists some thing that has an s:foo property that is 'bar', an s:nearbyShop property of 'Amazon Coffee', and l/l/a of 1, 1, 1...

23:52:31 * DanCon thought danbri was asking for cyc in RDF; oops...

23:52:49 <danbri> If we go the route you prefer, and have in

23:53:21 <danbri> a.rdf: '<s2:Something g:lat="1" g:long="1" g:alt="1" s:foo="bar" ...'

23:53:23 <danbri> and

23:53:30 <DanCon> I'm not convinced by argument with foos and bars. If you had something that you're really using, I might find it more persuasive.

23:53:43 <DanCon> did you consider my person/airport scenario?

23:53:52 <danbri> b.rdf: '<s2:SomethingMaybeElse g:lat="1" g:long="1" g:alt="1" nearestAirport="BRS" ...'

23:54:17 <danbri> This was in part motivated by wanting to be more precise about locations that nearestAirport was allowing us to be, yes.

23:54:24 <danbri> livesNear etc

23:54:41 <danbri> Also to give the location of the airport whose code is BRS

23:55:03 <Jibbler> Jibbler is now known as Paul|sleep

23:55:11 <Paul|sleep> Paul|sleep is now known as Jibbler|sleep

23:56:03 <DanCon> not sure i follow you

23:56:05 * danbri revisits http://www.jibbering.com/2002/8/foaf-people-map.svg

23:56:31 <danbri> some of us were (following your example for webont group, using nearestAirport in foaf files.

23:56:50 <danbri> wanted to have a more detailed way of associating people with places, including airports.

23:57:12 <danbri> and exchanging airport location info, eg. for consumption by http://www.jibbering.com/2002/8/foaf-people-map.svg

23:57:39 <zool_> DanCon: re tedious and unnecessary glue between lat/long and local co-ordinate systems, may i ask what application it was that caused you annoyances? would be good to knowwhat to avoid

23:57:51 <danbri>http://www.jibbering.com/2002/8/foaf.2.js

23:57:52 <dc_rdfig> V: http://www.jibbering.com/2002/8/foaf.2.js from danbri

23:58:14 <danbri> V:|Javascript source for foaf people map

23:58:14 <dc_rdfig> Titled item V.

23:58:53 <DanCon> zool_, people/airports app. see http://lists.w3.org/Archives/Public/www-archive/2003Jan/0065.html

23:59:06 <danbri> V:Uses [http://www.w3.org/2003/01/geo/wgs84_pos#|http://www.w3.org/2003/01/geo/wgs84_pos#] namespace, amongst others.

23:59:06 <dc_rdfig> Added comment V1.

23:59:17 <zool_> i'm only dealing with high-level approximations like http://space.frot.org/rdf/tubemap.svg (incomplete)

23:59:28 <danbri> V:See also [http://www.jibbering.com/2002/8/foaf-people-map.svg|svg src] (may not always work, depends on rdf harvester...).

23:59:29 <dc_rdfig> Added comment V2.