Semantic Web Interest Group IRC Chat Logs for 2003-02-19

00:00:56 <DanCon> er... I was following the people-map scenario, but I don't see a problem with using lat/long on people and airports yet, danbri.

00:01:17 <zool_> you can't get more sspecific than >10metres i can see is problematic, but i still dont see why conversion between grid systems is so problematic

00:01:33 <danbri> from http://www.cyc.com/cycdoc/vocab/spatial-vocab.html """In any case, the space region itself is not to be confused with a physical object or other spatially localized (non-space-region) thing that might happen to be #$cospatial with it"""

00:01:50 <danbri> I fear that just using flat properties encourages just that sort of mistake.

00:02:49 <DanCon> well, let me know if/when you change your mind, I guess. I guess I'll need to make my own namespace.

00:03:01 <danbri> With g:lat/long/alt attached to a class g:Point, it is easy to determine whether some rdf description of such a point is true. We can say 'all points exist' or something...

00:03:22 <DanCon> and how is that useful?

00:03:35 <danbri> with flat:lat/long/alt, I have no idea how I'd set about investigating the world to find out whether some claims written in that vocab were true or false.

00:03:43 <DanCon> so?

00:04:07 <danbri> So a vocab that doesn't give its users any criteria for knowing when they've screwed up... is of limited use.

00:04:17 <DanCon> er... hang on, it's clear how you'd investigate: you go to the described lat/long and look around.

00:04:27 <danbri> go on...

00:04:32 <danbri> and then what?

00:04:50 <DanCon> limited use: I think you're being closed-minded.

00:04:59 <DanCon> then you look and decide if it'a true

00:05:05 * DanCon family time

00:05:15 <danbri> I have modest eyesight, I can only see a few hundred yards. Is that enough?

00:05:25 <danbri> I started out w/ same design as you...

00:06:03 * danbri shrugs (btw feel free to add in 3 flat properties to the ns I made if that seems cheaper than a whole new one...)

00:06:09 <danbri> night

00:06:32 * zool_ feels faintly culpable

00:06:38 <zool_> zool_ is now known as zool

00:06:46 * danbri grins at zool

00:06:46 <mortenf> nn

00:07:17 * mortenf is reading http://www.cyc.com/cycdoc/vocab/time-vocab.html

00:08:02 <zool> that kind of ontological certainty unnerves me deeply

00:08:36 <mortenf> well, is was told to read it...

00:09:07 <danbri> Yeah, we'll be testing you on your reading tommorrow mortenf ;-)

00:09:13 <mortenf> :-)

00:09:15 <danbri> nice photos from a'dam btw

00:09:35 <mortenf> thanks - no lat/long though...

00:09:50 <zool> what does it have, if anything, to say about recurrence?

00:10:31 <mortenf> it has a section on "Repeated events", but i'm not a calendar expert (either).

00:11:24 <mortenf> heh, actually the #$IrregularlyRepeatedEvent has no definition...

00:11:26 <danbri>http://www.newsmonster.org

00:11:27 <dc_rdfig> A: http://www.newsmonster.org from danbri

00:11:34 <SethR> gig .. just installed NewsMonster in Mozilla for windows and now i can't find where the thing is ... anybody know?

00:11:35 <danbri> A:|NewsMonster!

00:11:36 <dc_rdfig> Titled item A.

00:11:48 <danbri> dunno for windows

00:12:21 <zool> #$WeeklyTemporalObjectType #$AnnualTemporalObjectType eeeuuuwww

00:12:27 <SethR> seem like it should be a side panel .. but dont know how to get it to display

00:12:34 <danbri> A:Includes [http://www.newsmonster.org/semantic.html|semwebbery]...

00:12:34 <dc_rdfig> Added comment A1.

00:12:49 <danbri> oh yeah seth, i thought you meant in the filesystem.

00:12:54 <danbri> yeah, its an extra side panel

00:13:17 <SethR> ok... im stupid .. how do i add the sidepanel in mozilla ?

00:13:45 <danbri> under 'view'?

00:14:02 <danbri> not sure

00:16:34 <SethR> kewl ... i found it ... never knew that was there ... its parallel to the sidebar :)

00:16:41 <mortenf> ""#$SpaceRegion-Empirical is in a way the spatial analogue of #$TimeInterval, whose own instances can be fully characterized by specifying their temporal properties; these two collections can be used, respectively, to talk about space and time as dimensions .""

00:17:29 * mortenf still doesn't get how a person can have a time-independent location...

00:19:43 <SethR> hmmm....now NewsMonster wants Java Web Start ... where's that?

00:21:13 <danbri> its Sun stuff

00:21:20 <danbri> .google Java Web Start

00:21:21 <datum> Java Web Start: http://java.sun.com/products/javawebstart/

00:21:28 <mortenf> nn. really.

00:21:30 <burtonator> yeah

00:21:32 <burtonator> just install a JDK

00:21:39 <danbri> hi burtonator, congrats on shipping!

00:21:44 <burtonator> thanks!

00:22:04 <SethR> but i got Java on my computer already

00:22:12 <burtonator> SethR: what version and on what OS?

00:22:39 <SethR> Windows ME

00:22:49 <burtonator> that should work if you have a modern VM

00:22:53 <burtonator> like 1.3 >

00:23:06 <burtonator> just go to Tools | NewsMonster aggregator and it should take off...

00:24:26 <SethR> thanks, its installing the web start :)

00:25:42 <SethR> its working .. its working ... :) :)

00:27:28 <burtonator> yay

00:27:29 <burtonator> you scared me

00:27:30 <burtonator> ;)

00:27:39 <burtonator> I am paranoid that people won't get it to work!

00:29:13 <SethR> how can i order the feeds myself .. i got this whole system of establishing prioity of the feeds i read ... ?

00:30:38 <SethR> ok i see ... i can rename them ... and you just alphabitize them for me

00:30:47 <SethR> woops .. rename dont seem to work

00:31:44 <burtonator> why won't it work

00:31:52 <burtonator> Beta2 will have categorizing and grouping

00:32:14 <burtonator> I didn't think it was necessary for Beta1...

00:32:24 <danbri> burtonator, have you changed 'newsmonster' script in linux version lately?

00:32:31 <SethR> i right click .. choose rename ... type a new name ... but nothing changes

00:32:33 <danbri> NEWSMONSTER_CLASSPATH=$NEWSMONSTER_CLASSPATH:lib/xerces.jar

00:32:33 <danbri> NEWSMONSTER_CLASSPATH=$NEWSMONSTER_CLASSPATH:lib/xml-apis.jar

00:32:33 <danbri> NEWSMONSTER_CLASSPATH=$CLASSPATH

00:32:37 <danbri> java -Xmx256m -classpath $NEWSMONSTER_CLASSPATH org.peerfear.newsmonster.swing.Main $@

00:32:44 <burtonator> oh...

00:32:49 <burtonator> I think I might have...

00:32:52 <burtonator> what happens?

00:33:00 <danbri> looks like it resets the NM classpath to be contents of my classpath

00:33:03 <burtonator> SethR: did you get the warning about aggregation?

00:33:04 <danbri> null in this case

00:33:25 <burtonator> subscription changes are reflected when you run the aggregator

00:33:26 <danbri> yeah! NEWSMONSTER_CLASSPATH=$CLASSPATH

00:33:29 <burtonator> I know it is confusing

00:33:31 <danbri> ...comment that out, it works ok.

00:33:36 <SethR> no warning after i click ok to the rename

00:33:37 <burtonator> danbri: hold on .. let me check

00:33:46 <danbri> its aggregating...

00:34:10 <burtonator> you must have already hit ok on that then...

00:34:15 <burtonator> hm... let me check

00:34:44 <burtonator> SethR: I have forgot to put the warning in that function

00:34:44 <burtonator> ug

00:35:21 <SethR> im also getting lots of "the file bla bla cannot be found

00:36:05 <burtonator> Can you believe it!

00:36:06 <burtonator> I did!

00:36:14 <burtonator> SethR: where?

00:37:03 <SethR> rename works in edit preferences, but not on right click in the f8 tab

00:37:56 <burtonator> SethR: right... that left pane is static output from an XSLT

00:38:18 <burtonator> this might be confusing for some people I think

00:38:25 <burtonator> I wish I had a better way of handling this...

00:39:13 <SethR> so if i reload mozilla will my rename that did take change the sequence of the feeds in the left pane ?

00:39:42 <burtonator> if you run the aggregator

00:39:45 <burtonator> it will changes

00:39:47 <burtonator> er change

00:40:10 <burtonator> I will spend some more time improving subscription management in Beta2

00:41:56 <SethR> just got a error in JVM.DILL while aggregator was running

00:42:27 <burtonator> WOW

00:42:31 <burtonator> that isn't my fault ;)

00:42:35 <burtonator> what JVM version ?

00:43:01 <burtonator> I had that problem on my XP box with JDK 1.3.1

00:43:03 <burtonator> I had to reinstall

00:44:29 <SethR> dont know .. ran it again and it went ok ... and the sequence changed :)

00:44:39 <burtonator> yeah

00:44:41 <burtonator> good

00:45:33 * burtonator uploads another standalone build with danbri's classpath suggestion

00:45:45 <SethR> so why am i getting file not found on like for example CNN - WORLD ?

00:46:15 <burtonator> probably because the aggregator had a problem building the file

00:46:25 <burtonator> you should at LEAST get an error report

00:46:35 <burtonator> Can you submit a problem report so that I can get your stack traces?

00:46:40 <burtonator> I can fix that ...

00:55:17 <SethR> ok i think i know what happened .. i had clicked in some kind of was such that the aggregator did not finish .. this time i just sat and watched it till it was done .. and now they are all there

00:59:12 <burtonator> yay

01:09:35 <SethR> well, burtonator ...it's kewl ... very kewl .... im gonna use it, thanks :)

01:10:39 <burtonator> SethR: sweet!

01:10:42 <burtonator> another convert ;)

01:10:58 <SethR> gotta go ... news all read .. bye

02:06:42 <burtonator> danbri: did RDFAuthor automatically layout the graph in your FOAF example?

02:49:28 <DanCon> "Since it is not feasible for the editors to check individual licenses, submitted software needs to come with a license listed on the "Approved Licenses List" from Opensource.org." -- http://www.semanticwebjournal.org/

02:49:31 <DanCon> nifty!

06:35:53 <minddog> minddog is now known as minddogzz

08:51:53 <zoyd> what's the license of newsmonster ?

08:52:02 <Jibbler|sleep> Jibbler|sleep is now known as Jibbler

09:02:06 Topic now RDF & Semantic Web hack'n'chat http://rdfig.xmlhack.com/

09:02:06 Users on #rdfig: logger_1 zoyd enta_habibi larsbot dmiles jllykifsh burtonator outs pixel Wack grove DanCon eikeon datum soccos|away minddogzz jang kham Mutiny golbeck SeeTemp jordan Jibbler tav dc_rdfig deltab danbri xover esigler arnarl em xower _joshua sbp` sandro greenman MarkB

09:02:06 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

09:48:12 <golbeck> golbeck is now known as golbeck_zzz

10:38:00 <em> em is now known as em-log

10:55:49 * dajobe wonders if em-log is chopping logs or something

10:56:20 <dajobe> are there forests in central OH?

11:05:54 <dajobe> all I can remember was it was very flat

11:11:22 <jang> you sound like obelix describing switzerland

11:13:25 <dmiles> jang: oh prolog on, do you know of the best "planner" system in prolog?

11:13:32 <dmiles> err prolog one

11:13:54 <Wack> Wack is now known as _Wack

11:14:25 <dmiles> i coded an inference engine data into a planning domain so that "planning" emulates infernce

11:14:53 <dmiles> example: [[it is true that, airplane(airplane1)], [it is true that, human(passengers)], [The , passengers, boarded the , airplane1, .], [it is true that, human(pilot)], [The , pilot, boarded the , airplane1, .], [it is true that, flight_path(chicago, austin)], [it is true that, flight_path(seattle, chicago)], [The plane taxiied to the , RunwayFn(seattle), .], [airplane1, took off from, seattle, .], [The plane cruised towards , chicago, .], [Th

11:15:12 <dmiles> but this was a small WARPALN prover

11:15:32 <dmiles> and seems to get loop and slow

11:15:45 <jang> it depends on the problem as to which planners are worthy

11:15:57 <dmiles> trying to decide to between pathological code correction.. or scrapping WARPLANNER

11:16:18 <dmiles> and starting out with something smarter

11:16:36 <dmiles> so really then none is general?

11:16:45 <jang> I'm not aware of any really successful generic planner

11:17:07 <jang> for the most part, they tend to be geared towards the kinds of problems they're trying to solve

11:17:39 <jang> the kind of optimisation that's beyond a partial specialiser;

11:18:07 <jang> (ignore ";", it's a C / java reflex, alas)

11:18:11 <dmiles> :)

11:18:54 <dmiles> i was thinking of trying to access my planning background throu predicates that fit the data into its speciazation paradym

11:19:32 <dmiles> like for example i have more states then actions

11:19:46 <dmiles> i wouldnt want to make 60k states part of the intial world

11:20:18 <dmiles> so i made an action called realized_fact(Fact)

11:20:34 <_Wack> _Wack is now known as Wack

11:20:38 <dmiles> that goes to the KB and sees if it's true

11:21:00 <dmiles> that way it only makes it part of the state if it's needed

11:21:15 <dmiles> (the curent state of the plan model)

11:21:36 <dmiles> it was such a cool mind blower

11:22:00 <dmiles> that is why above it "[it is true that, flight_path(seattle, chicago)]"

11:22:38 <dmiles> but the hard thing is .. WARPLAN would treat log:implies smartly

11:23:03 <dmiles> you'd need to abstract that into say prove_consequant()

11:23:21 <dmiles> and errm anteceedant or something

11:23:37 <dmiles> so right into that context its the wrong planner

11:24:19 <dmiles> or i dunno.. i am thinking of just making it better and better.. but it would be sad if i was the only developer

11:24:32 <dmiles> quite a burden

11:25:16 <dmiles> so must find a way of recruiting people who know prolog into some project

11:25:24 <dmiles> but also what planner ot really start with

11:25:57 <dmiles> i am trying to understand the "types" of planners that are out there

11:26:20 <dmiles> Partial Order planners, SHOP, etc

11:26:40 <dmiles> to know which to begin with

11:28:54 <dmiles> you with prolog experiance .. where would you start?

11:30:11 <jang> hm

11:30:14 * jang digests...

11:31:02 <dmiles> so much lisp tools out there.. but i am still kinda convinced that a prolog implimnentation could only be topped by someone who took a prolog interptor in C and specialized their unifiers

11:31:21 <dmiles> (so would be content in normal prolog)

11:32:09 <jang> I've just had a look at my bookmarks (a bit out of date)

11:32:17 <jang> what you say is true...

11:32:37 <jang> ... I don't have a reference to a single prolog tool.

11:32:54 <jang> they're all lisp or scheme or some variant.

11:33:03 * jang goes looking.

11:33:17 <dmiles> thank you (btw)

11:40:23 <dmiles> the problems i had when looking people love to say "prolog" in their webpages about planning.. yet not implented.. then when i think i am finding a planner they are using 5 lines of code version of solve/5 :) and didnt invest into making it smart

11:45:26 <sandro> IMIHO (I=ignorant) Prolog is good only when you don't want smarts -- it's smarts are good enough. If you want your own smarts, use a programming language.

11:46:21 <dajobe> hint: expect some interesting OWL & SWI-Prolog stuff to be announced soon...

11:46:37 <dmiles> dajobe: part of my renewed interest

11:47:20 * dajobe hunts for #ifdef to exclude various bits of OSX brokenness

11:47:58 <dmiles> sandro: i have used prolog as a programming language (its VERY difficult to not try to just use it as a database) very easy to get it confused with an infernce engine

11:48:37 <dmiles> alot of sentiment towards prolog comes that there are actually very few prolog applications out here

11:48:56 <dmiles> becasue programming in it is very wierd

11:49:14 <sandro> Yeah - I was being a little catty, I suppose. I think prolog is great as a rules engine and mostly-lousy as a programming language.

11:49:29 <sandro> that is, I love Prolog in its place.

11:49:42 <dmiles> yeah definetly true.. into the ammount of time you have to spend to make it do programming tasks

11:49:44 <dajobe> ouch

11:50:15 * jang sighs

11:50:17 <dmiles> well.. i had a job once with another prolog programmer it was very wierd

11:50:25 <dmiles> he say it as a rule engine

11:50:29 <dmiles> err saw it

11:50:33 <jang> well, despite the inherent nondeterminism in just about every planning algorithm out there...

11:50:43 <jang> ... I can't find a single system that's not based on lisp (!)

11:50:55 <dmiles> i never had thought that way :P.. i always thought it was a programming language

11:51:19 <dmiles> so our code syles were 180.. but it was amazing

11:51:44 <dmiles> he actually tried to process the problemn in prolog.. i tried to write a program in prolog to process the prblem

11:52:05 <sandro> And looking back, do you now think he was right...?

11:52:34 <jang> prolog's not smart enough. You need to know too much about how it works under the hood to get it to go properly

11:52:35 <dmiles> hrrm well he was laid off 4 months before i

11:53:11 <dmiles> but.. hrrm i dont think eighther was more right

11:53:22 <dmiles> well.. i do

11:53:37 <dmiles> i could go back into prolog and leverage its rule engine parts

11:53:49 <sandro> It seems to me tabling in prolog (eg XSB auto-table) changes it from being smart enough 20% of the time to smart enough 80% of the time.

11:53:59 <sandro> ... but I'm not thinking about planning there.

11:54:17 <dmiles> yes god tabling is important

11:54:23 <dmiles> planning has a built in tabling

11:54:40 <dmiles> becasue it has a "state" so transitive closure is not an issue

11:55:06 <dmiles> planners dont loop on logical problems

11:55:28 <dmiles> becasue they are goal orientaed .. not backchain greedy

11:55:46 * sandro really know nothing of planning, except in the total abstract.

11:55:53 <dajobe> #ifdef _APPLE_CC__ /* OSX gcc is broken */ #else /* rest of universe */ #endif

11:55:55 <dmiles> heh i am just learning this week

11:56:18 <dmiles> i feel like i wasted too much time on propositional resultion now

11:56:56 <dmiles> after seeing a planner do logical implication with all those neat infernce rules

11:57:38 <dmiles> mainly the heuristic of an infernce engine becomes your "planning domain"

11:58:02 <dmiles> so you deal first with the shell of your search .. before you do it

11:58:47 <dmiles> but its very hard to add planner to an existing infernce engine

11:58:55 <dmiles> from what i have been trying to do

11:59:08 <dmiles> but a bit easier to impliment proof theory in it

11:59:14 <dmiles> (in a planner)

12:00:07 <dmiles> also you kinda can represent time relations

12:00:17 <dmiles> (and have them mean something)

12:01:06 <dmiles> jang: well i found somthing written by MURRAY SHANAHAN

12:01:16 <dmiles> (July 1998)

12:02:09 <dmiles> i am messing with that.. but maybe later you'll have thoughts or suggestings

12:02:20 <sandro> Logic terminology question: I'm trying to disambiguate two meanings of "constant": (1) a symbol which denotes one object in the domain of discourse, or (2) as in (1) PLUS function symbols, predicate symbols, proposition symbols..... I guess proposition symbols are just nullary predicates, and 1-constants are just nullary function symbols; hrm.....

12:03:27 * dmiles reads that 5+ times

12:03:34 <sandro> heh, sorry

12:03:39 <dmiles> nbo its cool

12:03:44 <sandro> I can try to rephrase....

12:03:52 <dmiles> i am trying to do something usefull right now :) that looks the most usefull

12:04:48 <dmiles> defineing the "constant" as in how we could use it in a setence?

12:05:17 <dmiles> Constant is supposed to be anything that always us to make a unique name assumption

12:05:33 <dmiles> wee that we have all agreed deserves the unique-name rules

12:06:31 <sandro> !! In prolog perhaps, but not in FOL, or in algebra, or programming languages. const height=100; const width=100; No UNA there.

12:06:33 <dmiles> MarkTwain and SamuelClemens are both unique constants

12:07:09 <dmiles> but yet are the same logical object

12:07:25 <sandro> Ah, sure. Okay.

12:07:43 <dmiles> PLUS i would think also is a constant

12:08:46 <dmiles> so constant would really be a scoping operator in logic

12:08:56 <dmiles> maybe?

12:09:33 <sandro> Let me phrase it this way: If you were looking at a parse tree for the logic sentence: forall x (p or q or f(x,a) = g(x,b)

12:10:05 <sandro> .... how would you talk about "p"? You'd say it was a proposition, right?

12:10:17 <sandro> ..... and "f"?

12:10:23 <sandro> ..... and "a" ?

12:10:24 <dmiles> p is a literal

12:11:13 <dmiles> f is a predicate constant

12:11:24 <dmiles> p is also a constant

12:11:42 <dmiles> x is the only thing that is not a constant

12:11:59 <sandro> what terminology would you use to describe the kind of constant "a" is, as distinct from the kind "f" is?

12:12:03 <dmiles> (i am guessing (also not expertt at this but enjoy it))

12:12:22 <sandro> [alas, me too. :-)]

12:12:29 <dmiles> a is an object constant

12:13:03 <sandro> Hrm. How about Individual Constant?

12:13:10 <dmiles> i think becasue it's quanified

12:13:31 <dmiles> i dont know how to decide what is an indiviual

12:13:56 <sandro> hm.

12:13:58 <dmiles> oh.. maybe Object==Indiviual

12:14:24 <sandro> right. I think that's the description logics view.

12:15:00 <dmiles> its scarey that a logic system might be ambiguous in notion

12:15:15 <sandro> notion or notation?

12:15:24 <dmiles> notation

12:15:42 <sandro> like English? :-)

12:15:56 <dmiles> well that f(..)=f(..) is the = the predicate?

12:16:08 <dmiles> (not that that was the ambiguity)

12:16:26 <sandro> Yeah, = is a n infix binary predicate.

12:16:32 <dmiles> yeah logic is no less obscure then natural language ;P

12:17:05 <dmiles> so f(..,..) are indivouls

12:17:12 <dmiles> individuals

12:17:33 <dmiles> a and b are as well

12:17:37 <sandro> I'd say the value of f(...) ranges over individuals

12:17:40 <dmiles> but x? i cant guess

12:18:08 <sandro> individuals == objects in the domain of discourse

12:18:41 <dmiles> what would x range over?

12:18:51 <sandro> same

12:19:31 <dmiles> could it be numerical constants?

12:19:48 <dmiles> f(1,a)=f(2,b)

12:19:48 <sandro> whatever.

12:19:59 <dmiles> erm

12:20:13 <dmiles> oh .. not two numbers

12:20:34 <dmiles> (or two things) that just was a typo

12:21:05 <sandro> Ah, no, "x" is a variable term; you can think about it being bound to integers in some solution/interpretation.

12:21:24 <sandro> and yes, of course x is bound to the same thing everywhere.

12:22:09 <dmiles> about forall in general... the x it denotes should mean all things that are objects in the entire theory?

12:22:45 <dmiles> or is it subserviant to the domain of f 1?

12:23:16 <dmiles> so like domain(f,numbers)

12:23:29 <dmiles> domain(f,predicates)

12:23:47 <dmiles> not(domain(f,logicalobjects))

12:24:15 <dmiles> whould make it weird

12:24:50 <dmiles> domain = arg1? range=arg2? (maybe i have it reversed)

12:25:42 <dmiles> i think i am adding something not part of your intial thing.. but i think its impossible to do context free parsing on logic

12:26:06 <dmiles> (which is plain terrible!)

12:26:24 <dmiles> why cant we have representations that are explicit?

12:28:53 * sandro laughs

12:32:55 <dmiles> one thing i s am doing in the planner.. with ACTION: consder("forall x (p or q or f(x,a) = g(x,b)") is that it has a precondition that it domain(f,_) V range(f,_)

12:33:30 <dmiles> erm meant ^ for V

12:33:38 <sandro> yeah....

12:34:04 <sandro> I can parse that statement, at least. :)

12:34:22 <dmiles> this is why i think we can really do better infernce with a planner

12:34:35 <dmiles> that each logic context has bussiness rules

13:46:27 * danbri renounces iPAQ and PocketPC, decides to go back to his trusty old black and white PalmOS Visor

13:47:17 <danbri> I'll keep the ipaq for showing off, roaming net access, mp3 playing, audio recording, occasional telephony...

13:47:34 <danbri> ...I just won't try using it as an organiser :(

13:48:05 <danbri> Or I could put Linux on it, but I'm not feeling that hardcore

13:53:21 * dajobe moves on to merging win32 patches for raptor

13:53:49 <dajobe> turns out I nearly got all the win32 filename code right, without ever running it on such a beast

14:08:15 <sandro> danbri, what was the fatal flaw with the iPAQ?

14:08:41 <Jibbler> my ipaq's sat on my desk for about a year

14:08:54 <Jibbler> i wrote a graph drawing program for it once, though

14:11:54 <danbri> ipaq only talks to MS Outlook

15:36:12 <minddogzz> minddogzz is now known as minddog

16:56:10 <eikeon> Does anyone know if there is a schema for describing a schema as a whole?

16:57:43 <eikeon> Say one that defines a schema class and maybe properties for name, description, location where schema can be found... type of stuff.

16:57:46 <danbri> people use dublin core...

16:58:47 <eikeon> Can you point me at an example?

16:58:52 <danbri> no :)

16:58:58 <danbri> sorry, nothing comes to mind...

16:59:59 <danbri> woah -- 4 rdf packages in CPAN (RDFStore, RDF::Core, RDF::Notation3, RDF::Service) plus a bunch of RSS stuff, and an alpha FOAF one; outside of CPAN there is EricP's code, Dave Beckett's Redland w/ Perl bindings, and CARMEN/CARA.

17:00:07 <eikeon> Wouldn't mind adding another section to rdfschema.info (after clarifying wording of what the store contains ;)

17:00:31 <mortenf> The schema at http://www.w3.org/2003/01/geo/wgs84_pos has some schematametadata (!)

17:01:43 <danbri> I was thinking of doing an Annotated Ontology Language ns with a few leftover things that aren't in OWL or RDFS, like StaticInverseFunctionalProperty and Debian-esque maturity levels for properties/classes

17:02:49 <mortenf> eikeon: it would be nice if schema.info evolved into a schema directory of sorts...

17:03:18 <mortenf> so one could say in response to "is there a schema for X": look at schema.info!

17:03:38 <eikeon> mortenf: That is the sort of thing I am looking for... am hoping to find something people are already using as a type for schemata too.

17:03:59 <danbri> i had vageuly similar aspirations for xmlns.com, but never shipped much...

17:04:13 <danbri> I like idea of linking schemas more explicitly to the people, organisations, projects that use them

17:04:17 <danbri> sample data, services etc...

17:04:21 <mortenf> +1

17:04:23 <danbri> rather than a dry database of DTDs...

17:04:39 <danbri> also provenenence of who says what...

17:04:39 <eikeon> mortenf: That is my main motivation for doing it... I kept finding myself trying to answer my own questions along those lines... and never found a good source of answers.

17:04:46 <danbri> eg. dan says that rss:title subproperty of dc:title

17:04:47 <danbri> versus

17:04:52 <danbri> RSS-DEV WG says that ...

17:05:08 <danbri> I think schemas should be digitally signed, as should mapping assertions.

17:05:27 <danbri> scuttered data seems a good place to start, re what's actually in use in the world

17:05:38 <mortenf> sounds like a major task, danbri...

17:05:45 <danbri> yes, it is

17:05:50 <danbri> but we can get there gradually

17:05:53 <mortenf> yeah.

17:06:12 <danbri> a lot of the things we build for foaf aggregators should work for RDFS and OWL schema repositories too...

17:06:27 <danbri> oh, wonder if MattB has seen NewsMonster

17:06:40 <mortenf> does it grok his photo-rss?

17:06:43 <danbri> ...could be a good thing to hook his Jena-based rdf harvester thing too...

17:06:49 <danbri> don't know, good question

17:06:59 <mortenf> or mine, for that matter...

17:07:24 <eikeon> Is a major task... I am just trying to see how much of a start I can make on the easy bits and go from there.

17:08:25 <eikeon> Any interest in sharing harvester info?

17:08:45 <eikeon> We could create a small schema for that too.

17:08:45 <mortenf> I'd say "much", but I don't have one yet.

17:08:51 <danbri> harvester info...

17:09:13 <danbri>http://rdfweb.org/2002/foaf/scutter/log/

17:09:14 <dc_rdfig> B: http://rdfweb.org/2002/foaf/scutter/log/ from danbri

17:09:23 <danbri> B:|Some foaf harvester logs

17:09:23 <dc_rdfig> Titled item B.

17:09:29 <danbri> B:I should crontab this...

17:09:29 <dc_rdfig> Added comment B1.

17:09:30 <eikeon> URLs discovered... response... errors... that sort of thing.

17:09:50 <danbri> B:Gives property and class occurance stats, not much else

17:09:51 <dc_rdfig> Added comment B2.

17:10:01 <eikeon> Cool... that are already in RDF :)

17:10:02 <danbri> B:Would like to do URIs too, and URI/property pairs

17:10:02 <dc_rdfig> Added comment B3.

17:10:24 <danbri> B:so I can use it for bulk service description metadata, query routing etc.

17:10:24 <dc_rdfig> Added comment B4.

17:10:54 <danbri>http://www.w3.org/TandS/QL/QL98/pp/distributed.html

17:10:54 <dc_rdfig> C: http://www.w3.org/TandS/QL/QL98/pp/distributed.html from danbri

17:11:01 <danbri> C:|Query Language Issues in a Distributed Indexing Environment

17:11:01 <dc_rdfig> Titled item C.

17:11:29 <danbri> C:by Peter Valkenburg, Dan Brickley (position paper, W3C query workshop, Boston 1998).

17:11:29 <dc_rdfig> Added comment C1.

17:11:32 <Jibbler> oh, i hate my irc client, going "bing bong" whenever someone sends a notice to a channel

17:12:05 <eikeon> Hum... first thing I should do is expose what I already have in RDF!

17:12:38 <danbri> C:see sectionon query routing... We should be able to do a nice job on that for RDF aggregators. List of all URIs, vs list of all hashed URIs, or list of URI/property or property/URI pairs...

17:12:38 <dc_rdfig> Added comment C2.

17:13:29 <danbri> C:"No single catalogue can offer complete coverage of the Web, so there is a need for a 'forward knowledge' mechanism by which search agents might discover services that offer 3rd-party descriptions and metadata annotations for some specified Web resource."

17:13:29 <dc_rdfig> Added comment C3.

17:13:59 * danbri totally stands by all that... just wish we'd finished building it after 5 years...

17:14:31 <danbri>http://rdf.burningbird.net

17:14:31 <dc_rdfig> D: http://rdf.burningbird.net from danbri

17:14:50 <danbri> D:|Practical RDF weblog - complete first draft released

17:14:50 <dc_rdfig> Titled item D.

17:17:15 <eikeon> danbri: Does the stat schema exist somewhere... I am assuming it is the schema I would use for the {Properties,Classes} by popularity sections I have?

17:17:47 <danbri> D:FOAF stuff in ch14 under non-commercial apps. Hmm its time to clarify some bits of the vocab (surname/familyname/first/given/blah) before its too late...

17:17:48 <dc_rdfig> Added comment D1.

17:18:20 <danbri> stat schema? xmlns:stat="http://xmlns.com/2002/07/stat#" is vapourware

17:18:48 <danbri> If there is interest in collaborating on something, we could try writing a schema

17:18:50 <mortenf> danbri: I'm still threatening to clean/fix/append the FOAF schema if you don't before I get to it! :-)

17:18:54 <danbri> hosted there on on w3.org whatever...

17:19:13 <danbri> morten, wanna try your hand at the firstname/surname/etc issue in particular?

17:19:25 <mortenf> yeah, that as well - any specific pointers?

17:19:26 <danbri> raise a bugzilla entry for it if there isn't one...

17:19:40 <danbri> -> http://rdfweb.org/issues/

17:20:01 <danbri> I'm not sure if the concern has been raised yet.

17:20:38 <eikeon> danbri: Yeah... I am interested in collaborating on a schema. Especially if we can put it to good use.

17:20:54 <danbri> DanC raised it very early on. Need it to reconcile existing tools (foaf-a-matic in particular), data (foaf-a-matic generated stuff esp), and i18n concerns (I think the wrong terms have snuk into widespread use), also the actual contents of the schema at the namespace.

17:21:10 <danbri> i'll create a bugzilla entry.

17:21:17 <danbri> mortenf, do you have a login on foaf bugzilla?

17:21:25 <mortenf> great, I'll look around (at use etc.)

17:21:33 <mortenf> ah, no?

17:21:45 <mortenf> creating...

17:21:50 <danbri> ok, i can't cc you in the inital bugzilla entry, feel free to add after

17:21:53 <danbri> or ok i'll wait

17:23:00 <danbri> logger_1, pointer?

17:23:00 <danbri> See http://ilrt.org/discovery/chatlogs/rdfig/2003-02-19#T17-23-00

17:23:35 <mortenf> created.

17:23:43 <danbri> whats your email addr?

17:23:59 <mortenf> mof-rdf at mfd-consult dk

17:26:43 <_joshua> whoop!

17:28:18 * DanCon tries making a how/why diagram for SWAD-EU...

17:28:55 <danbri>http://rdfweb.org/issues/show_bug.cgi?id=7

17:28:55 <dc_rdfig> E: http://rdfweb.org/issues/show_bug.cgi?id=7 from danbri

17:29:13 * DanCon swaps in http://www.w3.org/2001/sw/Europe/plan/workpackages/live/rdf/

17:29:32 <danbri> E:|FOAF bugzilla entry (firstname/surname/etc issue)

17:29:32 <dc_rdfig> Titled item E.

17:29:38 <danbri> E:aka finding out whether Bugzilla works for collaborative RDFS editing

17:29:38 <dc_rdfig> Added comment E1.

17:29:44 <danbri> mortenf, is that ok?

17:29:47 <mortenf> great.

17:29:50 <danbri> :)

17:30:03 * mortenf leaves for "The Ring"...

17:30:08 <DanCon> <pm:wpStartDate/>

17:30:10 <DanCon> hmm...

17:30:14 * danbri notes that DanC always starts on interesting projects as the uk day draws to a close

17:30:29 <danbri> hmm, looks missing data

17:31:05 <danbri> btw dan, since we talked yesterday i have some .doc and excel stuff from Kate Sharp to look through, re updating swad-e effort planning.

17:31:23 <danbri> I expect to have more to say on that before monday, but nothing right now... not looked at it yet.

17:46:45 <DanCon> yes, remote collaboration is kinda cool, but the timezone issues can suck

17:52:32 * DanCon reviews http://www.w3.org/2001/sw/Europe/plan/workpackages/live/Makefile

17:53:16 <mhgrove> mhgrove is now known as mike-lunch

17:53:33 <DanCon> ooh... http://www.w3.org/2001/sw/Europe/plan/workpackages/live/_gantt.html

17:53:51 <DanCon> bummer... no "how did he do that" link

17:55:37 <danbri> hm s/he/she/ i suspect

17:56:20 <danbri> it was bin/extract_projdata.pl -- libby edited my ugly script

17:56:37 <danbri> i think made it even uglier, i remember grouching at the time, but i might just have been in a bad mood!

17:56:54 <danbri> [[

17:56:54 <danbri> # extract_projdata.pl $Id: extract_projdata.pl,v 1.10 2002/07/25 15:08:54 lmill$# hackparse the xhtml SWAD-Europe Workpackage descriptions

17:56:54 <danbri> #

17:56:54 <danbri> # hackier version by libby 2002-04-26 with a few cross-checks and

17:56:54 <danbri> # writing individual html files for earch deliverable

17:56:55 <danbri> # Soon to be superceeded by xslt.

17:56:57 <danbri> ]]

17:57:20 * danbri thought I'd outgrown such optimism -- "Soon to be superceeded by xslt" yeah right

17:58:37 <danbri> hmm maybe not that script actually. there was aprevious worse attempt at a gant chart.

17:59:50 <danbri> ah http://www.w3.org/2001/sw/Europe/plan/workpackages/live/bin/deliv-start-end.rb

18:03:27 <golbeck_zzz> golbeck_zzz is now known as golbeck

18:07:05 * DanCon chucles at ... outgrown...

18:17:28 * DanCon goes back to timezone rules...

18:20:58 <DanCon> hmm... I wonder if tzid should be a URIref...

18:22:25 <DanCon> gee... we don't have any good test data that uses Vtimezone components, do we?

18:23:19 <DanCon> ah... cal01...

18:26:29 <DanCon> "This

18:26:29 <DanCon> property parameter specifies a text value which uniquely identifies

18:26:29 <DanCon> the "VTIMEZONE" calendar component to be used when evaluating the

18:26:29 <DanCon> time portion of the property."

18:35:14 <mike-lunch> mike-lunch is now known as mhgrove

19:02:18 <minddog> heya dajobe

19:02:54 <dajobe> hello

21:31:31 <DanCon> hmm... there's something closed-world about tzid..

21:32:14 <danbri> howso?

21:32:42 <DanCon> I think tzids are local to a .ics file

21:33:07 <DanCon> i.e. my current mapping to .rdf won't "work" if you merge to calendars.

21:33:17 <danbri> they're a name taken from a set of names local to a doc?

21:33:26 <DanCon> yes, I think so.

21:34:14 <DanCon> I think they should become fragids.

21:34:18 <danbri> eg http://www.w3.org/2002/12/cal/tzd/America/Chicago#tz

21:34:26 <danbri> oh, i thought you were using fragids already

21:35:01 <DanCon> this isn't a fragid: <tzid>/softwarestudio.org/Olson_20011030_5/America/Chicago</tzid>

21:35:12 * mortenf is away: I'm busy

21:35:35 <danbri> ah, i was confusing w/ this: <Vtimezone rdf:about="#tz">

21:35:56 <DanCon> yes... look at http://www.w3.org/2002/12/cal/test/cal01.rdf

21:36:08 <DanCon> there's a reference from ...

21:36:18 * danbri scans http://www.w3.org/2002/12/cal/test/cal01.rdf

21:36:20 <DanCon> <dtstart rdf:parseType="Resource">

21:36:20 <DanCon> <dateTime>2002-06-30T09:00:00</dateTime>

21:36:20 <DanCon> -

21:36:20 <DanCon> <tzid>

21:36:21 <DanCon> </tzid>

21:36:23 <DanCon> </dtstart>

21:36:31 * mortenf is back (gone 00:01:19)

21:36:38 <DanCon> to the Vtimezone component.

21:37:20 <DanCon> now evolution seems to use world-unique tzids. But I'm not sure the ical spec requires that.

21:37:32 <DanCon> it certainly doesn't specify how to avoid collisions.

21:37:34 <danbri> I see <tzid>

21:37:34 <danbri> </tzid> as a property of <Vtimezone>

21:37:50 <danbri> and later, of a <Vevent>

21:38:29 <danbri> i was wondering if tzid was inverse-functional, but I guess not.

21:39:08 <DanCon> I think I have to either (a) assume tzid is inverse-functional , or (b) make tzids into URI refs

21:39:44 <DanCon> and (a) leads us to conclude that a Vtimezone component is equal to, say, the dtstart value of some Vevent

21:40:19 <danbri> If it is inverse-functional, then the <Vtimezone> described near the start of cal01, and the Vevent, are one and the same resource.

21:40:22 <DanCon> hmm... I could assume tzid is inverse functional only over Vtimezones

21:40:34 <danbri> or sorry, yes as you say, the start of the event.

21:40:41 <DanCon> not the vevent, but -- yes.

21:40:43 <danbri> ooh, that'd be fancy.

21:41:09 <DanCon> anyway... see what I mean about closed-world-ish?

21:41:59 <danbri> I'm considering scoped inverse functional stuff for the foafcorp dataset. It's kinda hacky. In general, foaf:name isn't uniquely identifying, but there are some names, ie those we use in foafcorp/theyrule data, that uniquely pick out individiauls. So currently I use a subproperty of name, so that the subprop can carry the inversefunctional-ness.

21:42:05 <danbri> ... http://rdfweb.org/foaf/corp/cola-corps-sample.xml

21:42:30 <danbri> I'm starting to, though I wouldn't have initially thought of it in those terms.

21:42:41 <DanCon> daml (and owl) can express "inverse functional over C" using cardinality.

21:42:56 <danbri> it's sorta scoped. i have trouble thinking about it as I don't know the cal vocab well enough yet.

21:43:10 <danbri> Oh, really? That's pretty handy.

21:45:19 <DanCon> hmm... I can see how to do "functional over C" but I can't work out the inverse bit... I'm pretty sure Ian showed me how in joint-committee, though

21:45:32 <DanCon> maybe just...

21:47:05 <DanCon> { _:nameI owl:inverseOf foaf:name. s:Literal s:subClassOf [ owl:onProperty _:nameI ] .... phpht. nope, I'm losing.

21:47:41 <danbri> thx for trying! i really want to have a crack at using OWL in real life

21:48:02 <danbri> Does the namespace URI change with every Working Draft btw?

21:48:02 <DanCon> you might ask public-webont-comments how to do it

21:48:15 <danbri> good idea

21:48:21 <DanCon> no, we've been sticking with 2002/07/owl I think

21:48:31 <danbri> good. i'm lazy...

21:48:53 <danbri> . http://www.w3.org/2002/07/owl <!-- $Revision: 1.3 $ of $Date: 2002/11/15 00:12:24 $. -->

21:49:14 <danbri> ...I guess the WDs are better reference than the namespace doc?

21:50:24 <DanCon> not sure.

21:51:12 <DanCon> the namespace doc is a copy of somthing in the WD, I think

21:52:49 <danbri> hmm, imports is still in there. could have some fun with that...

21:54:54 <DanCon> ugh... this ical schema is a pain to work with... I'm trying to write a rule ala start(E, UT) :- start(E, LT), tz(LT, TZ), gmtime(LT, UT, TZ)

21:55:17 <DanCon> but I don't think a Vevent can have more than one dtstart property

21:56:19 <danbri> you want multiiple dtstarts, each using a different zone?

21:56:56 <DanCon> or one dtstart with its object related to multipe date-strings, on a per-timezone basis.

21:57:10 <DanCon> but I can't do that either with this schema

21:57:11 <danbri> sounds reasonable...

21:57:42 <danbri> any obvious fix come to mind?

21:57:57 <DanCon> I think I'll make conclusions using the cyc vocab

21:58:41 <danbri> the conclusion could be a fresh instance of Event, just temporally overlapping..., perhaps?

22:00:27 <DanCon> having to make a new event thingy is the sort of thing I hate.

22:00:55 <danbri> yup. conceptually, its the selfsame thing, we just know more about its properties...

22:01:07 <DanCon> on lat/long... how about saying that lat/long are uniquely-identifying over geo:Points without constraining their domains?

22:01:38 * danbri ponders

22:02:01 <danbri> how would I know if the following were true?

22:02:03 <DanCon> well, without constraining their domains beyond SpatialThing

22:02:34 * DanCon don't feel any obligation to answer "how would I know if the following were true?"

22:02:54 <danbri> (no obligation) <wordnet:City rdfs:label="Bristol, UK"><lat>40</lat><long>50</long><alt>24</alt></wordnet:City>

22:03:27 <danbri> ie. can you write definitions for rdfs:label/comment of lat, long, alt that would help people judge if they'd used it correctly?

22:03:30 <DanCon> that's true if bristol spatiallyIntersects the point at 40/50/24

22:03:45 <danbri> I couldn't, so I wrote the one I'm using now.

22:04:09 * danbri considers that

22:04:32 <DanCon> I don't put much stock in people reading rdfs:comments. I believe in deploying code that makes it easier to do it right than to do it wrong.

22:05:54 <DanCon> I happen to like the rdfs:comment-ish stuff in the cyc docs a whole lot, meanwhile. But I believe those docs came after lots of code and experience.

22:05:59 * danbri too, but also in deploying vocab; forcing spatiallyIntersects on people makes them less likely to expose imprecise data. And precise Geo data is rare, expensive.

22:06:26 <danbri> I guess one could fall back on the nearTo vocab I was anticipating

22:06:36 <DanCon> s/if bristol/iff bristol/, in case that wasn't obvious

22:07:11 <DanCon> there's a cyc:near. I think it's only useful in combination with some context machinery.

22:07:31 <DanCon> ala... "if doc1 says ?who near ?where, then ..."

22:07:43 * danbri nods, agrees

22:08:04 <danbri> without contextualising, i can only think of things like within100Metres

22:08:15 <DanCon> right....

22:08:48 <DanCon> ... of course, the cyc folks will say you've just fallen back on a context that, while very widely agreed, isn't any sort of logical necessity.

22:09:06 <danbri> quite likely

22:09:34 * danbri imports the Good Old Fashioned Textbook Physics context

22:09:38 <danbri> <DanCon> on lat/long... how about saying that lat/long are uniquely-identifying over geo:Points without constraining their domains?

22:09:53 <danbri> Sure, let's try that. Fancy making the changes to the RDFS?

22:10:06 <DanCon> ok...

22:10:35 <danbri> If I get squeamish and want to a have a point-based version, I reckon best progression path would be for me to add in a new set of lat/long/alt properties to the schema with the domain.

22:11:27 <DanCon> 2003/01/geo/wgs84_pos.rdf , right?

22:11:40 <danbri> . http://www.w3.org/2003/01/geo/wgs84_pos# yup, er yes, .rdf

22:11:57 <danbri> definitions are "The WGS84 longitude of a Point." etc at moment

22:12:24 <danbri> -> "A wgs84 longitude of a resource.", or similar?

22:12:27 <DanCon> shall I introduce a SpatialThing class?

22:12:51 <danbri> Do you have a need for it?

22:12:52 <DanCon> or just leave domain of lat unconstrained?

22:13:18 <danbri> Are there things that lat/long/alt are never truely descriptions of?

22:13:26 <DanCon> yes, ideas.

22:13:28 <danbri> (mangled sentence, hope intent clearish)

22:13:31 <DanCon> and numbers.

22:13:38 <DanCon> and colors.

22:13:56 <danbri> ideas: that's a philospohy of mind rathole. But I buy it for numbers and colors.

22:14:28 <danbri> Why not? this is hardly a massively bloated schema...

22:14:30 <DanCon> I don't need SpatialThing... but if you want lat to be related to Point, I'm willing to say lat's domain is SpatialThing, a superclass of Point.

22:14:57 <danbri> Ok, let's try that. If nothing else, it'll document our thinking...

22:17:11 <danbri> hmm you might wanna strikethru or edit last three parags from intro to http://www.w3.org/2003/01/geo/

22:17:34 <danbri> as might I, i guess ;)

22:21:21 <_joshua> use TGN names for cities

22:22:57 <_joshua> canonical and numbered to boot

22:23:20 <danbri> makes sense, unless TGN constraints make it tricky to exchange bulk data that uses their names.

22:23:45 <DanCon> anybody got rules for scraping TGN to RDF?

22:24:44 <DanCon> did you mean to have a # in the title statement?

22:24:44 <_joshua>http://lists.w3.org/Archives/Public/www-rdf-interest/2002Jun/0063.html

22:24:45 <dc_rdfig> F: http://lists.w3.org/Archives/Public/www-rdf-interest/2002Jun/0063.html from _joshua

22:25:27 <DanCon> pht. http://ietf.org/internet-drafts/draft-pstickler-voc-01.txt has gone 404

22:25:42 <danbri> danc,which title statement?

22:25:45 <danbri> <dc:title>WGS84 Geo Positioning: an RDF vocabulary</dc:title>

22:25:52 <DanCon> yes, that one

22:25:53 <danbri> is what i get ffrom mozilla view source

22:26:03 <DanCon> <rdf:Description rdf:about="http://www.w3.org/2003/01/geo/wgs84_pos#">

22:26:03 <DanCon> <dc:title>WGS84 Geo Positioning: an RDF vocabulary</dc:title>

22:26:11 <danbri> a '#' sign? or a number?

22:26:20 <DanCon> a '#' sign

22:27:09 <danbri> "http://www.w3.org/2003/01/geo/wgs84_pos#"<-- this one? its just the ns uri... as used in xmlns:g="...#"

22:27:16 <danbri> so yes, i guess i did

22:27:52 <danbri> we could s/rdf:Description/owl:Ontology/...

22:27:59 <DanCon> wierd. I would have expected tht title to apply to the document, i.e. http://www.w3.org/2003/01/geo/wgs84_pos

22:28:14 <DanCon> I don't usually use URIs like http://www.w3.org/2003/01/geo/wgs84_pos#

22:28:35 <danbri> no?

22:28:39 * danbri open to suggestions

22:28:57 <danbri> it was an <Ontology rdf:about=""> sorta thing...

22:28:59 <danbri> (phone)

22:29:18 <DanCon> about="" expands to something without a #

22:29:31 <DanCon> Revision 1.13 2003/02/19 22:27:27 connolly

22:29:31 <DanCon> relaxed domain constraints on lat/long/alt from Point to SpatialThing

22:29:49 <danbri> (thanks)

22:29:54 <DanCon> sure

22:33:26 <DanCon> hmm... [[ The optional "TZURL" property is url value that points to a published

22:33:26 <DanCon> VTIMEZONE definition. ]]

22:40:58 <DanCon> ugh... more closed-worldness in ical: timezone rules only specify the when you *start* observing daylight time or standard time; no ends are given.

22:41:50 <DanCon> you have to conclude from the syntax that there are no more Vtimezone subcomponents; i.e. it should use parseType="Collection" in .rdf

23:01:00 <Jibbler> Jibbler is now known as Jibbler|sleep

23:36:48 <DanCon> woohoo! my timezone rule finally worked...

23:36:50 <DanCon> "2002-06-30T09:00:00" <#_g0> "2002-06-30T05:00:00Z" .

23:37:02 <DanCon> where g0 is the ny timezone

23:37:06 <danbri> congrats

23:37:42 <danbri> .google lbasic

23:37:43 <datum> lbasic: http://www.cyber-matrix.com/lbasic.htm

23:38:02 <danbri> bummer, hoped it wasn't used...

23:38:03 <DanCon> hmm.. how to say "nytime relates T1 to T2" without using literals as subjects...

23:38:07 <DanCon> this is nary.

23:39:40 <DanCon> zooming out, it converted...

23:39:42 <DanCon> :dtstart [

23:39:43 <DanCon> :dateTime "2002-06-30T09:00:00";

23:39:43 <DanCon> :tzid "/softwarestudio.org/Olson_20011030_5/America/New_York" ];

23:39:45 <DanCon> ... to ...

23:39:50 <DanCon> k:startingPoint [

23:39:50 <DanCon> dt:date "2002-06-30T05:00:00Z" ];

23:39:51 <DanCon> .

23:44:17 <DanCon> hmm... to get around literals as subjects, I think I'll use timezones to relate TemporalThings to local-time-strings.

23:46:58 <DanCon> there... now we have:

23:47:08 <DanCon> <#_g0> :tzid "/softwarestudio.org/Olson_20011030_5/America/New_York".

23:47:13 <DanCon> [ <#_g0> "2002-06-30T09:00:00";

23:47:13 <DanCon> dt:date "2002-06-30T05:00:00Z" ].

23:52:09 <minddog> minddog is now known as minddog[out]

23:57:49 <DanCon> ok, I cut the cyc:startingPoint and some indirection out...

23:58:07 <DanCon> :dtend [

23:58:07 <DanCon> dt:date "2002-06-30T06:30:00Z";

23:58:07 <DanCon> :dateTime "2002-06-30T10:30:00";

23:58:07 <DanCon> :tzid "/softwarestudio.org/Olson_20011030_5/America/New_York" ];