#topicmaps@irc.freenode.net log for 2003-04-20

This log is automatically generated by an IRC bot from the traffic on the #topicmaps IRC channel on the irc.freenode.net IRC server. This file has the traffic for 2003-04-20. If you have questions regarding this log, please contact larsga@ontopia.net.

03:41:11 mariyo mariyo has quit None (Remote closed the connection)
05:24:07 FloK FloK has joined #topicmaps
05:42:59 FloK_ FloK_ has quit None (Read error: 110 (Connection timed out))
06:16:21 mariyo mariyo has joined #topicmaps
06:29:57 GWa GWa has joined #topicmaps
06:48:00 drrho drrho has joined #topicmaps
07:46:03 mariyo mariyo has quit None (Remote closed the connection)
07:59:55 GWa GWa has quit None ("leaving")
08:42:15 the^chosen^one the^chosen^one has joined #topicmaps
08:42:20 the^chosen^one the^chosen^one is now known as erica
08:43:10 FloK FloK is now known as FloK-away
08:43:45 erica hi drrho
09:09:58 mariyo mariyo has joined #topicmaps
09:14:46 erica hi mariyo
09:39:27 mariyo hi erica :)
09:39:35 mariyo you still there?
09:40:14 mariyo i'm busy working on the National body comments - lost in concentration :)
09:40:30 mariyo this is not easy!
09:42:34 mariyo erica: if you see that i am online and i don't reply just send me a private message; i get a telephone ring and it brings me to attention.
10:17:58 erica erica has quit None ()
12:17:43 drrho drrho has quit None ("Client Exiting")
13:00:02 the^chosen^one the^chosen^one has joined #topicmaps
13:00:07 the^chosen^one the^chosen^one is now known as erica
13:00:27 mariyo hi erica :)
13:00:32 erica hi mariyo :)
13:00:39 mariyo did you get the templates?
13:00:45 erica have not check my mail
13:00:47 erica * erica checking mail now
13:01:37 erica nope did not get any mail
13:01:49 erica * erica -> dinner
13:28:18 mariyo erica: i sent the mail out again and cced myself. it seems ok. check your esantosa bond mail.
13:33:47 erica erica has quit None ()
15:53:51 FloK-away FloK-away has quit None (Remote closed the connection)
16:28:32 tmkal tmkal has joined #topicmaps
16:41:29 larsbot hi there, kal
16:41:46 tmkal hi larsbot
16:42:21 tmkal i was wondering if you were actually online :)
16:42:30 tmkal wonder if you could have a look at something for me...
16:42:59 larsbot sure
16:43:03 tmkal http://www.techquila.com/tmsinia.html
16:43:24 tmkal I've been playing around with UML notation for describing topic map constraints
16:43:38 tmkal http://www.techquila.com/tmsinia2.html describes it in more detail
16:43:53 tmkal but I also finally got around to doing PSIs for thesaurii amongst other things too
16:43:57 larsbot oh, cool
16:44:07 larsbot actually, I'm doing a TM ontology for a customer as UML, too
16:44:13 larsbot have a script that converts it to XTM and OSL
16:44:17 tmkal (see the news item http://www.techquila.com/index.html)
16:44:20 tmkal cool!
16:44:28 the^chosen^one the^chosen^one has joined #topicmaps
16:44:34 the^chosen^one the^chosen^one has left #topicmaps ()
16:44:44 tmkal I just use it for digramming at the moment
16:45:59 larsbot very cool. we need something like this
16:46:39 tmkal aye - it came up while I was doing some work for a client
16:47:05 tmkal more patterns would be a good thing
16:47:22 larsbot absolutely, and especially a convention for how to write them up
16:47:26 tmkal BTW I think that the hierachy pattern can extend ontopia's transitive PSI
16:47:41 tmkal way to write them up: yes
16:47:45 tmkal ideally a DTD :)
16:47:52 larsbot uh, well :)
16:47:59 tmkal but at least a set of guidelines as to what goes in there
16:48:21 larsbot I think what most people need to get started is an example
16:48:25 tmkal yeah
16:48:31 tmkal thats why I haven't announced it yet
16:48:45 larsbot ?
16:48:47 tmkal I'm going to do an example or two of each pattern and put them into the document
16:48:57 larsbot ah, I see :-)
16:49:04 tmkal oh you mean example of pattern write-up ?
16:49:15 tmkal or example of pattern ?
16:49:15 tmkal :)
16:49:19 larsbot the former :)
16:49:33 larsbot but they do need both
16:50:02 larsbot if you'd used <link rel=next ...> I could have read this as if it were a single document in Opera 7.1
16:50:06 larsbot something to consider, perhaps
16:50:28 tmkal indeed
16:50:57 tmkal I added code to my site generation stylesheet to stick in link rel=home
16:51:10 tmkal so many features to implement...so little time :)
16:51:36 larsbot yeah :)
16:52:20 larsbot hmmmm. isn't it better to not scope the perferred term at all?
16:52:40 larsbot scope is restriction on applicability, so to leave it unrestricted seems to me best
16:52:47 tmkal makes sense, yes
16:54:02 larsbot I like your approach with two alternatives, though
16:54:17 tmkal cule!
16:54:43 larsbot this is great stuff
16:54:45 tmkal i think translating between them would be pretty easy with a topic map toolkit
16:54:47 tmkal thanks!
16:55:00 larsbot yeah. once TMQL has modification it should be dead simple
16:55:23 larsbot btw, one thing to add may be XTM documents that add names to the topics
16:55:29 tmkal tell you what, I might just do a jython script using TM4J - I guess is might eb quicker :)
16:56:01 tmkal the topics that reify the published subjects ?
16:56:30 larsbot certainly would give you the result a lot sooner :)
16:56:34 larsbot yeah
16:56:46 tmkal thats a good idea
16:56:59 larsbot and maybe even an OSL schema to document the constraints...
16:57:19 tmkal yeah, thats a good point too
16:57:45 larsbot hmmm. why represent the facet-has-root relationship explicitly?
16:57:49 larsbot it's redundant...
16:58:18 tmkal in what way ?
16:59:08 larsbot if you can reach the same topic by traversing up the hierarchy you already have that info, no?
16:59:14 tmkal * tmkal scribbles suggestions on a bit of paper
16:59:23 tmkal yes, that is true
16:59:43 tmkal I just thought it would be good to be explicit
16:59:45 larsbot btw, you may want to look at this: http://xfml.org/
16:59:50 tmkal I already did
16:59:54 larsbot aha :-)
16:59:58 tmkal I got most of my ideas from it :)
17:00:08 larsbot because an XFML->XTM converter using these PSIs would be wildly cool
17:00:17 tmkal * tmkal scriblles note to ad xfml to bibliography
17:00:24 tmkal it would indeed
17:00:34 tmkal prob. easy enough with XSLT
17:01:05 tmkal the reason for having explicit root is that it might be that you would have two hierarchies
17:01:10 tmkal with the same hierarchy association
17:01:15 tmkal but different facets
17:01:16 larsbot yep. already have one in the OKS CVS tree that uses SAX and builds TM directly
17:01:27 larsbot how so?
17:01:48 larsbot see also http://xfml.org/spec/1.0.html#topicfacetid
17:01:52 tmkal say you are classifying car parts
17:02:06 tmkal use a part-whole relationship to identify part components
17:02:22 tmkal but also a part-whole relationship to identify place of manufacture
17:02:43 tmkal (e.g. longbridge plant is part of birmingham, birmingham is part of uk)
17:03:15 tmkal so two facets could be using the same hierarchy association type
17:03:30 tmkal the only way to distinguish the hierarchies is by root
17:03:34 larsbot sure
17:03:48 larsbot but you can always find the root through traversal, so...
17:05:23 tmkal I don't understand - where would you start traversal from
17:05:48 larsbot the topic itself
17:06:10 larsbot you could make a tolog inference rule that would always get you the root to which a topic belongs
17:06:24 larsbot or maybe I didn't understand you
17:06:36 tmkal ok I see what you mean
17:06:53 tmkal it would be possible to do that, yes
17:07:11 tmkal but I think if you wanted to construct a top-down view then having the
17:07:17 tmkal facet roots explicit would be a good thing
17:07:51 tmkal actually without knowing the roots, it would be impossible to tie a hierarchy to a facet
17:08:01 tmkal back to the car parts - if I have a part topic
17:08:25 tmkal and it is associated with a location topic and an assembly topic
17:08:34 tmkal which are both categories in different facet hierarchies
17:08:42 tmkal how do i tell which one belongs to which facet ?
17:09:11 tmkal just the association types are not going to be enough cos they are both using the part-whole hierarchy type
17:09:27 larsbot by traversing from the location/association topic upwards to the root you'd find the root
17:09:36 tmkal you would
17:09:41 tmkal but then which facet is it ?
17:10:00 tmkal you could traverse up to "World" at the top of the location hierarchy
17:10:10 tmkal and to "Car" at the top of the assembly hierarchy
17:10:18 tmkal but which is which ?
17:10:29 tmkal the root != facet topic
17:10:38 larsbot aha. that's the piece I was missing
17:11:16 larsbot oh, I understand now. I thought you followed XFML, but it seems you don't
17:11:22 tmkal ok, that would probably have been a whole lot easier if I had made an example :)
17:11:30 tmkal XFML has markup for a "facet"
17:11:30 larsbot yeah :-)
17:11:47 larsbot you connect World to Location, but not, say, Swindon to Location, right?
17:11:58 tmkal yes
17:12:12 larsbot ok. sorry, then everything I said was just wrong
17:13:01 larsbot anyway, the sooner you publish this, the better, I say :-)
17:13:02 tmkal I think that XFML requires that every topic in a hierarchy refers to a facet
17:13:05 tmkal :)
17:13:18 tmkal I guess I should just do some simple examples
17:13:33 larsbot it does, which was what led me astray here
17:13:52 tmkal maybe it is worth pointing that out in a note somewhere then
17:14:19 larsbot probably is. an example would also help
17:14:23 tmkal yeah
17:14:33 larsbot do you want this in the blog, btw?
17:15:21 tmkal why not ? :)
17:15:31 tmkal I already put the PSIs on easytopicmaps
17:15:50 larsbot that's what I noticed :)
17:15:59 larsbot btw, having this thing as a single page would be very nice
17:16:18 larsbot tmbot: url: http://www.techquila.com/tmsinia.html
17:16:28 larsbot tmbot: title: Topic Map Patterns For Information Architecture
17:16:54 tmkal cule!
17:17:22 larsbot tmbot: comment: Paper by Kal Ahmed which introduces a UML-based pattern notation for topic maps and uses it to define published subject sets for thesauri, faceted classification, and hierarchical classification systems in general.
17:17:43 larsbot will show up at http://topicmaps.it.bond.edu.au/weblogs.mc within the hour
17:18:14 tmkal excellent
17:18:36 tmkal BTW: do you know if there will be BOF sessions at XML Europe ?
17:19:23 larsbot nope. I have to contact IDEAlliance about that
17:19:34 larsbot did you see the TM/RDF QL one in the blog?
17:19:47 tmkal no I didn't...
17:20:24 tmkal aha...now I have
17:20:56 tmkal I was considering a topicmaps/IA BOF
17:21:07 tmkal hence the paper - as a discussion point
17:21:26 larsbot oh, I see. that would be very cool
17:21:55 larsbot I'll cc you on the email to IDEAlliance, then
17:22:02 tmkal that would be great, thanks!
17:23:49 larsbot now sent
17:23:57 larsbot also, btw, will you attend any of the ISO meeting?
17:24:22 tmkal yeah
17:24:39 larsbot oh, great. we need support :)
17:24:43 tmkal :)
17:25:04 tmkal I'm afraid that I am likely to be not very politically correct :)
17:25:10 larsbot how so?
17:25:31 tmkal I don't think that the RM/TMM is ready for standardisation at all
17:25:36 tmkal I don't think its needed
17:25:51 tmkal I don't think it standardises anything - cos there is no prior art
17:26:01 tmkal SAM standardises a model that is already implemented by different tools
17:26:23 tmkal TMCL/TMQL will standardise languages for constraints and queries which are already implemented by different tools
17:26:23 larsbot privately I agree with all of that
17:27:14 tmkal I can see why people want to develop a TMM/RM
17:27:24 tmkal but it should be developed first, then standardised
17:27:52 tmkal unfortunately I don't think that that is the kind of thing that will go down too well...
17:28:00 larsbot there are many people who agree with you
17:28:08 xover xover has quit None (Excess Flood)
17:28:31 larsbot in fact, there is a possibility that some people will try to throw it out of 13250 altogether
17:28:39 larsbot don't know whether it will be attempted or not
17:28:51 xover xover has joined #topicmaps
17:29:17 tmkal I guess that it will be a political decision at the end of the day
17:29:30 larsbot yeah, definitely
17:30:21 tmkal anyway, I do agree that SAM should go to CD as soon as possible
17:30:31 tmkal things need to start happening
17:30:33 larsbot excellent. that's the important thing, IMHO
17:30:50 tmkal yeah, I can live with TMM/RM being done, as long as things don't get held up by it
17:30:59 larsbot exactly. that's my view as well
17:31:03 larsbot I'm really looking forward to getting started on the TMQL/TMCL work
17:31:09 larsbot I think that will be a lot of fun
17:31:17 tmkal yeah I agree
17:31:41 tmkal and CXTM needs to be done and announced
17:32:03 larsbot yup. Pepper-san has made good progress on CXTM
17:32:07 larsbot it's not ready, but getting there
17:32:31 tmkal yeah, I really would like to get time to implement it before the meeting, but its not looking too likely at the moment :(
17:33:09 larsbot don't think it's actually possible to implement it as it stands right now :-)
17:33:13 larsbot so don't worry :)
17:33:29 tmkal :)
17:34:03 tmkal have you got an issues list for it anywhere ?
17:34:20 larsbot nope
17:34:30 larsbot that reminds me I should get hold of the XML and process it into tm-standards.xtm
17:34:37 larsbot in fact, I need to update tm-standards.xtm
17:34:48 tmkal might be good to start a list somewhere too then...
17:35:03 larsbot can pull that out of the <issue> elements in the spec
17:35:22 larsbot actually, I should make an OKS web application for that TM
17:35:41 tmkal ah, that reminds me
17:35:51 tmkal another bar-discussion should be on a topic map Wiki
17:36:03 larsbot oh yes yes yes :)
17:36:06 tmkal I am thinking that a combination of Wiki + LTM
17:36:17 tmkal would rock
17:36:17 larsbot FloK, who hangs out here now and then, is into that, I think
17:36:24 larsbot it certainly would. no question about it
17:36:48 tmkal it would also be a nice test app for building with tm4web
17:36:52 larsbot it's just crying out for somebody to do it, IMHO
17:36:55 tmkal :)
17:37:14 tmkal I agree ...
17:37:42 larsbot there's also Sakurai-san in Japan who's keen on it
17:37:45 larsbot or at least was keen
17:38:29 tmkal cule
17:38:52 tmkal well, it probably won't get started before XML Europe, so perhaps we can have a design session :)
17:39:01 larsbot would be lovely :-)
17:39:14 larsbot btw, I did a rough draft of an XSLT-for-topic maps, based on tolog. wanna see?
17:39:27 tmkal yes please!
17:39:56 larsbot momento
17:41:30 larsbot mailed to you now
17:41:43 larsbot note that the thing is implemented, so I have the example actually running
17:42:02 tmkal cule!
17:42:15 larsbot * larsbot was surprised at how easy it was
17:44:02 tmkal excellent stuff!
17:44:18 tmkal this is exactly the kind of thing that is needed to get non-programmers making pages with topic maps
17:44:32 larsbot very glad to hear you approve of it
17:44:40 larsbot was thinking this might belong as TMQL, part 3
17:45:04 tmkal why not make a proposal for a TMTL standard ?
17:45:23 tmkal actually, I wonder how this would play with XSLT2 ?
17:45:43 tmkal * tmkal hasn't actually read the whole XSLT2 spec yet
17:45:56 larsbot we could do a separate standard as well. not sure it matters
17:46:03 larsbot hmmm. what are you thinking re XSLT2?
17:46:09 tmkal Isn't there some extensibility stuff in XSLT2 ?
17:46:13 larsbot oh, I see
17:46:18 larsbot there is in 1.0, too
17:46:23 tmkal I wonder if it is possible to create a "tmql::" axis
17:46:29 tmkal or something like that
17:46:43 larsbot humm. would be cool to see an attempt, at least :)
17:47:44 tmkal I had a discussion a while ago with Tony Coates about doing something similar in XSLT 1.0
17:48:03 tmkal basically by hacking Xalan to add some topicmap-specific axes
17:48:14 tmkal which would be kind of cule
17:48:56 larsbot if you could get it to work it would be very cool
17:49:44 tmkal aye
17:49:56 tmkal its that lack of time thing again... :(
17:50:02 tmkal http://www.w3.org/TR/xslt20/#extension
17:50:03 larsbot yeah, or lack of people
17:50:20 tmkal looks like you can define your own XSLT elements
17:50:31 larsbot you can in 1.0, too
17:50:43 larsbot in my book I show how to implement extensions in several XSLT processors
17:50:46 tmkal * tmkal learns something new every day
17:50:54 larsbot both extension elements and extension functions
17:51:28 tmkal I though in XSLT1 extension mechanism wasn't defined by the spec ?
17:51:50 larsbot it is: http://www.w3.org/TR/xslt#extension
17:51:57 tmkal ah
17:52:23 larsbot not that it makes your idea any less interesting :)
17:52:27 tmkal :))
17:52:47 tmkal well, I would support a move to make a TMTL standard that builds on TMQL
17:53:10 tmkal rather than delay TMQL by wrangling over the mechanics of a templating language
17:53:23 larsbot yeah, agreed. though they should probably be designed together
17:53:30 larsbot the one interacts strongly with the other
17:53:51 tmkal perhaps
17:53:57 larsbot the TMTL proposal I put forward is subtly different from XSLT because tolog is different from XPath
17:54:38 tmkal TMTL seems to essentially be extracting strings rather than sets of objects
17:54:49 tmkal except in the for-each / if tags
17:55:33 tmkal if you did an XPath extension for topic maps, it might look more like SAM-access
17:56:09 larsbot prolly would, yes
17:56:20 larsbot TMTL: it does both, actually
17:56:29 larsbot sometimes you pick out objects, other times strings
17:56:45 larsbot note that I introduced occurrence predicates in tolog
17:56:47 tmkal I can't wait for the new tolog spec ;)
17:56:57 larsbot doing it as my Montreal paper :-)
17:57:04 larsbot occurrence-type(TOPIC, LOCATOR/VALUE)
17:57:20 larsbot so you can say email(kal, $EMAIL)? and get it
17:57:38 tmkal yeah, thats more or less how I hoped it would work
17:58:22 tmkal works nicely with the "occurrence as association" view of the world too
17:58:44 larsbot it does
17:58:53 larsbot tolog is actually a bit like the RM: everything is a predicate
17:59:00 larsbot and a predicate is pretty much an association
17:59:09 larsbot what's cool is that the RDF QLs seem to follow the same model
17:59:30 tmkal great minds...
18:00:38 tmkal * tmkal is off to make some tea
18:00:44 larsbot yeah. actually it's just datalog adapted to a different model
18:09:08 larsbot * larsbot starts on presentation for XTM discussion at ISO meeting
18:09:57 tmkal * tmkal is finally writing his slides!
18:10:04 larsbot ooh :)
18:10:10 tmkal :)
18:10:18 larsbot haven't started on mine yet, but I have *piles* of them to do
18:10:18 tmkal thats why i needed the tea first :))
18:10:24 larsbot makes sense :)
18:10:30 larsbot * larsbot drinking tea as he writes, too
18:13:53 larsbot * larsbot wonders how to get rid of OpenOffice assistant
18:15:06 tmkal feed it to clippit !
18:15:20 larsbot clippit?
18:15:31 tmkal the ms office assisstant
18:15:39 tmkal put them in a room and let them fight it out
18:16:14 tmkal alternatively, simply use emacs :)
18:16:29 FloK FloK has joined #topicmaps
18:17:04 larsbot attractive idea, but I'd have to install Windows to do it :-)
18:17:06 larsbot hi FloK
19:27:30 FloK hi larsbot
19:28:00 FloK i've been sleeping zzzzzz...
19:28:39 larsbot no problem
19:28:47 larsbot were you thinking of doing a topic map-based wiki?
19:29:02 FloK I had a nice but strenuous day with my girlfriend and her children :)
19:29:14 FloK wiki: yes :)
19:29:17 larsbot sounds very good :-)
19:29:21 larsbot * larsbot thought so
19:29:26 FloK :)
19:29:28 larsbot tmkal is interested in the same thing
19:29:34 larsbot could you tell us about what you've thought/done?
19:29:44 FloK I need it. Sometimes i lose my way in my wikik ;)
19:30:07 larsbot yeah, I also think it would be fantastically cool
19:30:11 FloK cool. - Hi tmkal!
19:30:43 larsbot how far have you progressed with the idea?
19:31:12 FloK i thought about inventing further relationships between pages..
19:31:19 FloK ie "see also"
19:31:46 FloK ZWiki already has a "isParent" rel.
19:32:01 larsbot right. perhaps adding 'type' to the pages would also be useful?
19:32:07 FloK so one can build a hierarchy
19:32:26 FloK what type for example?
19:32:47 FloK the type is the rendering type at the mom.
19:33:11 FloK stx, hmtl, plain, sh, sp, hp, blabla
19:33:24 larsbot things like FloK : person, tmkal : person, Ontopia : company
19:34:05 larsbot do you have an implementation of this, btw?
19:34:31 FloK about the page title - thistitle: topic
19:34:43 FloK no:(
19:34:57 FloK I cant get ZTM to work
19:35:03 larsbot too bad :-(
19:35:13 FloK and i havent heard anything from arnarl
19:35:18 FloK yes :(
19:35:21 larsbot he's probably on vacation
19:35:39 FloK i see. good for him :)
19:36:04 tmkal hi FloK! Sorry - I was engrossed in playing with fonts on my other machine :)
19:36:06 FloK I am not good enough in zope do to it myself.
19:36:17 FloK not yet... :) soon!
19:36:42 larsbot "do it" meaning "implement ZTMWiki" or "get ZTM running"?
19:36:49 FloK next week i talk with a company about working for them...
19:37:11 FloK they do zope and Knowledge M., and i wrote them about the idea..
19:37:24 FloK and they invited my the same day :)
19:37:36 FloK ..em for next week
19:38:03 FloK get it running, even
19:38:22 tmkal I had the same problem with ZTM...
19:38:38 tmkal I hacked at it a bit, but had to give up in the end
19:38:50 FloK hmm :(
19:38:52 larsbot oh, that's really cool. hope you get a job there :-)
19:39:03 tmkal shame cos I would like to have a running ZTM installation on my machine...
19:39:14 FloK i will. i Really want it!
19:39:51 FloK what did you do tmkal? same as me , something else?
19:40:07 tmkal yeah
19:40:22 tmkal I ended up doing what I needed with Struts + TM4J + Velocity
19:40:40 tmkal but if I had got it working with ZTM it would have had loads more features...
19:40:41 FloK ?
19:40:52 larsbot what was it you did?
19:41:05 tmkal just an example KM portal thingy for a client
19:41:10 tmkal but it ended up being read-only
19:41:21 tmkal cos I didn't have time to implement any form of editing
19:41:36 tmkal thats when I started thinking about tm-based wikis
19:41:57 FloK the portal was in zope?
19:42:13 tmkal the client is not at the stage of choosing a technology yet
19:42:18 tmkal so it was up for grabs
19:42:28 tmkal in the end they will prolly go for a commercial package
19:42:45 tmkal but right now they need something to do more proof-of-concept type work
19:43:14 FloK where are you loc. tmkal?
19:43:19 tmkal oxford, uk
19:43:32 FloK ok.
19:43:49 tmkal how 'bout you ?
19:44:13 FloK next city to munich, s-germany
19:44:40 tmkal ah!
19:45:06 tmkal I guess if you are looking for work you are prolly not planning on getting to XML Europe :(
19:45:32 FloK Do you want to try ztm again? I'll setup a new zope, maybe I'll get some help from germans.
19:46:02 FloK when and where?
19:46:07 larsbot 5-8 May, London
19:46:32 FloK yep, probably not because of work.
19:46:35 FloK :(
19:46:41 larsbot too bad :-(
19:46:45 tmkal shame
19:46:49 FloK would be a good opportunity to visit london ..
19:46:54 FloK never been to UK!
19:47:17 tmkal have to arrange it some time then!
19:47:19 tmkal :)
19:47:40 tmkal ztm: prolly going to see if I can implement something with TM4J instead
19:47:56 FloK TM4J ?
19:48:04 tmkal http://tm4j.org/
19:48:07 larsbot would be vastly easier with ZTM, methinks, but competition is good :)
19:48:15 tmkal :))
19:48:33 tmkal you overestimate how hard it would be with TM4J ;-)
19:48:34 FloK :)
19:48:54 FloK * FloK checks site
19:49:16 larsbot possibly :)
19:49:18 tmkal plus using tm4j, I can make use of the cool new SVG and VRML renderings for topic maps :)
19:49:25 larsbot whoo :)
19:49:45 FloK VRML ? wow.
19:50:11 FloK is VRML still ontopic?
19:51:21 tmkal well I must admit I don't know too much about VRML - another contributor to the project developed the renderer
19:51:26 FloK FloK has quit None (Remote closed the connection)
19:51:38 larsbot FloK crashed in astonishment :-)
19:51:42 tmkal but apparently there are a bunch of plugins that can display it...
19:52:59 FloK FloK has joined #topicmaps
19:53:11 larsbot do you have any examples of the visualizations?
19:53:16 FloK uff xchat crashed :(
19:53:52 FloK I dont get along with the TM4J site navigation.
19:54:06 FloK Is there content in the FAQ? Cannot see any.
19:55:45 tmkal there isn't a lot of content there yet
19:55:50 tmkal yeah, navigation needs sorting out
19:56:02 tmkal I have a plan for that, just need to find the time to do it
19:56:15 tmkal examples of visualisation: nope
19:56:42 tmkal Harald posted a message about it a couple of days ago, I haven't had chance to d/l and play with it...
19:57:22 tmkal prolly the best docs to look at about tm4j is the dev. guide
19:57:58 tmkal http://tm4j.org/docs/devguide/
19:58:40 tmkal TopicMapViewer announcement: https://sourceforge.net/mailarchive/forum.php?thread_id=1984501&forum_id=30220
19:58:42 FloK ah! :)
20:01:54 larsbot very cool :-)
20:02:19 tmkal I'm d/l the WAR file and an updated tomcat now
20:02:41 tmkal so maybe tonight I'll be able to put some sample SVG files somewhere if you are interested
20:08:37 larsbot I would be
20:09:18 larsbot btw, I'm finishing up a set of slides for the ISO meeting about the new XTM
20:09:25 larsbot anyone want to have a look through them?
20:10:09 tmkal I might take a quick look later on tonight if you can put them somewhere/email them to me
20:10:52 larsbot will email them, I think
20:11:03 tmkal cule
20:11:03 FloK * FloK installs ZTM on a fresh zope.
20:17:53 FloK FloK has quit None (Remote closed the connection)
20:18:16 tmkal gotta go have some dinner...back in a while
20:18:20 tmkal tmkal has quit None ("Client Exiting")
20:21:13 FloK FloK has joined #topicmaps
20:21:27 FloK re
20:25:12 larsbot xchat 2 is not yet stable, I see :)
20:26:01 FloK LOL
20:26:29 FloK It's called xchat-2, because it crashes two times a day.
20:26:41 FloK so for today i'm ok
20:26:43 FloK :)
20:27:41 larsbot lucky you :)
20:27:54 FloK :9
20:30:18 FloK FloK has quit None (Remote closed the connection)
20:32:51 xover xover has quit None (Excess Flood)
20:33:05 xover xover has joined #topicmaps
20:42:52 FloK FloK has joined #topicmaps
20:46:00 FloK arg! xchat upgraded to v.3 ;-)
20:46:01 FloK larsbot: xmtal quit?
20:46:01 FloK need his email. - ah, it's on the TM4J site, ok
20:46:01 FloK I'm fine.
20:46:02 FloK cu.
20:46:02 FloK FloK has quit None (sterling.freenode.net irc.freenode.net)
20:47:21 FloK FloK has joined #topicmaps
20:52:26 larsbot version 4 released already? :)
21:06:40 larsbot * larsbot finishes presentation about XTM
21:07:11 FloK probably coming soon... :-/
21:08:13 FloK But i check out for today, before i am checked out by xchat... :) ... cu!
21:08:34 larsbot bye :)
21:08:46 FloK It was nice to cu, Lars!
21:08:51 FloK bye
21:08:54 FloK FloK is now known as FloK-away
21:34:00 tmkal tmkal has joined #topicmaps
21:51:49 larsbot hullo
21:58:55 larsbot * larsbot installing batik
22:05:59 larsbot grrrr! batik won't run
22:10:56 larsbot ontopia.svg is not well-formed XML: encoding declaration is wrong
22:11:35 tmkal doh!
22:11:46 tmkal Acrobat's plugin doesn't care :))
22:12:12 tmkal I guessed that the char encoding was the problem
22:12:24 larsbot looks pretty nice, though
22:12:31 tmkal does oks.svg or graham.svg work ?
22:12:34 larsbot they still have some work to do, but this quite cool even so
22:12:39 larsbot * larsbot got them all to work
22:12:50 larsbot fixed ontopia.svg by hand :)
22:12:54 tmkal yeah, there are some things that could be done to it to make it nicer
22:12:55 tmkal :)
22:13:23 tmkal but its interesting to see that this was all built with TMNav even in its current pre-pre-pre production state :)
22:14:33 larsbot TMNav?
22:14:59 larsbot * larsbot discovers that sodipodi is a nice drawing tool using SVG as native format
22:15:37 tmkal TMNav started out as a desktop app for displaying topic maps
22:15:48 tmkal but evolved into a framework for topic map applications
22:16:02 tmkal using an abstraction/renderer model
22:16:17 tmkal abstractor generates a graph from the topic map using the abstraction rules
22:16:39 tmkal the renderer turns the graph into VRML/SVG/HTML/TouchGraph nodes
22:16:55 tmkal there is even a WS interface built on it !
22:17:08 larsbot oh, cool
22:17:27 larsbot how do you specify abstraction rules?
22:17:40 tmkal they are coded in Java at the moment
22:17:49 tmkal you can chose an abstractor implementation
22:18:02 tmkal they all use the same interface so in theory are interchangeable at run-time
22:18:28 larsbot right. and in theory you could have a language for writing them in
22:18:42 tmkal yeah
22:18:59 tmkal in fact you could probably make something quite interesting using Tolog
22:19:41 tmkal the nice thing about the abstraction layer is that the abstract graphs are easier to process than the source topic map
22:19:44 larsbot actually, if you had an XML language for them you could do TMTL -> language-X
22:19:57 larsbot anything is easier to process than a TM :-)
22:20:03 tmkal :))
22:20:27 tmkal you could do TMTL -> Language-X in theory I suppose
22:20:38 tmkal depends on how expressive you make Language-X
22:21:27 larsbot yeah. I don't know what your graphs look like, so I don't know how practical it would be
22:22:03 tmkal thats what I am wondering
22:22:14 tmkal they are basically labelled nodes with labelled arcs
22:24:00 larsbot sounds pretty easy, if you ask me
22:24:49 tmkal aye
22:25:15 tmkal the XML rendition should be really easy for users to work with (e.g. in XSLT)
22:25:36 tmkal I think that the task of defining an abstraction would be more of a programming thing
22:25:53 tmkal but it would be nice if you could make the language as close to the topic map model as possible
22:25:54 larsbot sure, but if you could do it declaratively that would be a win
22:25:57 tmkal yep
22:26:16 tmkal so something based on a TM QL would be ideal
22:26:38 larsbot yeah. we need a standard TMQL, and we need it soon
22:26:51 larsbot * larsbot working on slides for TMQL presentation in London right now
22:27:27 tmkal * tmkal messing about with fonts when he should be writing his presentation content :)
22:28:06 tmkal make sure you add a bullet in about abstractors for TM4J
22:28:07 tmkal :))
22:28:11 larsbot hehe :-)
22:28:18 larsbot which reminds me, did you look at the slides I sent out?
22:28:28 tmkal I haven't yet...
22:28:38 tmkal * tmkal is opening them now
22:32:01 tmkal N0398 = ?
22:32:10 tmkal I lose track of all these numbers :)
22:33:15 larsbot N0398 is the latest XTM syntax draft
22:33:19 tmkal ok
22:33:40 tmkal xtm-href-pointer: Is the proposal to only *require* bare name XPointer syntax
22:33:48 tmkal or to only *allow* base name XPointer syntax ?
22:34:21 tmkal my feeling is that an XTM processor should only be required to implement bare names
22:34:22 larsbot to only allow
22:34:27 larsbot * larsbot fixes wording
22:34:33 tmkal ah
22:34:46 tmkal then I'm not sure I like the recommendataion then :)
22:35:14 larsbot why not?
22:35:41 tmkal It is unecessarily constraining - surely if people want to put full XPointer syntax in there they should be allowed to
22:36:07 tmkal so XTM should allow them to do it
22:36:18 larsbot but then we have to support it...
22:36:22 tmkal no
22:36:31 tmkal you can say that a processor need only support bare names
22:36:45 tmkal and say nothing about what a user is allowed to put in the field
22:36:51 larsbot what's the point if none of them are going to support it?
22:36:59 tmkal none of them right now are
22:37:16 larsbot can't imagine that any of them will, either, since to support XPointer you have to build a full DOM
22:37:17 tmkal but perhaps someone will implement support for a subset of XPointer that can be processed in a streaming parser
22:37:32 tmkal memory gets cheaper all the time :)
22:37:46 tmkal one day the hardware will catch up with the DOM :))
22:37:49 larsbot sure, but where's the point? I don't see any gain anywhere
22:38:04 tmkal exactly, there is no gain in being over restrictive either
22:38:10 tmkal so be less restrictive
22:38:19 tmkal say what is *expected* of XTM processors
22:38:21 tmkal and set the bar low
22:38:31 larsbot * larsbot is uneasy about optional feature
22:38:37 larsbot XML did that, and look where they ended up
22:38:52 tmkal aye, but look at the take-up
22:39:07 tmkal cos the cost of entry is low, but the headroom is there for additional power
22:39:24 tmkal how many browsers support full XPointer now ?
22:39:28 larsbot none
22:39:39 tmkal so why bother with XPointer then ? :)
22:39:57 larsbot that's what we're proposing, to not bother with it
22:40:08 tmkal it was a rhetorical question :))
22:40:30 larsbot * larsbot now thoroughly confused
22:40:40 tmkal XPointer exists cos *some* people need it
22:40:53 larsbot sure, but in XTM everything has IDs
22:41:22 tmkal but making a blanket restriction of xlink:href stops you using XPointer in a resourceRef
22:41:30 larsbot ah, sorry
22:41:39 larsbot it's only going to be disallowed in topicRef and mergeMap
22:41:42 larsbot I should have written that
22:41:48 tmkal you should ! :)
22:41:55 tmkal ok, then I can live with that
22:42:14 tmkal you have subjectIndicatorRef in an example too
22:42:34 larsbot only because it was difficult to make a sensible example involving topicRef
22:42:44 tmkal it is possible that you would want XPointer there and in resourceRef
22:42:51 larsbot I agree
22:43:03 tmkal ok
22:43:09 larsbot * larsbot now glad he sent this out for review
22:43:25 larsbot you probably shortened the meeting by a good 30 minutes now :-)
22:43:29 tmkal :))
22:45:34 tmkal slide 12 - second bullet under Considerations - shouldn't that be *not* backwards compatible ?
22:46:13 larsbot if we deprecate (as opposed to take out) it will be backwards compatible
22:46:18 larsbot will try to make that clearer
22:46:22 tmkal ok I see
22:46:51 tmkal perhaps this meeting should start by agreeing the principles of goodness and badness
22:47:03 tmkal i.e. make sure that every one agrees that backwards compatibility is good
22:47:18 tmkal otherwise you are in for a long meeting :))
22:47:42 larsbot we've pretty much done that already
22:47:55 larsbot I do agree. we've had some long meetings already :-)
22:48:03 tmkal :)
22:48:25 tmkal does anyone have a chess timer ?
22:48:55 tmkal you should bring one to the meeting and give one to each attendee
22:49:09 tmkal after they use up their ten minutes, they have to stay quiet and just vote :))
22:49:49 larsbot excellent idea :)
22:50:05 tmkal slide 16: xtm-mergemap-reference
22:50:20 tmkal has it been discussed what the target element of a mergeMap reference should be
22:50:28 tmkal i.e. should it be an error if the target is not a topicMap ?
22:50:39 larsbot it hasn't been discussed, no
22:50:43 tmkal that seems to be what this is saying, but from a different approach
22:50:55 larsbot it is; it sort of takes that for granted
22:51:05 larsbot do you think I should put in a slide about that before it?
22:51:15 tmkal yes
22:51:18 larsbot * larsbot do
22:54:22 tmkal slide 25: xtm-unknown-elements
22:54:31 tmkal all unknown elements or unknown elements in the xtm namespace ?
22:54:52 larsbot all unknown
22:55:05 tmkal I want xml in resourceData :P
22:55:17 tmkal * tmkal stamps foot repeatedly
22:55:32 larsbot argue for it at the meeting, then :)
22:55:37 tmkal :)
22:55:47 larsbot it's on slide 33...
22:55:47 tmkal I shall save my chess-clock time for that then :))
22:56:06 tmkal perhaps a forward ref should be put on slide 25 then
22:56:17 larsbot * larsbot do
22:59:53 tmkal slide 33 is xtm-subjectidentity-children...
23:00:06 tmkal slide 32 is xtm-resourcedata-markup
23:00:28 larsbot not any more; I added a new slide
23:00:41 larsbot at your request, even :-)
23:00:51 tmkal :)
23:01:15 tmkal I think allowing a topic to reify multiple resources is asking for trouble
23:01:54 larsbot referring to the last one?
23:02:01 larsbot well, not allowing it is also asking for trouble, unfortunately
23:02:12 tmkal what kind of trouble ?
23:02:25 larsbot quite often you'll find the same resources in multiple places
23:02:38 larsbot OASIS XMLvoc has this problem with specifications, for example
23:02:52 larsbot it also does away with the merge conflict problem
23:03:12 tmkal its a nasty way to do it though cos it mucks up the whole conceptual model
23:03:35 larsbot well, does it?
23:03:42 tmkal I like the fact that two copies of the same page are two separate subjects
23:03:49 tmkal it makes sense and it works
23:04:20 larsbot sort of, kind of, and then again not
23:04:58 larsbot I feel this is the right resolution, but I'm open to argument
23:05:03 tmkal if you want to talk about some abstract concept of the resource, then surely the thing you get by dereferencing the URI is actually a subject indicator, not the subject
23:05:19 tmkal if URL A and URL B are mirror copies of a document
23:05:29 tmkal and I want a topic which reifies the document
23:05:37 tmkal than URL A and URL B are subjectIndicators
23:06:04 tmkal the topic that reifies the resource at URL A is reifying a specific instance of the document
23:06:51 tmkal if you allow multiple resourceRefs, then you end up with two distinct ways of looking at the reification of resources being conflated into a single syntactic construct
23:07:13 larsbot I'm not sure you can draw the distinction that sharply
23:07:34 larsbot I used to think the same way myself, but looking closer at this I feel it's not so clear
23:07:51 tmkal its never going to be absolutely clear
23:07:56 larsbot there are many levels of interpretation of a resource and there's no such thing as the level where there is no interpretation but just the resource an sich
23:08:03 tmkal you are into the realms of URL metaphysics
23:08:21 larsbot exactly, and that's one reason why I felt it was better not to restrict too much
23:08:28 tmkal the existing way of dealing with that is to present a heuristic
23:08:45 tmkal no, I feel that in this case a solid model is needed
23:08:59 tmkal you need to say *this* is the way in which topic maps consider resources
23:09:13 tmkal and it has this effect on things like mirror copies
23:09:19 tmkal and if you don't like it, tough
23:09:26 tmkal :)
23:09:35 larsbot * larsbot considering argument
23:09:54 tmkal if you are too loose with this fundamental definition of subject identity, you end up in the position of RDF
23:09:57 tmkal :-&
23:10:01 larsbot hmmmm. I'm not convinced
23:10:15 tmkal thats ok, cos I know I'm right :)
23:10:31 larsbot I think this needs to be discussed. we'll see what the others say
23:10:48 larsbot make your own set of slides if you feel like it :)
23:10:54 tmkal :)
23:11:13 tmkal I might do that for the couple of issues I'm going to fight on :)
23:11:49 larsbot it would be welcome
23:12:06 larsbot otherwise there's a risk that gra and I will steamroller over the committee because we are prepared and they are not
23:12:21 tmkal who will be there ?
23:12:34 larsbot don't have a precise list
23:13:06 larsbot SRN, MB, Patrick, mariyo, drrho, Steve, Geir Ove, me, you, gra, I'd assume Bernard, Martin Bryan, at least
23:13:21 larsbot probably not everyone will be there all the time
23:13:30 larsbot oh, and I guess Sam Hunting will be there
23:13:44 larsbot naito-san will not :-(
23:13:45 tmkal aye, but probably enough resistance to the lmg/gra steamroller :))
23:13:48 tmkal :(
23:13:58 larsbot with you there: maybe
23:14:02 larsbot without you: no
23:14:21 tmkal * tmkal blushes
23:14:37 larsbot Steve is the only one who has much of an opinion
23:14:49 larsbot though admittedly this might be different with the syntax
23:14:58 larsbot we haven't actually discussed XTM before
23:15:35 tmkal it can get people going when you discuss attribute names and values and all that
23:15:42 tmkal as I recall from XTM 1.0 :(
23:15:49 tmkal * tmkal shudders with pain
23:16:03 larsbot yeah. thank god we don't have to come up with new names for anything :)
23:16:11 tmkal :))
23:16:12 larsbot we're just trying to agree what the hell it means
23:16:55 tmkal I think that the discussions around my sticking points will be interesting
23:17:07 larsbot I agree. a lot of the other stuff is pedestrian
23:17:14 tmkal especially the resourceRef one
23:17:26 tmkal I mean it will be interesting, not pedestrian
23:17:33 larsbot * larsbot understood :)
23:17:37 larsbot though the xtm-topicMap-and-resourceRef issue is an absolute bastard
23:17:55 tmkal yes
23:18:41 larsbot actually, I think the added themes stuff was a mistake
23:18:57 tmkal on mergeMap ?
23:19:00 larsbot yeah
23:19:08 larsbot both that and <mergeMap> smacks too much of end-user features
23:19:12 tmkal possibly...
23:19:14 larsbot rather than interchange syntax
23:19:30 tmkal it might be that its late, but I had trouble working out what the issue there is
23:19:36 larsbot HyTM was much simpler, and IMHO the better for it
23:19:48 tmkal agreed on that
23:20:06 tmkal at least it was simpler in that particular aspect ;-)
23:20:29 tmkal do you think you can add an example of the problem to the slides ?
23:20:30 larsbot yeah, archforms and HyTime made it not quite so simple after all :)
23:20:35 tmkal :))
23:20:41 larsbot I can put in some more text
23:20:49 larsbot and split the slide in two
23:20:56 tmkal cule
23:21:22 tmkal thinking about it a bit more, I think I'm not seeing the problem cos its not a problem for TM4J :-)
23:21:43 tmkal there is a potential to merge and then find that you have to un-merge
23:22:13 larsbot unmerge?
23:22:21 tmkal aye
23:22:25 larsbot oh, you mean, to add the themes?
23:22:28 tmkal yes
23:22:41 larsbot well, that won't work unless you preserve non-SAM information
23:22:59 tmkal yes
23:23:04 tmkal which TM4J does
23:23:16 tmkal cos each Topic object maintains a list of merged Topic objects
23:23:26 tmkal rather than modifying properties of the base Topic directly
23:23:55 tmkal so I can pick out the set of characteristics which came from topic map A and replace them
23:24:03 tmkal and when I do, TM4J works out the new merge set
23:24:11 tmkal all by magic :)
23:24:37 larsbot that's why you are 5-10 times slower than us and use lots more memory :-)
23:24:43 tmkal :)
23:25:18 larsbot it's a tradeoff, and going back to apply the scopes will be *very* costly
23:25:31 tmkal yep
23:26:23 tmkal I think that in truth the right way to handle this is to say that there are two inclusions of the topic map
23:26:38 tmkal one with no added themes and another with added themes
23:27:08 larsbot that raises a related issue: is it an error to have contradictory <mergeMap>s?
23:27:19 larsbot that is, refs to same external doc with different added themes?
23:27:41 tmkal not a processing error - but its something that a user might like to know about
23:28:00 larsbot I tend to feel the same way
23:28:03 tmkal in TM4J, no mergeMap or topicRef merge processing is done implicitly
23:28:15 larsbot what do you mean?
23:28:20 tmkal you can get a list of what topic maps were referenced and what added themes were specified
23:28:42 tmkal then choose to merge in the referenced topic maps on a one-by-one basis
23:29:06 tmkal so you can end up with a partially merged topic map if you want
23:29:21 tmkal of course, there is standard code to just merge everything
23:29:38 larsbot ah, I see. that makes sense
23:29:51 tmkal but would require user intervention
23:30:02 tmkal so must apps prolly end up doing the automatic merge of everything
23:30:11 larsbot pretty sure 99% of them will
23:30:16 tmkal yeah
23:30:23 larsbot we've assumed users are not very interested in the form of their XTM
23:30:26 larsbot so far we've been right
23:30:44 tmkal I think its prolly true for almost all apps
23:31:05 tmkal I think for topic map editing tools and maybe some gui tools you might want more fine-grained control though
23:31:21 larsbot you would, but probably not over the XTM
23:31:47 tmkal no, but you would want better control over the way that it is processed.
23:32:11 tmkal e.g. "Don't merge in language.xtm from the topicmaps.org site"
23:32:24 tmkal or "Never merge topic maps from www.techquila.com" :)
23:33:01 tmkal gotta go now
23:33:06 tmkal might be around tomorrow
23:33:11 larsbot okay. thanks a lot for the review. it was very useful
23:33:28 tmkal no problems, was nice to think about something other than my slides :)
23:33:32 larsbot :-)
23:33:35 tmkal tmkal has quit None ("Client Exiting")