Re: [tm-pubsubj] mixed feelings

[Lars Marius Garshol <larsga@garshol.priv.no>]

* Bernard Vatant
| 
| We've got lately lively and interesting debates about paradigmatic,
| generic, genetic or distributed PSIs ... but that does not make our
| practical work move forward an inch I'm afraid.

Agreed.

| A month has passed since Seattle meeting, and we have still that
| bunch of issues posted in the current Deliverable 1 version, and not
| a heap of proposals to address those issues so far.

Why not try a different approach, like the one David Megginson used
for SAX: start a thread for each issue, with a suitable pause between
each thread? 

This divides the workload up into more manageable pieces and makes it
easier for the hard-pressed TC members to contribute. Approaches that
require large investments (in terms of time and energy) before giving
any tangible result are much less likely to work.
 
| I've been trying to review the document those days, and figure how
| to make things move forward, and all I got to is adding those new
| questions, about non topic maps applications of PSIs, or distributed
| PSIs ...

That shouldn't worry us. These issues are hard, and will take some
time to be resolved. Anyone who thinks otherwise is (IMHO) nursing a
pleasant illusion. I don't think this is *too* hard, but it will take
time. 
 
| I wonder if we've not been far too deep into the details of the
| "what and how" without having a real agreement on "why and what
| for". Those have been expressed in a too fuzzy way so far. 

Absolutely. I've been unhappy about the requirements document all
along, believing it to be far too fuzzy around the edges, but felt it
would be more productive to shut up and let the work continue.

However, I don't see any reason why we cannot improve it as our
understanding of what we want to achieve improves.

| In fact, I think there is a deep, but not clearly expressed,
| disagreement on objectives between
| 
| (1) those who consider PSIs essentially as necessary logistics for
| interoperability and wide-scale support for topic maps technology -
| letting the use of PSIs outside TM universe as a secondary and minor
| objective.
| 
| (2) those who consider topic maps like maybe the reference early
| users, but only one among many technologies that could use PSIs.

Position (2) is actually a new thought to me. I didn't think anyone
was holding it, so I'm surprised to see you "come out of the closet"
in this way.

Your phrasing of (1) overstates my personal version of that view, but
I do think we should start with topic maps, and that the only
concession to other technologies we should make is to take care not to
make PSIs unusable to them.

If the user/developer communities of other technologies were to become
enthusiastic about our work I'd be happy to see them join it and bring
their own agenda items, but as long as they remain quiet the way they
are now it is very hard to speak on their behalf. So I think we will
have to settle for some form of (1) whether we want to or not, for
simple practical reasons.
 
| 1. Have our objectives been defined clearly enough?

I don't think so. I don't think we should restart the requirements
process, however. I believe we can do much more productive work by
starting discussion on and settling the issues you have brought to
light so far.

More will be coming from the GeoLang TC in the not too distant
future[1], but I consider that a good thing. I think we can settle
many of these issues now, and achieve momentum that way.
 
| 2. If yes, are we going the right way so far, or are we misled?

I think we are going the right way, we just have to keep going. I am
not in the least worried.
 
| 3. Do we have to reconsider the objectives, in the light of recent
| debates, or do we stick to the initial ones?

What recent debates? I see no need to reconsider, but perhaps I've
missed something (having been away for four weeks).

[1] <URL: http://www.oasis-open.org/committees/geolang/docs/todo.html >

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
ISO SC34/WG3, OASIS GeoLang TC        <URL: http://www.garshol.priv.no >