[topicmapmail] what I mean by "context" and "bottom-up"

[Martin Bryan]

Bernard

>It leaves me with new questions about the human aspect of the cooReso
project, let alone the technical viewpoint. Assuming there is no clear
coincidence between a context and a community, however we define both of
them - and I surrender under your strong arguments on that point - then who
is to be responsible and competent enough for setting ontology in a
definite context ?

Ah, the million dollar question, which we are struggling with at ebXML.  Only user communities are competent enough. Whether they are responsible enough is another question! The approach we took was of a "starter set" of obvious things and a user definable way of extending these. So we provided Transport/Air and the user can add to this Passenger/Ticketing. Before we agree that Transport/Air/Passenger/Ticketing is of general application then a "responsible body" (at present the Core Components team itself, but in the longer term a "formal mechanism" based on the use of "registration authorities". This is an area still under discussion.

> And BTW who is bound to define contexts and give
responsibilities or validate competences to begin with ?

Someone with years of experience in the area. We are using the UN CEFACT team, who have been working in the area for 20+ years. (Whether working on a problem without solving it for this long consitutes compentence I cannot say, but at least they have the most complete experience of the problems.)

>The answers to these questions if any, are not technical, but juridical,
not to say political.
Don't forget the aim is to build a cooperative, wide-range, public
web-index.

The problem is how to transfer to this knowledge into the web environment in a way that can be used interactively.

>The only reference for that matter, as far as I know, is dmoz. And it's
quite clear that dmoz  has not given
satisfactory answers to these questions : dmoz editors are more or less
"auto-declared experts". They volunteer to edit a category, and are
accepted or not on the basis of a short CV,
capacity to add two or three relevant URLs to the database, and write for
them a relevant description respecting spelling and typographic rules. The
range of their activity can quickly extend to wider categories when they
have proven able to follow guidelines. I became myself, after barely a
month of editing in
World/Français/Sciences/Astronomie, editor in World/Français/Sciences and
in Science/Astronomy, then in World/Français, meaning I could add resources
and create new categories in fields in which I barely knew anything. For
all these "promotions", I had to justify of no more competence than to
begin with.
Just asked, and was accorded the privilege, never knew on what ground nor
by whom anyway,
filled a form and two days after my login field was extended  ...

OASIS have set up a formal registries group that may provide a better solution at XML level.

>In dmoz, I heard often "we've done this" or "we should do that" ... but
never could figure who were "we", given all the previous considerations.
You have as I understand a long-standing experience in building and
implementing ontologies in quite different contexts. So when you say  "we
have to ..." "we need to ..." you know what you mean, I suppose, by "we"
..

>Honestly, who do you design by "we" ? What community ;-) ?

It needs to be some body representing a large set of communities. Which is whty the e-business work is being done at UN and CEN (the European Standards Body) level rather than at consortium level. 

Martin Bryan