Re: [xtm-wg] [XTM-CMS] Is there no data model?

["Chris Angus" <chris.angus@btinternet.com>]

Lars

Can I respond to your email as someone who has recently become involved with
the process as an invited expert to the XTM-CMS subgroup.  Along with
Matthew West (who also work for Shell) we have had a number of meetings with
members of the conceptual group to participate on the production of the
'conceptual model' as part of the XTM effort.

I agree pretty much with the core sentiment of your statement "To me, the
data model seemed absolutely crucial, the interchange syntax merely useful
and no doubt the rest will also be useful. However, as I see it, all else
pales into insignificance compared with the data model.", particularly the
first part of the statement.

I think that for 'conceptual model' you should read 'conceptual data model'
or 'abstract data model'.  Matthew and myself have been involved because of
our background in data modelling and our interest and work in areas such as
data and data model integration.

The abstract data model is currently being refined and produced.  Following
a number of meetings over the last two or three weeks Graham Moore is today
producing a revised set of diagrams and I am planning to spend much of the
weekend working on the accompanying words.  We hope to be circulating the
revised model during the course of next week for review and discussion.

You say "What further worries me is that the community does not seem to
consider this very important. (Ref the TMQL discussion as one example.) In
my opinion the community absolutely MUST make sure that we get an abstract
data model with proper specification of important operations like merging."

I agree.  I think that the people working on the model have refrained from
participating in the discussions on TMQL, merging, etc. because it is more
important to us at this point to complete the model so that we can then have
sensible discussions on such issues.

Chris Angus
Home: +44 (0) 16977 41504
SSI: +44 (0) 207 934 4960
Fax: +44 (0) 16977 41666
Chris.Angus@BTinternet.com or Chris.C.Angus@opc.shell.com
----- Original Message -----
From: "Lars Marius Garshol" <larsga@garshol.priv.no>
To: <xtm-wg@egroups.com>
Sent: 03 November 2000 12:46
Subject: [xtm-wg] [XTM-CMS] Is there no data model?


>
> [Initial note: I have largely been standing outside the XTM process,
> so it is quite possible that I have not gotten the story straight
> here, and I know that.  If I haven't got it right, please tell me.]
>
>
> As far as I understood, XTM was to produce an XML-based interchange
> syntax, a data model and some other stuff.  To me, the data model
> seemed absolutely crucial, the interchange syntax merely useful and no
> doubt the rest will also be useful. However, as I see it, all else
> pales into insignificance compared with the data model.
>
> It is now my impression that XTM _may_ produce a syntax specification
> by December 1st, that it has a so-called conceptual model and that
> there is no data model in sight at all. To be frank, this scares me!
> We have one implied-but-very-inadequately-specified data model in ISO
> 13250, and now it looks like we are going to have another in XTM 1.0.
>
> We have at least five different topic map implementations (tmproc,
> TM4J, the Ontopia TME, the Empolis TME and the Mondeca TME) based on
> these inadequately-specified data models, and there are even
> discussions of new standards to be built on this very sand-like base.
>
> How can we properly assess and discuss the changes the XTM syntax
> modifications do to the underlying model? How can we make sure that
> the five (or more!) topic map engines implement the same model? How
> can we make sure that applications built on top of the engines in
> various installations will not make wildly conflicting assumptions?
> Where is really the documentation of all the unspoken assumptions in
> these various standards?
>
> What further worries me is that the community does not seem to
> consider this very important. (Ref the TMQL discussion as one
> example.) In my opinion the community absolutely MUST make sure that
> we get an abstract data model with proper specification of important
> operations like merging.
>
> All this syntax stuff is all very well, but it is the data model that
> really matters and that just HAS (where is the 72pt purple blinking
> text when you need it?) to be done, and done soon.
>
> So, what can be done? And how quickly?
>
> If it sounds like I'm panicking it's probably because I am.
>
> --Lars M.
>
>
>
> To Post a message, send it to:   xtm-wg@eGroups.com
>
> To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com
>
>


-------------------------- eGroups Sponsor -------------------------~-~>
eGroups eLerts
It's Easy. It's Fun. Best of All, it's Free!
http://click.egroups.com/1/9698/4/_/337252/_/973263592/
---------------------------------------------------------------------_->

To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com