[sc34wg3] Almost arbitrary markup in resourceData

[Michel Biezunski]

On Mon, 2003-11-17 at 14:03, Mason, James David (MXM) wrote:

> I'd like to hear more from Martin and the other two original editors on what
> they thought their goals were for TM interchange.

Jim, 

Here is my view on this issue:

The original intent was to enable a very large basis for interchange,
and using HyTime as a basis guaranteed that there was much freedom
possible for interchange.

What happened later is the realization that HyTime tools were not
readily available and that providing real interchange would not
be very practical in the absence of HyTime engines. Also, the shift
from SGML to XML pushed us in the direction of trying to make things
simpler, more syntactically focused for interchanged (one fixed DTD)
and something that would be easy to implement from a Web perspective.

I believe XTM pretty much fulfills that requirement and that when
the SAM will become available, we'll have a strong basis for 
interchange.

Now I think the important is what's coming next. My view on this
is that now that we have several significant real TM applications
available, we need to take into account in detail the current
requirements for interchange. It may mean that we should enable
any markup instead of PCDATA and that the XTM processors should
treat them as processing instructions are handled by XML. 

My perspective on the interchange is that XTM should continue
to provide sound interchange for linking topic properties, and
the rest (local ontologies, formatting, display, etc.) should be
left to the applications -- at least temporarily. 

If there are other standardization layers that need to happen
we should discuss those as part of longer-term perspectives.
For example, I believe we should head in the direction of making
customized versions of XTM-like applications, where users could define
their own rules for merging, their own structure, while still
maintaining a certain level of interchange between topics.
When I say long term it's because I believe the market is not
ready yet for those, and we need both to take into account
the current requirements and make sure we are not getting
trapped in something that may become underpowered when users
will want something stronger.

As far as I am concerned, I am looking for the broader possible
level of interchange between applications, while still providing
a usability value for application designers. It's kind of a
catch 22 problem, but I think as time matures, and we all the
experience we gained, we are in a better position than ever
to start studying these possibilities. I don't think this
is very different from the original intent we had when we
started the topic map effort.

Michel
 
-- 
==================================
Michel Biezunski
Coolheads Consulting
402 85th Street #5C
Brooklyn NY 11209
Email: mb@coolheads.com
Web  : http://www.coolheads.com
Voice: (718) 921-0901
==================================