Peter Murray-Rust peter at
Thu Dec 25 14:29:46 GMT 1997

At 16:44 24/12/97 -0800, Andrew Layman wrote:
>I have to agree with Paul here.  If the interface is only implementable in a
>single language, then you've failed.  Programmers need to make engineering

I'd hate to use the word 'fail' (I'm serious about this). If we meet David
Megginson's criteria, then we have succeeded, possibly only in a smallish
way, but it's a success. It would be a success because:
	- it would show that there is a communal willingness amongst early
adopters of XML to collaborate. This, in itself would be a remarkable
achievement, since members of this list are not bound by the same
commitment as members of the W3C (and de facto the XML-SIG). It would show
that there are people and organisations that wish to work with
interoperable software and which are - in some measure - prepared to put
resource and effort into the communal pool.
	- it gives experience and guidance for what works in this environment. IOW
if David's criteria can be met, then they may be useful criteria for the
next time round.
	- it gives **working software** from which we can gain experience for the
next 'virtual project', whatever that is.

>tradeoffs among a number of factors, and will sometimes very reasonably
>choose one language over another.  We could expediently pick one language

I am very sympathetic to this concern. I wrote my earlier SGML system
(costwish) in tcl, and I'd be far happier writing the current (JUMBO)
graphics in tk than Java. 

>and ignore all others in order to simplify the problem a little, but that
>would be letting the solution dictate the problem. I don't recommend it,
>since it would simply mean that the parsers written in other languages would
>be guaranteed incompatible.

OK - Andrew and Paul have made a point. The question is what is to be done,
**practically**.  I assume that their postings, like all on this list, are
constructive.  What we cannot so, IMO, is to lose the momentum that we have
at present for Jan 12. Therefore I'd suggest that Andrew and Paul propose a
way of taking the Jan 12th result (which is only 2 weeks away) forward.

A primary point to realise is that the Jan12 milestone will address many
language-independent points. It will basically define what concepts come
under the simple API and what terminology is to be used. If we agree on
those then I would assume that creating a language-independent API/IDL
would be relatively straightforward. 

>From my point of view (I can't answer for DavidM or TimB) I would see a
possible way forward as:
	- Jan 12: Sax-J
	- shortly after: IDL for Simple API (SIDL?) based on experience with Sax-J
	- formulation of SIDL-J (Java binding for SIDL)
I have got sufficient public and private mail to suggest that this list
plays a useful role, but a central motivation - at least for me - is that
it is more than a talking shop. Discussion is easy, implementation is much
harder.  Therefore, for those who propose that we should have an IDL
approach (and this is not the first time it's been discussed) any
implication that they would be prepared to contribute some effort would be
extremely welcome :-)


BTW - and this is not meant negatively - a common interface would be very
useful for MSXML. I downloaded it yesterday and I'm still trying to find
out how to get off the ground with it.  I'll get there, but - like all
software at this level - there is quite a lot of exploration to be done.


[... redundant xml-dev signatures deleted ...]

Peter Murray-Rust, Director Virtual School of Molecular Sciences, domestic
net connection
VSMS, Virtual Hyperglossary

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as:
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list