Daniel Veillard wrote:

>   It's not too late to *at least* emit a warning when the
> conversion occurs. Coming from a Working Group member this reflects
> badly, no ? Why should others take the pain of fixing their bugs
> when you don't.
A little strong, I think.  David already pointed out that the 1.1
draft is worded such that to emit a warning would in fact be a
non-conformance.  (I think this is a mistake in the spec, too FWIW).
The Saxon documentation states very clearly to what extent it conforms
to the 1.1 draft.  If you want precise XSLT 1.0 conformance then you
should be using an earlier version.  You will miss out on some bug
fixes in doing so, but these are few and clearly documented.  (Is the
same true of Microsoft?)

>  IIRC Microsoft rolled back this feature, if they
> can do it to their customer base in name of the compatibility,
I think the situation is a bit different.  Miscrosoft seems to have
had this feature either through simple omisson or because they thought
it a good idea - I am not aware they were trying to implement 1.1.

> I don't see how you can justify your decision.
Hindsight is a wonderful thing.  Mike has already said he would have
done things differently had he known what was going to happen.  As a
user of Saxon I would rather he spends his time making sure XSLT 2.0
is good.  And if he has spare time after that, can we have a 2.0 Saxon
please?

Regards,
Trevor Nash
--
Traditional training & distance learning,
Consultancy by email

Melvaig Software Engineering Limited
voice:     +44 (0) 1445 771 271 
email:     tcn@[...].uk

 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list