RE: RE:read-write same url in xslt 2 [was appendig to

>In practice, the result of the transformation could be used to
>replace the original file but only after the transformation is complete.
> I
>don't think Bryan was suggesting that side-effects should be allowed. This
>shouldn't be an error, IMHO.

yeah I was of course thinking that there wouldn't be any side effect because
I figured the processing model would always be that a document is read in
and the result is put out at the very end of the transformation. but I guess
from what Michael was saying was that this is not a required processing
model. In fact I guess when I think about processors like Napa and the use
of saxon:preview that there could be side effects from improper use of such
a method. Also I remember reading, probably in the xslt programmers
reference, theoretical discussion about stylesheet processing optimizations,
whereby parts of a transform could be put out before the whole transform is
done, right? The problem I think is that this is one of the nascent
possibilities in xslt(nascent in that it was not initially considered but
has in fact become somewhat widespread), that one can use it to build quite
decent little document management systems and such. And the thing that
strikes me sort of strange with disallowing something that will work with
some processors, and probably with the greatest number that support multiple
output extensions, is that in such a case one is making the 2.0 version less
powerful.



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