Re: use cases for d-o-e

David Carlisle <davidc@xxxxxxxxx> wrote:
> Sniff.
Sorry!

> Quick quiz, what do each of these MathML expressions represent:
> 1)  <mo>&Integral;</mo>
> 2)  <mo>&#8706;</mo>

That's a question only relevant for humans. Software doesn't benefit
from using entity references.

I suppose you use XSL to manipulate documents which are intended to
be examined and potentially changed by humans. I know you are converting
some non-XML sources. I'm not quite sure whether there are further
reasons: are you enhancing an editing system, using XSLT for automation
or such, or do you use XLST do keep the documents compliant to an
evolving data schema, or something else?

As for your task of generating XML documents with entity references:
i could have easily proposed an element <xsl:entity-ref name="..."/>.
However, i still think XSLT has never been intended and is not designed
to produce documents which will be read and changed by humans on a
regular basis. I don't like the idea to apply XSLT to *every* task where
the template mechanism might be useful. Especially i'm not fond of the
idea of making XSLT a swiss army knife tool. Since i have no authority
on this list, i'll invoke Mr.Kay's one who recently gave the advice:
   "- remember that XSLT is not the only tool in your kitbag"
In the case of the quiz you gave above, i mean for the task of producing
meaningful names for characters, you *can* use a stream editor for
postprocessing. But ideally, reading and changing XML files by hand
should be rare tasks and reserved to a few specialists, ordinary
users should use comfortable tools giving WYSIWYG display and such.

> I suspect there would be howls of pain from users if the feature went,
> with possibly fatal effects on xslt2 deployment.

I didn't meant to drop it immediately, without warning. Even i would
consider this unfair :-). But it should be possible to threaten to remove
it in the V3 revision, it surely will take another 2-3 years to get
this out anyway. This should be enough time that everyone can adapt their
writing style and existing style sheets while doing normal maintenance.
Mandate printing big fat warnings on encountering d-o-e meanwhile.

> You also missed out the other main use of d-o-e: outputting start and end
> tags in different templates thus avoiding teh entire tree processing
> paradigm of XSLT (This is rather common amongst beginners, I expect).
Easier grouping should remedy this. I don't find it worth mentioning,
lest some people believe they've invented something clever...

Regards
J.Pietschmann

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