[
Lists Home |
Date Index |
Thread Index
]
On Friday, Feb 20, 2004, at 13:27 Europe/Berlin, G. Ken Holman wrote:
> At 2004-02-19 23:25 -0800, Mark Lentczner wrote:
>> At 2004-02-18 11:11 +0100, Burak Emir wrote:
>>> No matter whether you are more concerned about documents or data,
>>> writing XML syntax to describe XML makes life easier.
>> Not clear to me. I don't find XSLT particularly easy or concise to
>> write lexically.
>> ...
>> It is unrealistic that a program wants to embed a literal that is an
>> entire XML document. It would be more likely that a program wants to
>> output an XML document composed of XML literals and data values from
>> the running program. Therefore, it is highly unlikely that the
>> compiler could do much in the way of ensuring schema conformance.
> ...
>
> For example, where would the following be checked?
>
> <?script for ($i = 1; $i < 10; $i++) { ?>
> <p><?script $i ?>
> <?script } ?>
> </p>
>
> There would appear to be nothing preventing the above from being
> considered incorrect until the result was created and then verified
> for being well formed. The well-formed requirement of XML as utilized
> by XSLT ensures that the above could not be written, therefore, the
> input syntax guarantees that the output syntax is well-formed, which
> means the output need never be checked for being well-formed and it,
> therefore, need not be cached when produced.
>
> This is why disable-output-escaping= in XSLT is so dangerous to use,
> as the stylesheet writer can shoot themselves in the foot. No-one
> should advocate d-o-e= except in the most dire of situations where
> nothing else will do.
>
> This is why *any* out-of-line system adds an entire level of
> complexity and a burden for the stylesheet writer because they cannot
> take advantage of well-formedness to help them ensure the
> well-formedness of the bits of XML they are working with.
>
would those who propose typed "out-of-line" expressions agree with
this? it depends on the out-of-line language. my impression is that the
papers on ml-based xml processing, the document by meijer et al which
was referenced in the initial post, and even x-query argue against that
claim,
> Also, I believe any out-of-line system treats XML as a syntax stream,
> not as a data model of nodes.
in that they posit a mechanism which establishes a type equivalence
between an xml data model and a process, expressed in a syntax other
than xml, which can be demonstrated to produce a correct data model
and/or a well-formed encoded stream.
> ... XSLT is *not* an angle-bracket processor, it is a node processor
> where the nodes usually (but not always) happen to come from and go to
> XML angle brackets. An out-of-line system is going to require users
> to consider syntactic issues rather than let the processors consider
> syntactic issues. XSLT relieves the users of this and lets people
> focus on their information, not on their syntax.
>
> The designers of XSLT make this claim up front and don't try to hide
> it: XSLT was not designed to preserve or manipulate the syntax of a
> document, it was designed to be totally general purpose with the
> information structure of a document: when the document is used by a
> processor downstream, the choice of syntax is irrelevant as long as it
> is correct.
>
> Now if you claim that the processor of an out-of-line system can take
> on the burden of checking the above for well-formedness, then you've
> given that development team a lot of responsibility (and opportunity
> for error) that could have just been handled by an XML processor ...
> just like XSLT does.
with the benefit, that the result "lets people focus on their
information, not their syntax." (the statement, after all, has more
than one sense.)
> An XSLT processor doesn't need to check well-formedness of templates
> because the XML processor inside the XSLT processor checks this and it
> is available off-the-shelf.
whereby, well-formedness is a rather low bar.
>
>> It is this very problem that lead Wheat to use templates. With
>> templates there is clear separation: The XML document looks, feels
>> and is authored as XML. The code is clear, concise and compact. The
>> trick is making sure that the code one writes to connect the two is
>> easy and clear.
>
> You don't describe the syntax of your system here, but I'm assuming it
> isn't like the script above ... I don't quickly see any examples on
> your web site of how you would do the above, but if you aren't using
> XML syntax then I'm not sure how you can guarantee well-formedness.
>
> So, I'm not trying to measure your system or comment on it, I just
> wanted to comment on your observations and derisions about XSLT. It
> seemed to me that your criticisms of XSLT were misplaced when you turn
> around and say "templating is better" ... XSLT *is* templating ...
from the descriptions, and the reference to zope/tal, one distinction
could be that they will be more monolithic with a recursive control
structure, in contrast to element/component and rewrite-based.
...
|
