Re: [xsl] Using or ignoring Types in XSLT 2.0 / XPath 2.0

----- Original Message ----- 
From: "Kay, Michael" <Michael.Kay@softwareag.com>

Michael,

Thanks for your thoughtful reply.  I recognize there has been rather more heat
than light generated around this issue.  That is both the reason I wrote and the
reason I hesitated so long.

> > 1) A range of XPath 2 functions display indeterminate
> > behavior in the absence of PSVI type annotation.
>
> I don't know what you mean by this. Could you please elaborate? I think that
> all the behavior is well-defined, except in a few quite particular cases
> where we have chosen to make it implementation-defined (e.g. collating
> sequences).

In the current toolset, if I ask for substring($date, 4, 1) I get it whether it
is junk or not.  The stylesheet doesn't fail.  I appreciate why many consider
this a problem.  I feel that it is a significant factor in XSL's success.  My
understanding is that if I ask for fn:get-year-from-dateTime($date), in the
absence of type annotation, $date will have a type xdt:untypedAtomic and asking
for a function whose type signature demands a xs:dateTime will cause the
processor to coerce $date and fail if it can't.

> I suspect that the biggest part of validation cost is the cost of compiling
> the schema into a form that constitutes an executable validator. It should
> be possible to do this once, in the same way as stylesheets are compiled
> once and executed repeatedly. If the compiled schema and the compiled
> stylesheet are combined, there is considerable potential for efficiencies. I
> think it would be a mistake to assume that performance impact will be a
> show-stopper, or even that the net effect of introducing schemas will be
> negative. It depends very much on the application.

Thanks for your insight.  I suspected that I was not seeing the whole picture
and you have confirmed that for me.

> You haven't really explained enough about your application architecture for
> me to comment. Where does XSLT fit in?

The presentation tier is a two-stage transformation generating web views or
printed reports.  XSLT is also used to manage data conversion.

> Converting the same data between three different representations -
> relations, objects, and XML - looks like the inefficient part of your
> process to me. But I can't really judge.

Performance of our current application is quite solid and scales very well.  I'd
like not to introduce validation in the production system.

> It's worth pointing out that the type information in the data model used as
> input to an XSLT transformation does not necessarily have to come from
> schema validation. Implementors might well provide utilities that create a

So I have learned from Mr. Rys.  If I could persuade you to comment on the
clarification I made in reply to his message, I would be obliged.

> Absolutely not. What gave you that impression?

The mistaken notion that type annotation cannot be applied in the absence of
schema validation.

But I still hang here on the idea that the behavior of any given function,
operator or instruction is dependent upon whether or not a type annotation is
available for the datapoint.  Doesn't this draw a much sharper distinction than
currently exists between a valid instance and one merely well-formed?  Doesn't
this prevent me from simply throwing a switch for validation?  My stylesheets
must know whether to expect annotation or not.

> I will refrain from speculation on this. Partly because I have no idea what
> you mean by "userspace".

The place where people who are not professional developers go to use a computer
for anything other than email and surfing.  I am a linguist by training, a poet
by avocation and a self-taught developer.  I fondly recall the stupid ease with
which I began to use a computer for real work when XML arrived.  I owe my
current career to markup.  Many pay lip service to the idea that XML gives
control of data back to users; few appreciate how enabling and ennobling that
experience can be.  I would hate for that potential to be lost.

> If you could make concrete suggestions as to how it could go further, that

I cannot.  And in light of the potential to annotate datatypes without
validation, there is little that would further.

> Clearly the languages are trying to meet a range of different needs. If you
> think the design can be improved, without reducing the range of requirements
> we are trying to satisfy, then we will welcome concrete suggestions.

I do not.  I guess I do feel that the range of requirements the WG is trying to
satisfy is part of the problem.  I recognize that train has left the station.

> Please - if you think the behavior is "unsafe" then you must tell us how.

I am not thinking of type safety per se.  I am thinking of the lexical failure
of an automatic cast performed by a *Basic XSLT Processor* as described in XSLT
2.0. Perhaps this is only an issue over the XSD built-in types and more of a
molehill than it seems at first blush.

> You
> obviously have some particular problem in mind here, please tell us what it
> is.

That my stylesheets must be aware of whether or not they are dealing with
annotated data.


With many thanks for your cogent response,

Mike