[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: XPath 2.0 - how much of XQuery should it include?

• To: "Michael Kay" <michael.h.kay@n...>
• Subject: Re: XPath 2.0 - how much of XQuery should it include?
• From: Jeni Tennison <jeni@j...>
• Date: Fri, 10 May 2002 17:44:14 +0100
• Cc: xml-dev@l...
• In-reply-to: <000701c1f834$2f6b6600$6501a8c0@pcukmka>
• Organization: Jeni Tennison Consulting Ltd
• References: <000701c1f834$2f6b6600$6501a8c0@pcukmka>
• Reply-to: Jeni Tennison <jeni@j...>

Hi Mike,

> There is no right answer, but I'm personally happy with having the
> line where we've put it. It does mean that some XPath expressions
> sit rather uncomfortably inside XML attributes, and we've looked at
> proposals that address that, but I don't feel that balancing the
> length of the XPath attributes against the number of XSLT elements
> is the right way to make the judgement. You can always break up long
> expressions using variables or function calls.

Of course there's no right answer -- this is a matter of design, and
people have different opinions on design. Usually that arises from a
difference in the criteria on which they assess a design, which
usually arises from a difference in the goals that they have for
whatever they're designing.

My criterion is usability, and my goal is to learn and to teach
(preferably in that order). I don't think that the current design is
very usable. That's not just based on the length of attribute values
(though I don't think that's an unreasonable heuristic), it's based on
my estimate of the effort involved in learning two ways to do
everything and what the advantages and disadvantages are of those
choices. My opinion is also based on the problems I anticipate in the
maintainability of stylesheets -- I think that people will avoid
creating user-defined functions to do relatively simple operations,
and that given the lack of variable-binding expressions in XPath,
we'll therefore get a lot of code with a lot of repetition in it,
which is hard to maintain.

From previous discussions, I believe that your goals are eventual
convergence with XQuery and the ability to improve the performance of
Saxon. From that, your criteria are similarity with XQuery and
optimisability of XPath and possibly XSLT. Is that a fair summary?

Do you think that I'm wrong when I characterise the current XPath 2.0
design as hard for users to learn and use? Do you think that the
changes I've suggested would make it less usable? Or do you just think
that usability shouldn't be the main criteria on which the design is
assessed?

Cheers,

Jeni

---
Jeni Tennison
http://www.jenitennison.com/

• References:
  • RE: XPath 2.0 - how much of XQuery should it include?
    • From: "Michael Kay" <michael.h.kay@n...>

• Prev by Date: RE: Tim Bray on "Which Technologies Matter?"
• Next by Date: Re: Tim Bray on "Which Technologies Matter?"
• Previous by thread: RE: XPath 2.0 - how much of XQuery should it include?
• Next by thread: Re: XPath 2.0 - how much of XQuery should it include?
• Index(es):
  • Date
  • Thread