Re: [xsl] distinguish whether variable holds string or node set

[TW <zupftom@xxxxxxxxxxxxxx>]

Thanks for sharing your expert insight.  I must say, I'm very
impressed by your activity and omnipresence all around XSL.  Thanks
for the great work you're doing.


2010/6/14 Michael Kay <mike@xxxxxxxxxxxx>:
>
>> Well, then I'd rather stick to the "adopted standard".  (Why is XSLT 2
>> so badly supported?)
>>
>
>
> Poor choice of words there. No-one supports XSLT 2.0 badly; they either
> support it well, or not at all.
>
> If you mean, why have so few people implemented XSLT 2.0, then I think you
> have to ask the question "Why has company|person X not implemented XSLT
2.0"
> for a number of values of X, and then see if there is any commonality in
the
> answers.
>
> For the browser vendors, I think the answer is that none of them is
> particularly interested in making the first move. You don't get competitive
> advantage in the browser space by being the first to support new standards.
> There's a built-in inertia here that's a great drag on the industry.
>
> For Microsoft, the answer is that they lost interest in XSLT, largely
> because they became over-enthusiastic about their own proprietary
languages,
> and because of various power shifts between the various internal groups
> interested in XML technology. But that's me trying to read the tea-leaves:
> only Microsoft can give you the inside story, and I doubt they ever will.
>
> For a lot of the people who produced open-source XSLT 1.0 engines, I think
> you'll find that they did the 1.0 version for fun, or in a vague hope that
> they might find some unspecified revenue stream as a spin-off, perhaps
> having greatly underestimated the size of the task; and having done it,
with
> a lot of hard work, a lot of fun, a lot of lost week-ends, and no revenue,
> they weren't inclined to start all over again.
>
> But five XSLT 2.0 engines have been shipped, or five-and-a-half if you
count
> Oracle, and I know of four more that are under development; if we get to
the
> point where there are nine separately-developed implementations then we
will
> have achieved far more than most programming language standards. And
> although some users are still sticking to XSLT 1.0 because that's the only
> thing supported in their favourite environment, a great many more have made
> the shift and are very pleased with the productivity benefits that it
> brings.
>
> Michael Kay
> Saxonica