Re: EXSLT 1.0 Proposed List

Pawson, David wrote:
>     > What motivation can there be to implement something as a 
>     1.0 extension 
>     > when it is available as 2.0 base functionality?
>     >   
>     
>     As you can see in the legend, there is a lot in EXSLT *now* 
>     that is also available in XSLT 2.0.  Why did people 
>     implement those extensions?
>
> Because of the timing Uche? 
> XSLT 2.0 is nearing rec status.
>   

What does the timing of XSLT2 have to do with EXSLT?  I for one will
almost certainly never implement XSLT 2, and I know of a lot of other
implementors who feel the same way.  In fact, right now it's easier to
find XSLT 1 implementors who don't plan to move on to XSLT 2 than it is
those who do.


>     Of course the answer is that implementing a handful of 
>     useful functions/elements that happen to be similar to ones 
>     available in XSLT 2.0 is *much* simpler than implementing 
>     XPath 2.0 and XSLT 2.0 itself. 
> Agreed. 
>     A hell of a lot easier.  Several implementors, including 
>     me, have expressed the preference to move the most useful 
>     and sensible extensions into EXSLT so that people don't 
>     have to deal with the XPath/XSLT 2.0 mess if they don't want to.
>
>
> *slowly* I'm becoming a convert to 2.0. Yes, there is a mess in there,
> but at the same time there are some very useful bits too.
>   

Well, that's fine for you, buit for those of us not willing to bother
with the mess, EXSLT should provide some of the more useful stuff in
XSLT 1.0.

XSLT 2 could have done a lot of useful things incrementally (the
node-set/RTF issue, grouping, dates...) without making the enormous leap
in complexity represented by XSLT 2 (mostly by virtue of XPath 2).  I
think EXSLT can and should offer that incremental step.


> Useful and sensible as the criteria? I guess that makes sense,
> even if its 4suite on the route to an XSLT 2.0 implementation
> it will allow for a cleaner migration?
>
>   

I don't know who is planning to implement XSLT 2 in 4Suite.  I'm not
sure I want anyone to.  4Suite's big enough.

> Do you think other implementations would show similar levels of interest?
> Daniel, Apache, Oleg?
>
> Spec'd without implementations could appear silly.
>   

EXSLT originally had no implementations until I whipped up 4EXSLT.

This is a spec built on practical needs and working with existing
material.  As such I expect it will invite implementations.


-- 
Uche Ogbuji                               Fourthought, Inc.
http://uche.ogbuji.net                    http://fourthought.com
http://copia.ogbuji.net                   http://4Suite.org
Articles: http://uche.ogbuji.net/tech/publications/