Re: [xsl] Haskell programmer's rant about xslt

[Alan Painter <alan.painter@xxxxxxxxx>]

One can debate the merits of such-and-such an approach to
transformations and the debate should be lively.

But verbiage such as "your bastard child of a language" has no utility
whatsoever and certainly no place in polite discourse.

I would suggest giving this guy's opinions no further attention.

cheers

-aljo


On Tue, Apr 17, 2012 at 6:19 PM, Michael Kay <mike@xxxxxxxxxxxx> wrote:
>>His point is that XSLT makes it especially difficult to debug (bad)
>> programs written by others. How would one best solve the problem he
>> describes?
>
> Yes, debugging stylesheets that have evolved over a period of time and
> contain lots of overlays can be difficult. The price of making it easy for
> stylesheets to evolve is that they can evolve badly.
>
> I generally use the Saxon -T option to resolve such issues. The output can
> be horribly large when the input file is large, unfortunately, but it
> usually reveals fairly quickly which rules are being invoked for which
> nodes. Perhaps a variant that only outputs the template rule matches would
> be useful (it's a very simple tweak).
>
> Michael Kay
> Saxonica