Re: [xsl] Math "functions" for XSLT 1.0

[Dimitre Novatchev <dnovatchev@xxxxxxxxx>]

On Wed, Mar 3, 2010 at 10:56 PM, COUTHURES Alain
<alain.couthures@xxxxxxxxxxxxx> wrote:
> Dimitre,
>>
>> To substantiate this statement, I ran a quick and dirty test for sin()
>> on 201 values in the interval [-1, +1], with x(i) = x(i-1) +0.01.
>>
>>
>> Here are the results (using Msxml4):
>>
>> CORDIC (Couthures): 93.810ms
>>
>> Taylor series(FXSL) : B  9.664 ms
>>
>>
>
> Thank you for these measures! Did you also compare results by the way ?

Yep, the results from both transformations are the same in 11 to 16
digits after the decimal point.


>
> About CORDIC, the quote in Wikipedia you gave is not in the French version
> of it but should be added. Today, the problem is not "to read more" but "to
> find what to read" and "not to find what not to read".

I'd put it in another way: Always verify any statement you read --
don't take it for granted.


>
> So, I will keep in mind to look at this later if necessary...


..................................


> I don't pretend to be smarter, the truth is I don't hesitate to do
> myself what I can't find, even if others say it's impossible or, at least,
> difficult.

This is exactly describing me, too, but after carrying out a lot of
such "impossible" tasks, even this becomes boring. After implementing
an LR parsing system in XSLT and creating with it two specific parsers
-- for JSON and for XPath 2.0, I cannot find anything that is at least
as interesting to implement as these. :(

>
> I love XSLT and, even if it's not always easy, I'm convinced that XSLT 1.0
> (the only one implemented by Microsoft, PHP and browsers,... if only there
> was a count of deployed XSLT engines, XSLT 1.0 would be the (almost unused)
> winner) is good for any kind of calculations too.


As for using XSLT processors from a particular vendor, I always use
the one most suitable to the task -- sometimes it is Saxon, sometimes
MSXML, a lot of the time it is .NET XslCompiledTransform. A very
important factor is also the XSLT IDE to use. For XSLT 2.0 Oxygen +
Saxon is nice, for XSLT 1.0 I'm still using the 10 years old
XSelerator.

These IDEs allow many user-specified XSLT processors to be invoked
most easily, so it is not a problem to use a specific one in every
case.


Cheers,

Dimitre.



>
> -Alain
>
>



--
Cheers,
Dimitre Novatchev
---------------------------------------
Truly great madness cannot be achieved without significant intelligence.
---------------------------------------
To invent, you need a good imagination and a pile of junk
-------------------------------------
Never fight an inanimate object
-------------------------------------
You've achieved success in your field when you don't know whether what
you're doing is work or play
-------------------------------------
I enjoy the massacre of ads. This sentence will slaughter ads without
a messy bloodbath.