Well, we have already seen this and have been
surprised with Sablotron poor results. A closer
look shows that Sablotron is pretty fast running
the most of test, but really very slow running a
few others (patterns itself takes 68% and bredth
16% of overall time).
That's still a pain - Sablotron performs well with
commonly used transforms, but benchmarks come with
a wide range of all possible stylesheets and
uncover Sablots weak points.
On the other hand, this kind of testing is very
helpful and allows to improve Sablotron a fast
way. It would be interesting to see the same test
results for upcoming versions.
Petr
"jan.larsson" wrote:
>
> XSLTMark now includes Sablotron:
>
> http://www.datapower.com/XSLTMark/res_2000_11_20.html
>
> > -----Original Message-----
> > From: Adam Doppelt [mailto:[EMAIL PROTECTED]]
> > Sent: Tuesday, November 21, 2000 8:22 PM
> > To: Sablotron Mailing List
> > Subject: [Sab] [SAB] XSLBench - have you seen these?
> >
> >
> > Kevin Jones ran several XSLT processors through some benchmarks recently,
> > and I don't remember seeing the results on the list. You can find his page
> > here:
> >
> > http://www.tfi-technology.com/xml/xslbench.html
> >
> > Sablotron performed solidly in the middle of the pack.
> >
> > Adam
> >
> >
--
Petr Cimprich
Ginger Alliance Ltd.
www.gingerall.com
