cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <stef...@apache.org>
Subject Re: Saxon as the default XSLT procesor
Date Fri, 02 Jan 2004 19:02:33 GMT

On 2 Jan 2004, at 09:16, Antonio Gallardo wrote:

> Hi:
>
> 2 recently threads make me think about the topic of the thread. AFAIK
> Cocoon does not try to stick on any product but Jetty performs better 
> than
> Tomcat:
>
> http://marc.theaimsgroup.com/?l=xml-cocoon-users&m=107213647015861&w=2
> another fact is Saxon better than Xalan:
> http://marc.theaimsgroup.com/?l=xml-cocoon-users&m=107243671502545&w=2
>
> JETTY:
> ======
> The license is compatible with Apache since we are already 
> distributing a
> small version of jetty inside Cocoon.
>
> SAXON:
> ======
> Saxon is Mozilla Public License 1.0. It is an OSI aproved license too. 
> But
> I am not sure if it is compatible with ASF license:
> http://www.opensource.org/licenses/mozilla1.0.php
>
> Question:
>
> If the MPL 1.0 is ASF compatible then can we include Saxon as part of
> Cocoon?

Yes. It is.

> AFAIK it will deliver better performance that for sure all the
> users will welcome.

Yep, nobody complains about better performance.

> I understand that Cocoon project is not sticking on any 3rd party 
> product
> from another project, but I think is very important enhance the user
> experience by default to the best products we can use in each area. 
> This
> is why I made a similar question somedays ago about the default XSP
> compiler:
> http://marc.theaimsgroup.com/?l=xml-cocoon-dev&m=107194493701061&w=2
>
> I appreaciate the replies from Vadim and Stefano about my request of
> changing the XSP compiler topic. They helps me understand that eclipse
> compiler was the currently best choice right now.
>
> Please note I am not requesting to remove Xalan nor close the door to
> Xalan. I am just requesting to include Saxon since it performs better 
> and
> seems to have less bugs.

I think Xalan is one example of a community that would die overnight if 
the corporate sponsor pulled the plug. It might be said the same about 
Saxon since it's a one man show, but knowing Xalan a little, I think 
Saxon would be easier as a codebase to patch if Mike Kay decided to 
quit (or got hit by a bus). [it's hard to think at something more 
complex than Xalan to work on]

What I like about this move is the signal is gives: imagine the press 
"Cocoon moves away from its own Xalan and uses Saxon". Xalan has been 
languishing and there is no community, the codebase is incredibly 
complex (yet not the most performant, which irritates me) and translets 
(which were a super-nice idea) are still left behind.

I've done some tests on my machine using the XSLTBench and here is what 
I got:

Saxon 6.5.3 -----> 49.52
XalanJ 2.5.2 ----> 19.89

couldn't test XSLTC

So, +1 to change to Saxon.

--
Stefano.


Mime
View raw message