W3C home > Mailing lists > Public > public-mobileok-checker@w3.org > October 2007

Re: Optimizing XSLT usage

From: Sean Owen <srowen@google.com>
Date: Mon, 22 Oct 2007 17:58:26 -0400
Message-ID: <e920a71c0710221458s572faf2cxe91d7165bf1a8c78@mail.gmail.com>
To: "Dominique Hazael-Massieux" <dom@w3.org>
Cc: public-mobileok-checker <public-mobileok-checker@w3.org>

I'll run the profiler again tonight to see where the hotspots are at
this point. These are all good ideas. I'd like to see how fast we can
make the current implementation and hope we don't have to resort to
ripping it up for performance reasons.


On 10/22/07, Dominique Hazael-Massieux <dom@w3.org> wrote:
> Hi,
> I just re-read a message from Sean stating that a huge part of the
> processing time of the library is spent in the XSLT processing.
> Reading http://saxon.sourceforge.net/saxon7.7/using-xsl.html#Compiling ,
> I wonder if the following has been explored:
> * pre-compiling the style sheets - since they are static, it may at
> least help reducing the time spent in parsing/interpreting them
> * making sure validation is turned off - it probably doesn't matter
> since the moki document doesn't have a doctype, and I don't think we're
> processing XHTMl files directly, but who knows...
> If it doesn't help, I guess we'll have to consider whether using DOM
> analysis rather than XSLT processing wouldn't be better - ideally,
> having a way to compare performances would help quite a lot.
> Dom
Received on Monday, 22 October 2007 21:58:40 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:21:19 UTC