[xmlroff] Re: xmlroff-list Digest, Vol 14, Issue 4

[Chris Bowditch]

Tony Graham wrote:
> On Wed, Jul 09 2008 10:46:33 +0100, phill at theactivitypeople.co.uk wrote:
> ...
> 
>>Are there any performance/memory/feature comparisons available for
>>xmlroff and fop?
> 
> 
> No.
> 
> xmlroff doesn't impinge upon the consciousnesses of the FOP developers
> much, if at all: there's more mentions of FOP on this list than there
> are mentions of xmlroff on fop-dev.  I don't recall many mentions of
> xmlroff at all other than a flurry back in 2002 when xmlroff was first
> announced.

I concur that none of the FOP developers or users have done anything 
like a performance test between xmlroff and FOP. If they did they stayed 
quiet about it. I would be interested to see the results of such a 
comparison :)

> 
> FOP's conformance is listed on the FOP website.  xmlroff's conformance
> is listed in the xmlroff manual in the distribution.
> 
> FOP and xmlroff have different approaches: FOP tries to flush each page
> as quickly as possible; xmlroff is still working towards being able to

That used to be true in FOP 0.20.5 but that's no longer true of the 0.9x 
versions where a best fit algorithm is used over the life of the page 
sequence.

> work on the area tree to get the best result.  (The eventual plan has
> always been to have loadable modules to do the tinkering with the area
> tree so you can determine your trade-off between tinkering and speed.)

That is a commendable objective. It would be great if the user could 
choose between first-fit/best-fit/total-fit by the mere adjustment of a 
config file :)

Thanks,

Chris