[Scribus] speed problem once more

[Marvin Dickens]

On Thursday 21 April 2005 11:46 pm, Craig Ringer wrote:
> On Thu, 2005-04-21 at 17:04 -0400, Marvin Dickens wrote:
> > Strip the scribus binary and things will speed up some. The
> > trade off is no debugging info in the event of a problem.
>
> That's interesting. My understanding was that stripping binaries didn't
> gain you much, if anything. Some quick research suggests that some
> extremely CPU intensive apps gain as much as 4% when stripped. That's
> about 1/10 of a second in a two-second operation, which I doubt would
> generally even be noticed.

The main focus of my work revolves around firmware. It has been my experience
that in single operations the gain is not noticeable to humans. However, 
operations that are reiterated hundreds or thousands of times to complete a
task benefit because the "task" is seen by the user as a singular event when 
in fact,  "under the hood" it is the sum total of multiple operations. Now, 
I've not really looked at the code that comprises Scribus. But, in an 
application as large as Scribus is, there is a high likelyhood that this  
paradigm applies. But, then again - I've not looked at the code.

Regards

Marvin