Alexander R. Pruss wrote:
> Well, anybody who wants to benchmark whether performance
> was degraded can do so.  My own feeling is that the amount of
> degredation is going to be miniscule. 

OK; I thought you were saying that it was noticeable in your
tests even without benchmarking.

> Moreover, access to YOffset type data doesn't happen once 
> per character.   It happens once per sequence of characters
> in the same style on the same line.

(and other messages said that bookmarks are a YOffset.)

Am I understanding this correctly?  In order to jump to 
#section5 which is 87% of the way through a document,
you first have to pretend you're rendering every single
line above it, to until the YOffset finally matches?

Does this apply to go-to-bottom as well?  If so, I'm pretty
sure that *my* page-opening time is dominated by 
uncompression, because jump-to-bottom isn't all that 
slow.  (Or does half-page down also recalc from the top?)

(and in another message)
> duplication of effort between WriteLine() and GetLineMetrics().  Under
> POSE/NR70 it speeds up my benchmark page from 26 sec to 10 sec.  On my
> NX70, it sped up the same page from 4.8 to 3.1 seconds.  

Do you have any idea why it would help so much more on POSE?  (60% vs 30%)

-jJ
_______________________________________________
plucker-dev mailing list
[EMAIL PROTECTED]
http://lists.rubberchicken.org/mailman/listinfo/plucker-dev

Reply via email to