Re: [orca-list] Firefox caching

[Jason White <jason jasonjgw net>]

Michael Whapples <mwhapples aim com> wrote:
 
> Regarding performance of orca and firefox, I wonder if some of the
> problem really comes from some of the work arounds and checking has
> to do because of firefox issues. I am partly basing this idea on how
> other screen readers which don't use virtual buffers perform (I
> believe voiceover and the firevox plugin both don't use virtual
> buffers) and I guess another clue to potential performance with
> direct access can be judged by how orca performs with webkit (I
> think I have seen messages which suggest webkit behaves better and
> may need less checking by orca).

Virtual buffers were discussed extensively once before, and the conclusion was
that they were not desirable for both user interface and implementation
reasons. I would respectfully suggest to anyone who wants to reopen that
discussion that they read the mailing list's archives and consider very
carefully whether they have anything new to add to what has been said already
on the subject.

As to performance, the Gecko scripts for Orca amounted to more than 700k of
Python code when last I looked, which gives a very broad indication of how
much work Orca has to do in order to make Mozilla applications accessible. I
am not surprised that all of this complexity comes with some associated
performance costs. Indeed, versions of Firefox prior to 4.0 have been
criticized for lacking in performance compared with Webkit-based alternatives
such as Chromium. I don't know to what extent these various contributors are
responsible for the performance issues that some people have been
experiencing.