Re: [GOK] gok and gnome 3.0

[Steve Lee <steve fullmeasure co uk>]

2009/9/23 Ben Konrath <ben konrath utoronto ca>:
> I see two options here: we can either move gok over to Python and use
> pyatspi2 or we can keep gok in c and port it to cspi2. For the second
> option, somebody, I guess that's me, will have to work on cspi2.
>
> My feeling is that moving GOK over to Python is a better long term
> strategy - this brings us inline with other GNOME accessibility apps and
> it will be easier to experiment when it comes time to add new features
> after GNOME 3.0. This, however, is only my opinion. I would like to get
> comments from other people and hopefully come to a consensus on this
> issue before I proceed. Gerd, I'm specifically interested in hearing
> your opinion on this.

+1 to Python

1) pyatspi is mature, stable, is very easy to use and well proven
though its use in Orca, Accesserciser and others.
2) Accerciser, Orca etc provide code examples that are up-to-date.
3) Orca encapsulates a large body of AT development experience
4) There is growing expertise in Python in the accessibility community
and wider GNOME community.
5) Having developed extensively with C/C++ and Python I find Python
offers many advantages for developing AT.

Python and C play well together so it might be possible to package
some existing GOK code as modules, depending on current architecture.
The reverse is also possible, embedding a Python interpreter in C, but
I suspect that is less likely to be useful..

The problems I found when working on Jambu were all due to the atspi
spec leaving much to the interpretation of individual programs and not
due to pyatspi (especially events and dynamic behaviour).

I'm also not sure of the status of cspi - I thought it was depreciated
(or was that before cspi2)?


Steve Lee