Howdy,
Just a small side note. As slow as python might look like, the biggest bottleneck might be transferring huge AT structures using dbus (as dbus is hell slow) or applications flooding events. And those issues needs to get fixed on other places in any way.About parallel computing - parsing AT data in an parallel manner does not work in any way. So yea, python has limitations, but none are hit by an screenreader.
Cheers chrys
Am 23.03.23 um 22:55 schrieb Linux for blind general discussion:
Orca is indoubtebly an amazing project, with a lot of good work behind it.
Though, as great as the final product is, there are certain limitations.
The screenreader is written in a language that can take eternity just to
finish a loop, and is not capable of true concurrency. I'm not
criticizing Python here, I use it myself and love the language, I'm just
saying using it for performance critical tasks has its caveats.
Rust is from ground up designed for high performance and parallel
computing. That's why I'm personally very excited to see these strengths
used for accessibility, they could yield very interesting results.
Odilia also handles input devices differently, and aims for supporting
touch screens in the future. That's also a very interesting feature,
especially now, as Linux mobile phones are on a rise.
We'll see how the project goes, and what results we get. But it's
certainly a very interesting and impressive work, that fully deserves
attention.
Best regards
Rastislav
Dňa 23. 3. 2023 o 12:26 Linux for blind general discussion napísal(a):
On 23/3/23 03:18, Linux for blind general discussion wrote:
Interesting, could you be more specific on this? From those I know, NVDA
is Python & C/C++, Jaws is written entirely in C++ if I remember correctly.
VoiceOver, just like the two is older than Rust itself, so I guess it's
written in the Apple's middle-level languages.
While Android has a Java centered accessibility system.
Indeed, I think this might be the first implementation of a screen
reader in Rust.
I'm not immediately persuaded that writing another screen reader for the
Linux desktop is a good idea over-all. I think Orca is expertly
maintained, and that most of the problems of graphical user interface
accessibility under Linux arise outside of Orca itself.
On the other side, implementing from the ground up with new technologies
might yield positive results in the end, if the project is sustainable,
and people can of course contribute to the community however they wish.
That includes working on a new screen reader project if it's what they
want to do.
The last time there was a second screen reader for the Linux desktop - I
think it was called LSR - it was maintained by IBM. Then IBM decided to
discontinue their work on the project, and no one took it over. So we've
been in the "multiple screen reader" situation before.
_______________________________________________
Blinux-list mailing list
Blinux-list@xxxxxxxxxx
https://listman.redhat.com/mailman/listinfo/blinux-list
_______________________________________________
Blinux-list mailing list
Blinux-list@xxxxxxxxxx
https://listman.redhat.com/mailman/listinfo/blinux-list
_______________________________________________
Blinux-list mailing list
Blinux-list@xxxxxxxxxx
https://listman.redhat.com/mailman/listinfo/blinux-list