search

nvaccess / nvda

NVDA, the free and open source Screen Reader for Microsoft Windows
https://www.nvaccess.org/
Other
2.11k stars 637 forks source link

Rapid integration of multiple infrastructure changes may cause user-facing instability and decrease confidence #9539

Closed ctoth closed 5 years ago

ctoth commented 5 years ago

I am raising an issue which may have been already discussed, feel free to discard.

Looking at recent development it looks like both Python 3 and the large work on speech refactor will be landing pretty close together, potentially on the same release. I absolutely understand wanting to perform all possible breaking changes concurrently, so staggering these temporally is not optimal.

With that considered, I propose both strong messaging around the eventual release, mentioning instability, and a much longer than usual Alpha period.

Really looking forward to this modernization work, and thank you for all you do.

josephsl commented 5 years ago

Hi, I second this motion. I will do the same for the add-ons community once Project Threshold officially kicks in. Thanks.

From: Christopher Toth notifications@github.com Sent: Saturday, May 4, 2019 3:42 AM To: nvaccess/nvda nvda@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: [nvaccess/nvda] Rapid integration of multiple infrastructure changes may cause user-facing instability and decrease confidence (#9539)

I am raising an issue which may have been already discussed, feel free to discard.

Looking at recent development it looks like both Python 3 and the large work on speech refactor will be landing pretty close together, potentially on the same release. I absolutely understand wanting to perform all possible breaking changes concurrently, so staggering these temporally is not optimal.

With that considered, I propose both strong messaging around the eventual release, mentioning instability, and a much longer than usual Alpha period.

Really looking forward to this modernization work, and thank you for all you do.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/nvaccess/nvda/issues/9539 , or mute the thread https://github.com/notifications/unsubscribe-auth/AB4AXEFGU2NYAKMYMQLN6W3PTVSAXANCNFSM4HKYTV3Q .

LeonarddeR commented 5 years ago

I also agree that we need to test carefully. There are some things pretty hard to test, for example, braille display drivers that are in core but not used very widely, appModules for some pretty old apps. Some appModules and drivers might even perish during the process.

Note however that project threshold will be hosted in a stand alone branch. This means that the Python 2 based code base of NVDA can still be developed further. I think that it is unlikely that 2018.3 will be based on threshold, and I think that even a 2018.4 based on threshold would be a race against the clock. That, unless the release schedule is slightly modified/paused for this reason.

Brian1Gaff commented 5 years ago

Well as the resident ludite around here, I'm happy to test Outlook Express and maybe some older versions of Office etc, if I'm poked accordingly.

Brian

bglists@blueyonder.co.uk Sent via blueyonder. Please address personal E-mail to:- briang1@blueyonder.co.uk, putting 'Brian Gaff' in the display name field. ----- Original Message ----- From: "Leonard de Ruijter" notifications@github.com To: "nvaccess/nvda" nvda@noreply.github.com Cc: "Subscribed" subscribed@noreply.github.com Sent: Saturday, May 04, 2019 11:56 AM Subject: Re: [nvaccess/nvda] Rapid integration of multiple infrastructure changes may cause user-facing instability and decrease confidence (#9539)

I also agree that we need to test carefully. There are some things pretty hard to test, for example, braille display drivers that are in core but not used very widely, appModules for some pretty old apps. Some appModules and drivers might even perish during the process.

Note however that project threshold will be hosted in a stand alone branch. This means that the Python 2 based code base of NVDA can still be developed further. I think that it is unlikely that 2018.3 will be based on threshold, and I think that even a 2018.4 based on threshold would be a race against the clock. That, unless the release schedule is slightly modified/paused for this reason.

-- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/nvaccess/nvda/issues/9539#issuecomment-489316575

Adriani90 commented 5 years ago

I agree we should stay with two parallel master branches, python 2 and python 3. If someone wants to change to python 3, they can do that. If not, they can still stay to python 2 for a while. But in any case we must inform the community widely that testing under python 3 is very importnat because one day the python 2 branch will not exist anymore. Then, also people using older braille displays or older software will be incentived to test their things under python 3. This work needs also a big contribution from translators worldwide. Because in many less developped countries people are still using old systems with old pieces of software or old braille displays. Even in Germany some people are still using 20 years old braille displays. So we have enough people to get feedback from. Don't forget to consider contacting translators from every country to promote this.

zstanecic commented 5 years ago

I as a croatian translator can help with this...

From: Adriani90 notifications@github.com Sent: Sunday, May 5, 2019 12:57 PM To: nvaccess/nvda nvda@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: Re: [nvaccess/nvda] Rapid integration of multiple infrastructure changes may cause user-facing instability and decrease confidence (#9539)

I agree we should stay with two parallel master branches, python 2 and python 3. If someone wants to change to python 3, they can do that. If not, they can still stay to python 2 for a while. But in any case we must inform the community widely that testing under python 3 is very importnat because one day the python 2 branch will not exist anymore. Then, also people using older braille displays or older software will be incentived to test their things under python 3. This work needs also a big contribution from translators worldwide. Because in many less developped countries people are still using old systems with old pieces of software or old braille displays. Even in Germany some people are still using 20 years old braille displays. So we have enough people to get feedback from. Don't forget to consider contacting translators from every country to promote this.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/nvaccess/nvda/issues/9539#issuecomment-489414996 , or mute the thread https://github.com/notifications/unsubscribe-auth/ACVCDE5XSQTZOZBE7UXFZPTPT24QRANCNFSM4HKYTV3Q . https://github.com/notifications/beacon/ACVCDEY2WPT7WKT24BPMFRLPT24QRA5CNFSM4HKYTV32YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODUV6CVA.gif

LeonarddeR commented 5 years ago

I have the feeling that this issue is pretty hard to monitor, as it appeals on our attitudes rather than on a particular change in code. Therefore, I"m closing this issue for now. If you have the feeling that messaging from either NV Access or other NVDA Council isn't what you expect, feel free to reach out again.

josephsl commented 5 years ago

CC @MichaelDCurran, @feerrenrut and others.