Actually, far from useless. December 6, 2021 is a credible date because:
- There is a forum (hosted on Groups.IO) where NVDA translators hang out, and members were asked to finish 2021.3 translation by end of November.
- After translators send in their final data, NV Access publishes release candidates, in this case I expect 2021.3 RC to be released anytime this week.
- At least a week after the first RC is released, the stable version is released. If translations schedule holds, next week is when I expect NVDA 2021.3 stable version to come out. Things can change if folks discover a really serious problem that must be fixed right away.
I said "credible date" precisely because of point 3 above: things can change. A rule of thumb is that:
- A given NVDA release will go through beta/RC phase for at least three weeks: two weeks while translators put finishing touches (translatable string freeze), then an RC week. NV Access can choose to declare string freeze (two weeks) right as they release beta 1 but typically developers have waited several days after releasing beta 1 to tell translators that they have two weeks to finalize translation data.
- We can guess as to when the next stable version of NVDA might be released, which takes place within a week after end of translatable string freeze. Folks can "guess" this date not only because NV Access has documented this in many places (including in various In-Process blog posts), but also since things can change without notice.
- For backwards incompatible releases, beta phase is longer - not only developers are finalizing the release, but beta is used by add-on authors to make their add-ons compatible. These days more add-on authors have realized that they must prepare their add-ons a bit early, going so far as willing to think about testing their add-ons against alpha releases (not often but happens; the practice of testing add-ons while using NVDA alpha builds is optional). I expect there will be announcements about NVDA 2022.1 after the new year.