Fixed development cycle.

[vaut]

Our development pattern is not very successful.
Releases once a year without intermediate versions. The last release was almost 9 months ago. Players have been waiting for ready-made improvements for more than half a year.

At the same time, the stability of the release does not differ much from the usual master build. Both are more than good.

Such a long cycle is bad because:

1. it is dangerous to change the balance. An unsuccessful change will be fixed for a long time
2. improvements are received by players with a long delay
3. quality deteriorates due to insufficient testing. Not found issues overlap.
4. the release is constantly being postponed in order to add another almost finished feature now, and not in a year.

I want to suggest moving to a shorter loop:
release every 3 months. The reason to postpone can only be open errors with regression. Several interim releases for tests in MT. Updating the version several times a day is problematic ...
And it's very problematic to install not the latest version.

Personally, I've been waiting for several months:
research progress share, key for external storage of player profile, headless version, multi-port, penetration refactor, splash refactor, new hotkeys.

[past-due]

You'll be pleased, then, that 4.0.0-beta1 was published today. 🥳 And I agree that more frequent, smaller releases would be better.