Rules of Minetest version numbers?

[Wuzzy]

What exactly is the logic behind the Minetest version numbers? Can someone of the core developers please explain?

So far, version numbers of Minetest seem to follow this pattern:

X.Y.Z

Where X, Y and Z can be any whole number equal or greater than 0.

For example: 0.4.16 or 0.5.0

And there seem to be the following rules:

1) Each number gets increased with the following rules:

• X: ???
• Y: milestone / really big update (?)
• Z: big / medium update

2) If Y is increased, Z is reset to 0.

First question: Is my analysis correct so far?
Second question: What is the meaning of X?
Third question: Do you get mad when people in the forums call the version 0.4.16 not “0.4.16”, but “4.16” instead? (I sure do!)

[rubenwardy]

X - always 0
Y - incremented on a breaking change (eg: 0.3->0.4 was modding, 0.4->0.4 is CSM)
Z - incremented on each other release. Currently contains features and bug fixes. Backwards compatible unless there's a mistake.

So yeah, you're right except Z isn't always a big/medium update. 0.4.17 will be a bug fixing release.

It's shit, but there seems to be resistance in changing it so meh

Personally I'd just drop the 0. and then add a patch number, to make 0.4.17=4.16.1 and 0.5=5.0.0

[Byakuren]

Maybe X is always 0 and is meant to remind us that development is a neverending cycle of improvement. The day X is incremented is the day our spirit dies.

[Casimir]

Personally I'd just drop the 0. and then add a patch number, to make 0.4.17=4.16.1 and 0.5=5.0.0

I like it. It would be semantic versioning.

[Sergey]

There is no need for three numbers at all, especially when one of them is always 0 (zero).

Just two numbers is enough — <major_improvements>.<minor_improvements/bugfixes>

Examples: 4.16, 4.17, 5.0, 5.1,…

[v-rob]

The X is for when we have more players than MC. So likely never. :)

[Byakuren]

I like haskell, so I propose we use the versioning scheme that has become standard on hackage: major.major.minor.patch, with four numbers. The advantage of this scheme is that we can keep 0 as the first digit and still have new major, minor, and patch versions.

[Sergey]

I like haskell, so I propose we use the versioning scheme that has become standard on hackage: major.major.minor.patch, with four numbers. The advantage of this scheme is that we can keep 0 as the first digit and still have new major, minor, and patch versions.

What's the point of that zero?

[Byakuren]

Maybe X is always 0 and is meant to remind us that development is a neverending cycle of improvement. The day X is incremented is the day our spirit dies.

[Sergey]

Maybe X is always 0 and is meant to remind us that development is a neverending cycle of improvement. The day X is incremented is the day our spirit dies.

As if software with version number with no zero in front is not developing. Not so funny joke, really.

[Linuxdirk]

Do you get mad when people in the forums call the version 0.4.16 not “0.4.16”, but “4.16” instead? (I sure do!)

To me it makes people look dumb. Plus: it is just plain incorrect because the official version number is 0.4.16 and not 4.16.

[paramat]

The leading zero is probably because MT decided to vaguely follow a conventional practice where you develop, and then later 'release' with 1.0.0, but of course that isn't particularly relevant to free open source open-ended development.
I guess we like to be humble and understated, and not over-confident and contented so we didn't rush towards 1.0.0 even though the game has been usable, 'released' and widely used for a long time.
I see the 0 as a reminder to stay humble and always feel like there's much to do and much to improve, which of course is the case. '1.0.0' actually scares me a little, personally i would prefer to always be '0.x.x' however good MT may get.

MT can't be judged to commercial standards because it's essentially around 5 busy adults who work or study full time and do this in their limited and precious spare time, unpaid.
However we get treated as if we're Mojang and MT gets compared to MC, many people really should lower their expectations.

[Wuzzy]

So, to summarize, the first number doesn't carry any meaning because it is always 0?
LOL.

[paramat]

Nah, going to1.0.0 has been suggested by celeron55 and we have considered going 0.4.16->5.0.0. Doing either of these would not suggest any sort of milestone.
But also yes the '0' isn't very significant.