Shouldn't Minetest 5.0.0 actually be called...

[SKCro]

Hey everyone,

I know I'm new here, but:
Shouldn't the new Minetest 5.0.0 release actually be called "Minetest 0.5"? If you actually made the first stable (neither alpha or beta) version of Minetest, it should be called "Minetest 1.0" not "Minetest 5.0".
EDIT: It should be called "Minetest 0.5" because the previous version of Minetest has "Minetest 0.4.17.1" in the title bar, and it doesn't make sense to skip all the way to Minetest 5.0...

This is the way it should be (but imagine it saying "Minetest 0.5" instead of "Minetest 0.4.17.1"):

VersionNumberPic1.PNG

This is NOT the way it should be:

VersionNumberPic2.PNG

Your new beta tester,
SKCro

[sofar]

Hey everyone,

I know I'm new here, but:
Shouldn't the new Minetest 5.0.0 release actually be called "Minetest 0.5"? If you actually made the first stable (neither alpha or beta) version of Minetest, it should be called "Minetest 1.0" not "Minetest 5.0".
EDIT: It should be called "Minetest 0.5" because the previous version of Minetest has "Minetest 0.4.17.1" in the title bar, and it doesn't make sense to skip all the way to Minetest 5.0...

This is the way it should be (but imagine it saying "Minetest 0.5" instead of "Minetest 0.4.17.1"):

VersionNumberPic1.PNG

This is NOT the way it should be:

VersionNumberPic2.PNG

Your new beta tester,
SKCro

no

[benrob0329]

The developers switched to a sane versioning scheme, and wanted to avoid the press of "the 1.0 release".

Its not wrong, its still a larger number and a lot of people got confused by the leading 0 anyways.

[SKCro]

Why would they want to avoid "Minetest 1.0"? Its weird that they skipped to 5.0 and never did 2.0, 3.0, or 4.0...

-SKCro

[benrob0329]

Because "1.0" would imply a fully finished product, which MT will likely never be due to the nature of it's development. So they decided to just drop the leading zero.

[sorcerykid]

Minetest actually underwent two changes to the versioning scheme almost concurrently.

The first change was to renumber development builds based on the forthcoming release (whereas before it was derived from the previous stable branch). I brought up this concern a couple years ago, as it was causing a great deal of confusion when distinguishing between stable and development builds. That seemed to generate enough interest that the core developers opted for a saner versioning scheme.

viewtopic.php?p=285668#p285668
viewtopic.php?p=288006#p288006
viewtopic.php?f=18&t=18331

The second change was to adopt a variation of SemVer, as both the engine and the game long since surpassed the alpha stage. Core developers discussed this thoroughly on GitHub and IRC before reaching a consensus. It was decided that removing the preceding zero would ensure a greater sense of continuity between releases. Also some users were already referring to Minetest 0.5.0 as "5.0".

viewtopic.php?p=322733#p322733
viewtopic.php?p=322837#p322837
http://irc.minetest.net/minetest-dev/2018-06-13

Here's the GitHub issue where both matters were addressed: https://github.com/minetest/minetest/issues/6073

[Andrey01]

Why would they want to avoid "Minetest 1.0"? Its weird that they skipped to 5.0 and never did 2.0, 3.0, or 4.0...

-SKCro

They used to skip nowhere. The core devs just got rid of unnecessary "0." in the beginning of each version, that`s all.

[v-rob]

Actually, this happened before as well. In the super early alpha versions, it went from 0.0.1 to 0.2, so this would make it the second time a leading zero has been dropped.

[MineYoshi]

You know, devs here do whatever they want and we comply with them without greater hassle.

Try to find some "real" bugs instead if you want to help the game's development.

[Gibeinumo]

Wow, it is called 5.0 and not 0.5. It took me six days until I realized it. I guess it does make sense because it already is a playable, good game.

[Linuxdirk]

It was decided that removing the preceding zero […]

The bogus version number now being used as base for future versions is always justified with “oh, we’re just removing the leading zero” but what was actually done is simply skipping to the 5th major version in the version number scheme that was used before.

Like with Chrome or Firefox releasing a new major version for every minor set of changes just to have a higher number. Or like Nvidia coming closer to major version 400 with their drivers with every release.

This is just ridiculous.

[Skulls]

If you are looking for more information in a version number than it being before the things after or after the things before you are overloading that poor number and should show some kindness to your data structures.

It's a number that goes up. Everything important is in the change log.

[sorcerykid]

The bogus version number now being used as base for future versions is always justified with “oh, we’re just removing the leading zero” but what was actually done is simply skipping to the 5th major version in the version number scheme that was used before.

That's not true because the version scheme changed. It wasn't as simple as just removing the leading zero (altho that was a justification for consistency's sake). But in truth, 0.4.x and earlier releases used a version scheme that was contrived specifically for Minetest development. Even core devs admitted this on IRC. Following that unofficial "convention", the second number represented a breaking change. In contrast, the current version scheme is more closely related to the SemVer standard, in which the first number represents a breaking change. Hence, 0.5.0 is correctly translated to 5.0.

[Linuxdirk]

Yes, this is how it was justified afterwards. Minetest never used semantic versioning and does not use it with 5.0.0. The version number is just nonsense, or as Skulls said: "It's a number that goes up. Everything important is in the change log."

[rubenwardy]

Minetest does now follow semver, with slight changes to how we label development versions

[sorcerykid]

Um no, this was not how it was justified afterward :P Please read the IRC log from June 3, 2017.

http://irc.minetest.net/minetest-dev/2017-06-03

You'll see there was an intensive discussion between Wuzzy, VanessaE, and the core devs about finally adopting semantic versioning a full year prior to the paramat's proposal on issue #6073 about dropping the leading zero from 0.5.0.The core devs were legitimately trying to adhere to the SemVer standard (except for labels of development builds as rubenwardy stated), not just using it as an excuse to bump the version to 5.0.

[Punk]

It's to match the kernel 5.0.

[rubenwardy]

It's to match the kernel 5.0.

Nah, pretty sure they copied us

[paramat]

The previous leading zero would have stayed with us forever, as MT is not beta, is not approaching some kind of release or 'completed' state or 'version 1.0.0'. So we removed the leading zero. MT hasn't skipped through 1.0.0, 2.0.0, 3.0.0 etc.
Previously it was '0.major.minor.patch', now it is 'major.minor.patch'.

Concerning '1.0.0': MT was fully functional in MT 0.3.0 in 2011 or even earlier, so the new version scheme gives the suitable impression that '1.0.0' was a long time ago and that this is the 5th major version, which it roughly is.

The issues, PRs and discussion show it wasn't justified afterwards, it was decided before. Linuxdirk is writing negative nonsense.

[Linuxdirk]

The previous leading zero would have stayed with us forever,

Not if Minetest followed SemVer from any release in the past. Which was obviously not wanted (otherwise Minetest would use SemVer as of today and not some sort of SemVer-influenced made-up versioning scheme).

MT is not beta […] MT was fully functional in MT 0.3.0 in 2011 or even earlier

It indeed is functional. But I don’t see where skipping from 0.4.17.1 to 5.0.0 makes any sense having the (feature-rich but) incomplete API and zero polishing in mind. Even if it was discussed in “dev space” before the change and justified towards the players afterwards.

Why wasn’t a clear cut made and said “1.0.0 is the new version and from now on we’re using SemVer”? From a user point of view “skipping” from 0.4.17.1 to 5.0.0 makes no sense. Even the explanation of moving from 0.4.17.1 to 5.0.0 by removing the leading zero seems made-up for “marketing reasons”. Why was it there in the first place? Why wasn’t it removed 2011?

… and why the hell are we discussing a stupid version counter? :)

[Jordach]

… and why the hell are we discussing a stupid version counter? :)

Because someone can't handle change for the better, clearly.

[Linuxdirk]

for the better

[citation needed]

[ShadMOrdre]

It is what it is. MT is a great game. What difference does a number make?

Versioning schemes should somehow divulge a method of easily identifying the latest version. A number just goes up, and well, which number do I need? Knowing that a product is on any given version is only relevant to the users of that version, and to those that help them....and to salespeople looking to market to new customers.

What I really want, is to be able to identify, by sight, what that means. A date, perhaps. I can look for the latest date. Otherwise, I have to weed through countless sites, just to find the version of a product that I need, cause, you know, the latest greatest don't always work on my oldie but goodie. And sometimes, I just want an earlier version.

And most especially, my customers need me to show them HOW, not just bilk them for mo' money.

[Festus1965]

Failuretest ?

[wziard]

[edit] Ah nevermind...

Version numbers are too unimportant to even discuss... :-)

Ik like the date system ubuntu uses (13.4 is from april 2013), for the rest I don't care as long as it goes up.