KSP should require a Standardized Mod Versioning File

[ferram4]

Guys guys guys, you have taken a very simple proposed idea, trivial for everyone to implement, and turned it into some kind of monster. What I am proposing is that KSP require that all mods contain one additional text file in their distributions that contains 3 trivial pieces of information:

1) The version of the mod in a standardized format

2) The version of KSP that the mod was developed to work with

3) The date the mod version was released

That's all! Very minimal, literally anyone with zero technical training could do it for you. But it would drastically improve indexing of mods and allow mod update utils for end-users. I'm proposing that Squad FORCE mods to have this file in order to function properly - not sure why everyone is getting bent over being forced to add a trivial indexing file - Squad 'forces' you to do all sorts of things to make your mods work - like using correct spelling and capitalization of intrinsics (!?!?!?). I think that folks must be misunderstanding what the original proposal was because this is pretty minimal stuff here guys and in no way deserving of the outrage that seems to be present here.

Actually, I'm kind of annoyed at the frequency with which you call this trivial. It does take time to collect all the data, make sure that it's correct, and make sure that KSP functions properly with it set (after all, you are requiring it to make the mod function). General trend that I've seen is that in programming, the more someone talks about it being "trivial" to implement and make work, especially when someone else has to code it and make sure it works, the more the person talking has no idea what they're talking about. Repeating that it's "trivial," "minimal" and "reasonable" again and again just makes me think that you don't actually have a compelling argument for this and feel like if you assert your "reasonableness" enough you'll eventually convince everyone against you to agree, despite the fact that you've been implying that all of our concerns are simply unreasonable and can be dismissed.

As to the difference between coding changes and this, when the game's API changes there is generally a good reason for it; a new feature got added that required it, code got refactored for performance, bugs got squashed. You're asking for, "do this to make your mod not break," and you still haven't responded to my point about how it will discourage new modders who don't know why their first attempt at a mod doesn't work for no apparent reason.

You do have to understand that there currently aren't any mod managers that require such a metadata file for them to work, and arguing that we need one because it could be used that way is really arguing for a solution to a non-existent problem.

Edited March 17, 2014 by ferram4

[Red Iron Crown]

Is there a reason why version numbers are needed? Couldn't the mod version be its release date? (Disclosure: I'm not a modder or programmer, it may be blindingly obvious to all of you why this is a bad idea.) I must admit, from the outside the three pieces of information Oddible is looking to have standardized do seem trivial and are information that is included with any reasonably well maintained mod, but I certainly could be wrong about that.

I think it is desirable to have some sort of mod management either built into the game or as a "meta-mod". It would make it easier for players to install, remove and update mods, which I can only consider a good thing. If what Oddible suggests is not the best way to go about it, what is a better way? What obstacles need to be overcome to have centralized mod management become a reality?

[Dodgey]

Is there a reason why version numbers are needed? Couldn't the mod version be its release date

I'm pretty sure it's just to help with organization.

[Amaroq]

Red Iron -- anything other than a version number is prone to breakage, for example, when some ambitious developer decides to release two or more versions on the same date. (Either due to ambition, or due to the need for an immediate bug-fix.)

O.P., you haven't really done much to indicate why you think this is such a "significant improvement for end-users"; rather than trying to minimize the impact your proposal will have on ferram and the other mod developers, why don't you paint a picture of how this improves things for all of us? What are you trying to accomplish with this?

[Guest]

Actually, I'm kind of annoyed at the frequency with which you call this trivial. It does take time to collect all the data, make sure that it's correct, and make sure that KSP functions properly with it set (after all, you are requiring it to make the mod function). General trend that I've seen is that in programming, the more someone talks about it being "trivial" to implement and make work, especially when someone else has to code it and make sure it works, the more the person talking has no idea what they're talking about. Repeating that it's "trivial," "minimal" and "reasonable" again and again just makes me think that you don't actually have a compelling argument for this and feel like if you assert your "reasonableness" enough you'll eventually convince everyone against you to agree, despite the fact that you've been implying that all of our concerns are simply unreasonable and can be dismissed.

Personally, I think it's trivial for a user to figure out when and if a mod has changed, or whether it can be used with the current version of KSP, or how to install it simply by doing a little bit of light reading, but we all know how well assuming that will go over. How it stands now works pretty well for me, at least, and I run a non-trivial amount of mods.

[avivey]

I think this thread is getting too overheated over an issue that shouldn't be very big.

Personally, I think a standard version file is a great idea, but (1) it only makes sense in the context of a mod manager, and (2) nothing that is forced ever works - people find ways around it. It's also annoying to mod writers.

A mod manager can require a version file (i.e. not work without one). A super-popular mod manager can be a force to make "all" mods agree on a format. Several managers that have a lot of users between them, and agree on a file format will also work.

I've seen a couple of mod-managers laying around, I don't know how much traction they have. My guess is, if a mod manager starts by supporting a big list of popular mods without the version file + any mod that has one, it can get popular.

[capran]

Had to register just to reply to this thread. I'm not a modder myself but as a user of the KSP program it's very tedious work just to find the right mods you want by itself (Spaceport is a clusterf***) only to have to go through 15+ zip files to install manually. which i know is a cut and paste for the most part. I only play occasionally, so keeping up with every updated mod is a exercise in futility. I basically have to do a clean install every time or I have tons of issues. Maybe I'm not doing it right. As modder shouldn't you be catering to the users of your mods making them more available to the users otherwise why even put your mod out there. How much work would it actually take to package it for a program like KSPMM?

With KSP's release on Steam they already have a very effective repository system in place AKA Workshop. Squad just needs to implement it in game. Which is likely not an easy task granted. That just works for the steam users though. Nexus (i think) and Curse both have repository systems in place as well but they function more like KSPMM. I think the problem really comes down to Squad either not wanting to or not able to implement this anytime soon due to a small dev team.

some of us modders are stubborn and dont like to be told how to work our magic. we would sooner see the world burn than change our ways.

that said if someone were to hypothetically develop the ultimate package manager that makes mods stupid-easy to install and also came with dev tools for quickly making packages to the correct format. its possible that such a thing could become a defacto standard that everyone uses just because it makes everyone's life easier. such a thing would have to cover all the bases though. it would need to install dependencies, manage updates, then resolve compatibility conflicts and all the other things that can go wrong with an automated installer. so its a pretty big task on its own. even then you would still get rebels who like to do things their way and ignore it completely, or worse have competing tools that do the same job but dont work interchangeably. the only way i see this working is if squad or other community leaders took the reigns.

I completely agree with you guys...Squad just needs to implement workshop or something similar! SOON. KSP stock is fun at first, gets old. It's the mods that make it FUN and ADDICTIVE. But it's a pain in the ass to have to keep them up to date! I have 53 folders within my GameData folder right now. FIFTY-THREE.

Has anyone put up a petition yet on this topic?

[sievers808]

I'm not a modder and I don't know the first thing about how they are made and regulated but usually the more popular mods are rather clear about what KSP version they're for and especially what dependencies they need.

If anything, I think modders should adopt the same file structure when zipping up the mods. Some mods have a root folder then Gamedata then the mod folder... some don't have Gamedata in there, some are totally random. For installing a couple mods at a time it's not really a problem, but when you're loading 50 mods into a fresh install and you have to unzip each one individually and properly place them it can be tedious. If every zipped mod had Gamedata/modfolder/modstuff then you could just unpack all 50 at once into the root KSP folder.

Not that I really mind, nor would there be any realistic way to enforce it, just a thought that I've had a few times.

[Justin Kerbice]

some of us modders are stubborn and dont like to be told how to work our magic. we would sooner see the world burn than change our ways.

that said if someone were to hypothetically develop the ultimate package manager that makes mods stupid-easy to install and also came with dev tools for quickly making packages to the correct format. its possible that such a thing could become a defacto standard that everyone uses just because it makes everyone's life easier. such a thing would have to cover all the bases though. it would need to install dependencies, manage updates, then resolve compatibility conflicts and all the other things that can go wrong with an automated installer. so its a pretty big task on its own. even then you would still get rebels who like to do things their way and ignore it completely, or worse have competing tools that do the same job but dont work interchangeably. the only way i see this working is if squad or other community leaders took the reigns.

Donkeys always need some proper motivation (either carrot or stick) .

Look at license thing, many modders add it because THEY HAVE TO, but they don't really know what they're doing (GNU licenses especially are not that's good) and many are too restrictive, and at the end, when author have gone away for good, license random choice could make life harder for all people because it may not allow to continue working on the mod, even if the author would agree with it.

As Majiir said too, plug-in have to include at least a version number, it help a lot for compatibility checking and bug reporting too.

(easy case: plug-in X come in mod 1.0, in mod 1.1 it is updated but have the exact same size, and author don't add any version in a readme or anything, archive name is the same: "mod X". A player install mod X 1.0, another mod X 1.1, both have too different issues, when they report them, which version they'll use ? Especially when a bunch of mods are installed at the same time.)

When new rules will be added to allow mod to be posted there, modders will follow them, 'til that day they still continue to carry their own bad habits.

(by the way rules should make a minimal documentation/readme mandatory, many mods are just files in an archive with a license users don't care about. If forum post vanished for a reason or another, ... what happened then ? No more doc on that's mod !)

Finally, there exist one good and long time validated model: the Debian package sytem, all tools can also be used for anything other thing as such a package is just a gzip'ed tarball. It's quite complex but it's robust, only when packager do their job well sadly (no error free system exists yet, even the ADA language with all it's crazy tight checks, and ultra rigid syntax, remember Ariane 501)

[blizzy78]

GNU licenses especially are not that's good

That's why the Linux kernel uses the GPL, right.

[Red Iron Crown]

It's too bad Curse has such a bad rep, making many modders choose not to host there. Their client is supposed to be able to track mod versions and help keep them up to date, but if the mods you're using aren't on Curse then that obviously won't work.

[sarbian]

Why don't you people do some real mods out before telling us how to "do it better" ? You'll see (as other said here before) that it require a LOT of time. We have to code/model/whatever, test, adapt to every KSP quirk, deal with endless user complain that we should have done x and not y, manage to write some doc, AND (and this is the part that some users seems to forget) have a life/work/study/whatever beside KSP.

I agree that some uniformity between mod install would help (like packaging the gamedata folder) but don't come and tell me what to do when you don't pay me s... anything.

Edited May 30, 2014 by sarbian

[Justin Kerbice]

That's why the Linux kernel uses the GPL, right.

What is this suppose to means ?

Linux kernel used to be a great project, but with more devs adding stuffes into it, it became messy. Did DevFS (in 2.4 version) troubles ring any bells ?

Anyway, it's not because one project, whatever it is, use a license, that's this license is automatically "tainted" as good as the project. GPL has rised a lot of very hot discussions on many projects related to Linux (distributions for example) due to its inheritance part.

[Greys]

Why yalls necro~ This thread was dead and burried and gone and forgotten but now it's back here with us and it sucks! The world is burning and corpses walk the street, thousands killed! You killed them! Ok I'm out of overdramaticjuice. Seriously, this was dead and gone and for good reason, three pages of it, but now it's back with people saying the same thing as always and quoting THE FIRST FEW POSTS IN THE THREAD!

let it die, there's nothing left to say and the idea has been rejected on all fronts.

If you really want a standard, compose it, and try to convince INDIVIDUAL modders to use it. I like my mod.xx.yy.zz.version file system, it's simple and easy, but I'm not interested in convincing other modders to take up the standard and maintain it. Conversations such as this thread do literally nothing except make people annoyed. It's like talking politics at work, nobody cares what you think and nobody is going to change their mind because of your words. Action makes the difference and general action isn't good enough, you can't wait for people to find you and agree with you and opt in to your standard, you have to go to them and convince them specifically.

Hope this thread burns in a pity full of grapes

Love Greys

[Eric S]

some of us modders are stubborn and dont like to be told how to work our magic. we would sooner see the world burn than change our ways.

Was about to say that that's true of a lot of programmers in general as well, but really, you can substitute "people" for "modders" there. So not a flaw specific to modders :-)

[Rich]

The topic of this thread has long since been brought to a conclusion, and as a result I will be locking this thread.