Failures of Core Dev and the Needs of Content Creators

[Jordach]

So it's that time again when the community, and Content Creators finds themselves at an impasse with the Core Developers.

Judging by these responses:
viewtopic.php?f=3&t=25507
viewtopic.php?f=3&t=25417
https://github.com/minetest/minetest/is ... -692473048
https://github.com/minetest/minetest/is ... -692978163

Content Creators (mod creators, game designers, artists, server operators) in Minetest have been long under appreciated by Core Developers and are seen by Core Devs as nothing but a nuisance. Should your screen name start with a p and end in t.

For Content Creators who on occasion fix bugs and submit PRs, four weeks on average is a horrifying prospect. Especially in cases where the PR stays open for years or more. Who wants to maintain a stale branch when than that time could've been spent more productively on mods or even bug fixes and better PRs.

The Unofficial Discord practically lit up like a Christmas Tree with activity proves that something is clearly wrong. To quote a few users off of UD:




Want to see the scope of this?


https://github.com/minetest/minetest/is ... 27t+add%22
https://github.com/minetest/minetest/is ... ng+term%22
Open each image, have a drinking game, where, if paramat closes an issue or applies the "won't fix" or "no core dev support long term" label. Take a shot. You'll be dead of poisoning in an hour.

Pictured, the deaths of Issues and PRs:




Core Devs need to strictly review Pull Requests and merge PRs and delegate a yearly or quarterly roadmap to the community; where server operators, game authors and modders figure out where the engine should be directed. Issues going foward should be used for bug reports only. Not Core Developers who immediately shut down a conversation the moment they don't like the idea or are too lazy to come up with an actual argument. Another problem is #minetest-dev and #minetest-hub: the cool kids clubs. By segregating normal users from the Core Developers, and by instituing a depressingly joyless channel; normal users have no want or need to interact with Core Developers. Should Core Developers go missing or inactive, their Core Developer role is stripped and returned to "Active Contributor" status.

Content Creators are constantly testing their content on live servers, while Core Developers rarely even log onto multiplayer servers unless it's a PR they have to test. Core Developers do not play Minetest as a game engine that can supply a theoretically infinite number of voxel powered experiences. But rather, use it as a testing ground, instead of creating or writing a game to experiment with the engine. The devtest built-in game used by core developers is evidence of this. Perfect on paper and in testing. Not perfect in actual gameplay.

Another problem with development is the over insistence of backwards compatibility. Nobody wants to maintain a 8 year old stale mod when there's quite literally much more productive things to be done. Leave compatibility to within one or two minor patches, then it's deprecated or obsoleted. Not kept around like a fetid, festering zombie that won't be put down as it's a beloved pet.

There are even cases where a Core Dev doesn't even know what it could be used for, despite the exact reasons being written in the Issue.

To cap:

[*] Core Developers step down from directing the engine roadmap, and is delegated to the community poll a few times a year to gauge interest.
[*] GitHub Issues becomes a bug tracker exclusively: feature requests are to be moved to the content creator roadmap.
[*] Core Developers must be active within the community both on Instant Messaging/Forums or become a regular contributor. Activity is measured by showing up a few times a week.
[*] To prevent stale Pull Requests, the requirement for merging is reduced to one one approval. The window for closing a Pull Request is 30 days. Core Devs can independantly test it to verify it works in both devtest and existing games and servers. Vetoing a merge can be done, but has a week before it expires and is merged irregardless of problems.
[*] Code reviewing should be moved to more automated testing services. As there are many free plans that offer code coverage testing. Stop wasting developer eyeball time on reading text when they could be testing for issues.

Any questions? (If it feels like a calling out on bullshit, it probably is)

[paramat]

It is disappointing to see you write stuff like this, i think it is beneath you to do this and does not seem in character, you are usually fairly reasonable.

90% of your post is nonsense.

[PolySaken]

It is disappointing to see you write stuff like this, i think it is beneath you to do this and does not seem in character, you are usually fairly reasonable.

90% of your post is nonsense.

I disagree. This post is accurate for the most part (albeit a bit anecdotal), and raises some important points (that may or may not have been raised many times before) in a slightly unorthodox way.
Just because the post mentions you in a critical light doesn't make it unreasonable. You do have a tendency to close things quickly.

[Zughy]

Just because the post mentions you in a critical light doesn't make it unreasonable. You do have a tendency to close things quickly.

I must say, though, in his defense, he's the only one who's at least try doing something about it. Like, every "supported by core dev" label is his. He's been judged by his own actions, which is fine, but let's also remember others core devs usually ignore what's going on, meaning they can't be really blamed because they don't take a stand. Imho that's worse

[PolySaken]

Oh yes, I agree. It's not really a fault of paramat that the only options that seem available is 'close immediately' or 'leave open and ignore forever'.

[Mantar]

If a PR is really old, won't merge without being rewritten, and the guy who wrote it isn't around anymore, then closing it is just the best option available. Well, short of "rework it yourself until it merges again," which doesn't scale to hundreds of PRs.
If it's a really good idea, somebody will come along and make a new PR, possibly basing it on the old one, and that one can be merged.
When the garden's turned into a jungle, sometimes you just gotta whip out the machete and start hacking away.

[Lone_Wolf]

I'd only agree with decreasing to 1 approval for simpler PRs like https://github.com/minetest/minetest/pull/9438 and https://github.com/minetest/minetest/pull/10441 (One of them had 2 approvals and still caused a crash that needed to be fixed afterwards)

[Jordach]

Just because the post mentions you in a critical light doesn't make it unreasonable. You do have a tendency to close things quickly.

I must say, though, in his defense, he's the only one who's at least try doing something about it. Like, every "supported by core dev" label is his. He's been judged by his own actions, which is fine, but let's also remember others core devs usually ignore what's going on, meaning they can't be really blamed because they don't take a stand. Imho that's worse

Core devs not chiming in when they should be, regardless of Issue/PR is another cause for concern. Lack of speaking outside of official business such as release announcements. What do we do when Paramat and Krock are physically unable to contribute due to outside reasons? When was the last time sofar actually acted as a core developer, and not as a server operator for Inside the Box? Where are the core devs who are literally the figureheads of the project? I'm not equating this to a job. But showing up for a couple of hours isn't too hard.

Hopping into the IRC channels for 10 minutes to announce a merge and leave immediately isn't what I'd call being a contributor. Whole point of #minetest-hub was to facilitate such endeavours between server operators, modders and core devs. Citing lack of time to avoid doing so is bad faith towards the users of the software you're effectively managing. There's a thing called IRC bouncers; not exactly hard to setup.

Just ask sorcerykid how she felt about being given the run around for https://github.com/minetest/minetest/pull/9717. GitHub inline reviews are piss-poor for actually explaining the changes you want to see made. Not everyone has the experience of years to understand exactly what you're actually attempting to convey.

[duane]

So it's that time again when the community, and Content Creators finds themselves at an impasse with the Core Developers.)

More hyperbolic ranting?

If what you say is true, the answer is simple: fork the project, merge the PRs yourself, and merge any future updates into your vastly superior fork, which will undoubtedly win the hearts and minds of the minetest community, because its developer listens and cares (and has plenty of time to devote to updates).

[emperor_genshin]

More hyperbolic ranting?

This is the result of pent up frustrations and disagreements within minetest's development. Clearly there is something wrong going on within it's development that's causing this to happen.

For those who thinks ranting is a nuisance: Instead of bashing it, gather around and try discussing why this is happening and do something about it to prevent further disappointment from occurring like civilized grown adults in a community. (Jordach isn't the only community member that's complaining you know)

This is a community project after all, welcome to democracy.

If what you say is true, the answer is simple: fork the project...

Making a fork of this project will never get things done, It only pointlessly tries to divide the community further (IE: Multicraft (MoNTE48), Minetest Final (Old Coder), Freeminer (xyz), etc...).

It will only waste more time for yourself at the same time since not everybody will be willing to stick to forked versions anyways.

NOTE: It's impossible to please everybody in a community-based project thanks to dozens of clashing opinions being on the table within it's development depending on the size of it's community. Which is why we can't simply point fingers at the core devs all of the time since all we are doing is relying on them for everything and constantly filling up more crap to their TODO list without lending them a hand, the core devs can't solve every single problem in the game engine itself, and some of them don't have the knowledge to implement certain requests without aid, which is why some of them state "Why can't you do this yourself?" (I am literally sick and tired of hearing this quote being used since 2015 while no one is doing anything about it).


It's not just the core dev's fault that this is happening. it's the entire community's fault (including the core devs) for failing to efficiently collaborate and understand one another and constantly disagreeing with each other's opinions. This is the main problem to why we can't have nice things for this project and why is it taking so long for something to get implemented or fixed.

Also do not forget Minetest's main purpose. it's suppose to be treated as a game engine for making games, not a full game.

[Zughy]

It's not just the core dev's fault that this is happening. it's the entire community's fault (including the core devs) for failing to efficiently collaborate and understand one another.

I don't agree. They're the ones who should start communicating in the first place. I didn't say a thing for months, I've been helping MT without complaining at all, but then I had enough and I decided to speak my mind. Then I apparently brought hell with my post as I manifested my dissent, but I've been keeping doing my part on the engine nonetheless (1) (2) (3) (4). Even if my C++ knowledge is basic and I don't aim to be a developer in real life, like, at all. I'm, again, doing everything I can for this project, and apparently even what I can't (C++).

But these topics go on without them. Do you want to talk with them? Knock knock on the minetest-dev channel or forget about it. This topic? It got (kind of) answered, on IRC, after I linked it... after days that had been posted. Lejo trying to be noticed? "yes". Matrix blessing request? Unnoticed. So I ask you, how exactly can you understand someone behind a screen if that someone doesn't speak at all and treats you like shit? If you have questions about code on IRC, they're super fast. But the rest? Tabula rasa. And seeing some sort of Stockholm syndrome on someone who's been trying to help makes me fucking angry, outraged and it makes me feel bad for that person, because he/she doesn't deserve this kind of treatment. Like, FOR FUCK SAKE, we're all human beings, aren't we? And yet, there is this mix of fanaticism ("the almighty Celeron") and caste ("we, superior core devs who don't bother about users") which is ridiculous.
Funny enough, though, without users Minetest won't be anything. And congratulations, you succeeded to piss off a member of the community who's been active since 2011 - the author. So, you know, think about that while you're lurking.

Also, yes, given the current situation, to me the best solution is a fork. Jordach is tired, Wuzzy complained, ruben is unsatisfied (and that screenshot: savage. Agree 100%), v-rob too, same for a few names I talked in private with. These numbers are incrementing, and honestly that gives me more hope than anything. And I don't want to be nihilist about the current team; simply take a look at the latest posts on here

[emperor_genshin]

Also, yes, given the current situation, to me the best solution is a fork.

I completely disagree entirely with you on forks. it doesn't invite features nor encourages them for minetest, it only divide the community further.

Name me at least one fork that did good for Minetest itself?

Majority of forks died except for Multicraft (Freeminer 2.0), want to know why?

Spoiler

multicraft is merging changes from MT

Still want to make a fork? It's probably going to die like the rest of them, have fun wasting your time...

Did any of these forks helped the community and the project?

Spoiler

Nope.avi. it only separated the community further. Notice the lack of users on minetest vs multicraft?

Sure it brought a few players on "some" minetest servers, meanwhile it's separating users from this community little by little, day by day. Sounds kind of like a vulture or parasite thing, right?

Do you want to know the "REAL" problem that we as a community are currently having?

Spoiler

We all seem to can't get along with each other, including the core devs.

Yes, core devs are part of this community, if they can't get along with us and we can't get along with them. Then we, as a community, are failing to collaborate efficiently with one another.

For fucks sake, we are a community, we should be working together rather than splitting up and fight about differences!

Instead of helping each other, what are we doing?

Spoiler

Fighting and separating away from each other. Because we utterly fail to set our differences aside and fail to help each other out.

Your recent posts are literally proving my point...

The community is separated already. Us, idiots, and them, people with power not giving a crap. Try to change that

Is this productive behaviour for a community project?

Spoiler

No. It's the beginning of the end of a community if this keeps up.

[Zughy]

Also, yes, given the current situation, to me the best solution is a fork.

it doesn't invite features, it only separates the community further.

The community is separated already. Us, idiots, and them, people with power not giving a crap. Try to change that

[emperor_genshin]

Also, yes, given the current situation, to me the best solution is a fork.

it doesn't invite features, it only separates the community further.

The community is separated already. Us, idiots, and them, people with power not giving a crap. Try to change that

Fact is there must be a change. If no one is going to do anything about this (not even c55 and the core dev team), then this community will undoubtedly fall apart. People will not stay in a community if it's being neglected this way, they will begin to leave little-by-little if this keeps up.

[Zughy]

Majority of forks died except for Multicraft (Freeminer 2.0), want to know why?
Because it's supported by core devs.

Fair point

Instead of helping each other, what are we doing?

Again, personally speaking I'm helping too while complaining. And we're here, debating about things this guy or the other guy don't like. Which is, at the end of the day, a positive thing. But frustration aside, you can't deny we're the only ones doing that. You say we must try to be more unite, civil, and I totally agree with you. At the same time, you can't also deny there is this huge wall in communication which is, like it or not, between the community and the majority of core devs, a wall that wasn't built by the former (see Ruben screen). Look at Paramat's answer here. He could have defended himself, but he didn't bother, maybe because he thought all this was pointless. And maybe for him it is, but for the people who've kept commenting is not, and there have been a lot of people saying similar things recently. So maybe these people - me included - are the problem, but is the silence treatment the right approach? Dialogue is key, always. Maybe core devs are so fed up, stressed, can't even breathe, fine: but let's talk about it, let's understand each other. These problems can't be solved one way.

[emperor_genshin]

So maybe these people - me included - are the problem, but is the silence treatment the right approach? Dialogue is key, always. Maybe core devs are so fed up, stressed, can't even breathe, fine: but let's talk about it, let's understand each other. These problems can't be solved one way.

Silent treatment and forks are never the solution. If there are issues that still exists in this community (which clearly there are issues present) then the only solution is to talk about it. Find what the problems are by understanding one another, formulate a civilized solution rather than making a civil war against each other with forks.

Understand why core devs are hiding behind a wall and refuse to accept help from others, understand why are people still complaining about the same problems and find a structural way to cease this disappointment from building up even though it probably won't make everyone happy in the end since this is still a community project after all (we can only do so much for everyone, but not everything). We all need to start opening up to each other and discuss what must be done to bring balance to this community.

Just like you've said Zughy. Let's talk about it, let's understand each other. These problems can't be solved if nobody is going to discuss about it and solve this problem once and for all.

I hope to see core devs participating in this discussion, because if they refuse to do so (AKA: Not giving a fuck about this, like many are accusing them for doing so) then it defeats the purpose of ceasing conflict and nurturing their community that they themselves are a part of. People will begin to leave this community and this project if that were to be the case which I truly hope not.

[Mantar]

A fork isn't a bad thing per se, and "supporting the community" is putting the cart before the horse, it isn't what a fork needs to succeed, it just needs devs.
If a fork has just one person working on it and he fails to attract anyone else, then sooner or later it dies. If the person who creates a fork manages to show that he has a clear vision and can get things done and attracts other developers, then the community will follow.

Look at Cataclysm, when Whales lost interest in the codebase he dropped it in favor of writing a whole new system, called Cataclysm 2. There were a couple of forks of the original cataclysm made around that time, too, but it was just one, Dark Days Ahead, that swiftly moved to the top because the guy who forked it showed he knew how to run a project and quickly got several other devs on board, and so the community moved over to it.

[emperor_genshin]

A fork isn't a bad thing per se, and "supporting the community" is putting the cart before the horse, it isn't what a fork needs to succeed, it just needs devs.
If a fork has just one person working on it and he fails to attract anyone else, then sooner or later it dies. If the person who creates a fork manages to show that he has a clear vision and can get things done and attracts other developers, then the community will follow...

"Yeah, let's split the community more because of unresolved dysfunctionality..."

That's all i'm hearing right now. Very problem solving indeed... (*sarcasm*)

It's actually kind of sad having to hear this all of the time.

[rubenwardy]

Core Developers step down from directing the engine roadmap, and is delegated to the community poll a few times a year to gauge interest.

I've included community comments from the roadmap brainstorm when drafting the roadmap. This process can be repeated when needed

GitHub Issues becomes a bug tracker exclusively: feature requests are to be moved to the content creator roadmap.

I have actually suggested this before - the problem is when technical discussion is needed on a particular feature. Perhaps when a feature is approved, an issue can be opened on the repository for technical discussion

Core Developers must be active within the community both on Instant Messaging/Forums or become a regular contributor. Activity is measured by showing up a few times a week.

I agree with this, but there should be leeway for periods of absence.

To prevent stale Pull Requests, the requirement for merging is reduced to one one approval. The window for closing a Pull Request is 30 days. Core Devs can independantly test it to verify it works in both devtest and existing games and servers. Vetoing a merge can be done, but has a week before it expires and is merged irregardless of problems.

I don't agree with this. I think that merging PRs can be made easier by automation. One approval can perhaps be extended for smaller PRs, but not for big architectural changes. It also shouldn't be possible for a single core developer to merge big changes by themselves.

Code reviewing should be moved to more automated testing services. As there are many free plans that offer code coverage testing. Stop wasting developer eyeball time on reading text when they could be testing for issues.

I am a massive proponent for greater automation, I have attempted to introduce completely automated code style linting. I think that having 95% of the code correctly styled justifies the edge cases. Currently, we have something like 5% correctly styled but at least a few edge cases are ok

Another problem is #minetest-dev and #minetest-hub: the cool kids clubs. By segregating normal users from the Core Developers, and by instituing a depressingly joyless channel; normal users have no want or need to interact with Core Developers. Should Core Developers go missing or inactive, their Core Developer role is stripped and returned to "Active Contributor" status.

#minetest-hub's intention is to create somewhere with devs, server owners, modders, and other contributors. I personally think that it doesn't serve much purpose now that IRC is much smaller. #minetest-dev is necessary to keep focus on development

[ShadMOrdre]

Core Developers must be active within the community both on Instant Messaging/Forums or become a regular contributor. Activity is measured by showing up a few times a week.

I agree with this, but there should be leeway for periods of absence.

I think in a more general sense, this means not any particular core dev, but the entire core dev group. In this way, the absence of one or two, out of 5 to 8, is not as detrimental as the absence of 4 to 5 out of a group of 5 to 8.

Celeron55 still lurks here and there. Paramat, rubenwardy, krock can be found here on the forums, for the most part. Out of the three, rubenwardy is by far the most responsive, followed by paramat. Rubenwardy and krock can be commended for their civil discourse, and willingness to at least acknowledge an issue and talk through possible solutions.

While IRC, Discord, reddit, and other websites contribute to the big conversation, those thoughts and ideas are not expressed HERE, on the MT forums, where the community actually is supposed to exist.

The MT website, as a whole, can and should provide all the necessary tools for users, developers, game-makers, and modders. By this, I do mean links to other apps that are useful, such as gimp and blender, but also should contain the conversation. Too often, parts of the conversation are pointed to offsite, third party websites, or to the IRC log. Logs are hard to read for those who are familiar with them, which is why there are so many tools for reading them. Pointing to reddit or Discord means a community conversation, all the IP involved, is now owned by the third party website hosting the tool. It also means that some may not even be able to access those sites.

While the MT dev wiki website provides links to gimp and blender, and to the IRC logs, we are constantly reminded that the dev wiki is not a good resource, when that is actually not true. Just because it isn't updated with the newest API features doesn't mean that users of MT 0.2 can't use that info.

There are so many issues.

The only solution is what everyone is saying, just in ways that are off putting to some, downright dirty to others, or just dismissive in general.

Community. IF we are a community, then participate in the community! In whatever way you can! Can you type? There are articles to be written. Can you use gimp? There are textures to be made. Not Texture Packs, but textures. Can you code? In C? In Lua? Offer code. Can you use blender? Offer models.

It is that simple folks. Communities are by definition a grouping of individuals of a shared interest working towards that interest.

Let's be the community. Complaints are welcome, when accompanied by solutions. Ideas are great, but are only useful with prototypes. Words are cheap. Action, not words, gets things done!

Since celeron55 has "released" the project, an argument could be made that he no longer can be the BDFL. The L still stands for leader. If you wish to not lead, then appoint someone please. Not necessarily a core dev.

Shad

[Termos]

Making a fork of this project will never get things done, It only pointlessly tries to divide the community further

What's wrong with dividing the community anyway?
Not that it's particularly uniform thing, people come and go and there are different crowds that want different things, sometimes mutually exclusive.

MT development is stale and there's no way of changing it, so why keep the dissatisfied crowd around only so they write rant posts and keep bugging developers about stuff they probably don't understand.
And even if progress were possible in the current setup, there's the crowd of people who like their game just as MT currently is - without too many moving parts, and even fiercely defend some bugs they've come to appreciate. Why ruin the fun for them by fixing MT and improving upon it?

A fork is the only way of moving forward, and it's a perfectly normal thing in open source, going all emotional about it and trying to make it look like a hostile act looks a bit silly to me.

[Lone_Wolf]

Making a fork of this project will never get things done, It only pointlessly tries to divide the community further

What's wrong with dividing the community anyway?
Not that it's particularly uniform thing, people come and go and there are different crowds that want different things, sometimes mutually exclusive.

MT development is stale and there's no way of changing it, so why keep the dissatisfied crowd around only so they write rant posts and keep bugging developers about stuff they probably don't understand.
And even if progress were possible in the current setup, there's the crowd of people who like their game just as MT currently is - without too many moving parts, and even fiercely defend some bugs they've come to appreciate. Why ruin the fun for them by fixing MT and improving upon it?

A fork is the only way of moving forward, and it's a perfectly normal thing in open source, going all emotional about it and trying to make it look like a hostile act looks a bit silly to me.

This seems to be mainly fueled by your anger over the collision PRs. If it makes you feel any better: https://github.com/minetest/minetest/is ... -714572073

MT development is stale and there's no way of changing it

I don't feel like this is the case at all. MT just needs to tie its shoes properly so that it doesn't trip up when it has a lot of walking to do.

trying to make it look like a hostile act looks a bit silly to me.

It's not hostile if pretty much everyone agrees development is dead. But most of the community don't agree on that, so creating a new fork is a hostile act to them.

[rubenwardy]

Creating a fork because you disagree with direction isn't necessary hostile. It's hostile if your intensions are hostile, or if you perform hostile behaviour like spamming or relicensing to avoid backporting. Forking is a natural part of open source, licenses allow derivative versions.

I support improving the movement and collisions, the current behaviour isn't especially great. There was a lot of controversy about this a few years ago, the compromise made was to keep old_move as a movement option

[sfan5]

And even if progress were possible in the current setup, there's the crowd of people who like their game just as MT currently is - without too many moving parts, and even fiercely defend some bugs they've come to appreciate. Why ruin the fun for them by fixing MT and improving upon it?

I get the impression you are referring solely to paramat, who often bends over backwards to construct arguments that bugs are features, should not be fixed and were intended to be that way all along. That's nonsense too in my opinion.

But I can't think of anyone else, isn't a "crowd" multiple people? You might be exaggerating the issue a bit.

I support improving the movement and collisions, the current behaviour isn't especially great. There was a lot of controversy about this a few years ago, the compromise made was to keep old_move as a movement option

I wonder how many servers are actually using old_move. The controversy about it was great but I doubt it's that many.

[Mantar]

That's all i'm hearing right now.

I don't know who you're listening to then, because that's not what I said, nor anyone else that I noticed here. A good fork can revitalize a dying project, (not to say Minetest is dying, but it's seen better days) it's not by definition either harmful or hostile.