The future of Minetest after RealBadAngel's departure

[MirceaKitsune]

I wasn't sure when I'd get around to making this thread. Since hearing the news of RealBadAngel's death, I've been rather demotivated to do anything Minetest related, since going back to the project feels rather strange now. I might remain semi-away for a while, but there's one important thing I wanted to bring up for the time being. Although what happened is very sad, Minetest is still moving forward, and what's to be done next has to be discussed at some point.

We must start considering who's going to manage the parts of the engine that RBA used to maintain, and finish the works in progress he didn't get the chance to. Particularly shaders, the minimap, textured skies, and a few other things. There were a lot of great upcoming features he was working on: Bloom and Depth of Field were close to being finalized, with phong-based realtime lighting planned afterward, together with dynamic water reflections and more. There was hope for even further visual improvements in the distant future, like dynamic buffer shadows and even more GLSL shaders (eg: motion blur).

As far as I know, the dev team doesn't currently have other members experienced with the shader and rendering code. I'm not that good with it either, despite making a few changes for the engine in the past. I'm hopefully not the only one who expects Minetest to someday reach at least the same visual quality as Terasology, so we can't let the shader system be abandoned in a corner. RBA started some of the best improvements Minetest has gotten, and without him it would still look like the early days when we barely had sound and a daytime cycle... now we must finish them.

I was wondering if there are any plans on who will take over shader management, finish the pull requests still active on his account, and continue it in the same direction in the future. Please share news as well as your ideas or suggestions here.

[celeron55]

Frankly, the shader system will very probably be abandoned in a corner. It will be maintained as-is and it might get some performance improvements, but there is absolutely no force that would move that part of Minetest anywhere now that RBA isn't around.

It's just what will happen, there's no need to sugar-coat it. But I guess one can always hope for a new contributor to appear.

[paramat]

We can't force it, someone has to want to do it, whether a new dev or an existing one, their interest will itself make this happen so we don't really need to do anything.

Est31 has plans for integrating the minimap in a better way.
Post processing shaders is WIP, depth of field needs a better implementation, this is something where the existing work could possibly be developed by someone else, i hope it is and that it is considered for merge.
New lighting is something wanted and often discussed by others and may well still happen, although probably in a different form.

All the other plans were essentially his plans for his fork MinetestHD, and are not necessarily sutable for Minetest, so i feel we are not obliged to add them. However it would be good to see someone continue the Meyencandy fork.

[MirceaKitsune]

Heh... that's motivating to know right off the bat >_> Though to be fair, it's understandable as much as it is unfortunate... considering RBA was pretty much the only developer working on the shader code. I'm hoping there will be motivation for someone else to fill this gap, but especially with FOSS projects that's not always the case.

I know it's not obligatory to add every feature in MinetestHD, and I don't suggest blindly doing so. Generally speaking, I was hoping Minetest might finally enter an era of becoming more modern and visually appealing. Features like bloom or DoF or realtime lighting are welcome because they give players who want modern graphics (myself included) the ability to have just that... which I think is a good thing to want.

RBA understood that and agreed with me here. Which is itself a reason I'm concerned about what we'll do without him: I'm not sure if others think this too, or still believe modern graphics are an useless obstacle in the way of running Minetest on very ancient computers.

[TumeniNodes]

not sure if this may be of any help? (or if it may even be weird to even post it but, here goes)
I can try to start researching into this but cannot make any promises

http://prideout.net/blog/?p=1

If it makes no sense for me to post this, please just go ahead and delete

[TumeniNodes]

I can also dig and see if I can get in touch with whomever created this https://www.youtube.com/watch?v=1nvNxlC3efg demo of really nice effects (shaders) they created with lua/irrlicht, maybe get some more info on..., "how?"

[TumeniNodes]

sorry so many posts but this seems to be a decent place to start. http://irrlicht.sourceforge.net/docu/example008.html
Not sure but I am going to give it a go at water effects. I'd like to be able to help in some way, especially since just coming into this community, I was looking forward to possibly getting to work with RBA at some point.
I am sad he passed. Knew nothing of him aside from his work here. But I know he was working on some items I am interested in, such as skybox and skydome (I have made a decent sky for skybox and only switched out my texures for his in the skybox mod, and set the day cycle to nil, and set the time rate to 1.)
But I am very interested in trying to help with water effects. As I said, I cannot make any promises, but I can definitely research and fiddle around and offer any assistance i can there

[paramat]

A new water surface shader was soon to be done, the current waving water was going to be removed because we now have water surface decorations that don't move (waterlilies). RBA's plan was a shader that left water surface stationary but had animated wave patterns as the surface texture. The work might be in his MinetestHD fork.

[philipbenr]

Well, this may be simplistic, but the whole issue I believe lies with our development team, and within Irrlicht itself. It isn't as easy to work in as other engines, which are much more streamlined. I know there was a thread opened that was talking about migrating Minetest to a different engine, but the thread was shot down pretty fast last time I checked. I know this isn't a very productive post, but I don't know to much about the Minetest engine, so I can't say that I can add anything better. -.-
Truthfully, that is probably why people aren't as excited about becoming a dev. The handmade appeal of the Minetest engine appears a bit daunting from a distance, at least from my perspective. There were also some big issues with the dev team not cooperating very well, and many members have left.

It is at times like this that I really wonder about how Freeminer is faring... I checked a while back, and it didn't seem to be moving very quickly...

[celeron55]

Well, this may be simplistic, but the whole issue I believe lies with our development team, and within Irrlicht itself. It isn't as easy to work in as other engines, which are much more streamlined. I know there was a thread opened that was talking about migrating Minetest to a different engine, but the thread was shot down pretty fast last time I checked.

It doesn't matter if a thread is shot down or not; what it actually takes for Minetest to use a different engine is someone to know what engine would actually be better and to have the skills and motivation to migrate to it. It's a near impossible task; practically a borderline rewrite of Minetest. Any benefit whatsoever may or may not be worth the effort. Nobody is telling anyone not to do it - apparently everyone is just deciding for themselves that they won't be doing it. I don't blame them.

[TumeniNodes]

I agree with that view. I have been researching into a couple of features in irrlicht I would like to explore, but documentation and information are not easily. nor readily available. Often it turns into a ghost hunt following link to link, only to hit a dead end with a 404 error (page does not exist) : /
But, the amount of work it would take to completely reverse engineer the entire Minetest entity, to maybe gain one or two neat features is a task I think most would not look forward to. This would entail every bit of code to be rewritten line for line, including every mod, and it would most likely turn into a can of worms.
There are a lot of cool and well documented, massively developed engines out there, but many are not free of cost, so that is another issue in itself.
If a group were looking to make a full featured, robust, fps game with profit in mind then sure, $ would be invested toward the best engine available, or they could even begin to develop their own.
But when a group wants to be able to offer a good game to share with the public, for free..., you would look for decent features, basic, but good, and cost would be a large factor..., so free is actually a really great main feature to look at.
Sometimes the best course is to make the most with what you have, and I think that fits well when thinking about Minetest. And I think that is exactly what the Minetest devs have done so far. But then what I see is many who like the "free", but then expect / want everything only a costly source product can provide. And also a lot of comparing to and expecting a free minecraft.
What I see here, is a group of devs and contributors who have put together a lot of really cool stuff, and still continue to tweak scripts and code to keep coming up with even cooler stuff, working around the engine's limitations, only still to have people expect / want more, and complain about what is not there, which is usually something minecraft related.
I think it would be very cool to exploit the shaders potentials as much as possible. For me, it is so cool to have a real looking sky and water but, people have to remember that shaders also use up a lot of power and can create lag and glitches. And while I think those features are so cool..., even if it did exist, I would most likely normally run Minetest with all the shaders off, etc. to get maximum performance. So maybe another way to go, is to focus more on the graphics and push the boundries of visual illusion, which is what I am working on. It is just as challenging to create a graphic which tricks the mind, as it is to create an actual effect which puts more stress on game performance.
Sometimes a developer can also even create a new ability in the engine and push the code upstream.
Anyway, I ramble a lot.... A LOT :D, sorry

[TumeniNodes]

A new water surface shader was soon to be done, the current waving water was going to be removed because we now have water surface decorations that don't move (waterlilies). RBA's plan was a shader that left water surface stationary but had animated wave patterns as the surface texture. The work might be in his MinetestHD fork.

Very good info, thanks. I'm going to check out his work. And stationary water makes a lot of sense when considering those decorations...
I'm really interested in looking into and attempting the mesh process, I am seeing in quite a few places on the web.
I have noticed the wave feature is kind of messed up anyway as it effects ALL water sources... even water falls, which ends up looking really weird so, he definitely had a point.

[philipbenr]

Well, this may be simplistic, but the whole issue I believe lies with our development team, and within Irrlicht itself. It isn't as easy to work in as other engines, which are much more streamlined. I know there was a thread opened that was talking about migrating Minetest to a different engine, but the thread was shot down pretty fast last time I checked.

It doesn't matter if a thread is shot down or not; what it actually takes for Minetest to use a different engine is someone to know what engine would actually be better and to have the skills and motivation to migrate to it. It's a near impossible task; practically a borderline rewrite of Minetest. Any benefit whatsoever may or may not be worth the effort. Nobody is telling anyone not to do it - apparently everyone is just deciding for themselves that they won't be doing it. I don't blame them.

viewtopic.php?f=7&t=12997

This is the thread I was referring to. There were many people who weren't a part of the dev team discussing the benefits of a different (and maybe better) game engine. Then somebody came in and said that there was work toward the possibility of getting rid of Irrlicht (abstracting it out, was the term I believe), but then somebody said that that dev quit Minetest, so it was not possible. Many devs I know were already against it.

Still, I do wonder how much work that one dev did... Would be nice to see if there was anything reasonable added that allowed for easier development.

Still, as I said, there is nothing I can do, and many are scared away due to Minetest's use of Irrlicht (very hard to find stuff on, trust me), and overall odd setup. I have looked at many other game engines, and Minetest's is definitely one of the oddest looking setups I have seen.

How hard would it be to base a new game (not exactly port Minetest, but build a similar one) to Antarctica and make a new "Minetest-ish game" ?

[rubenwardy]

It shouldn't be too hard as it is a fork.

[benrob0329]

Didn't said dev come back? (I keep up with Github)

[rubenwardy]

RealBadAngel has died.
One of his friends has been logging into his accounts.

[benrob0329]

I meant the dev that was abstracting MT away from Irrlicht, I think it was Zeno?

[rubenwardy]

Oh yes, zeno is back now.

[philipbenr]

Ah... well, I'll have to take a look at the Github a little more closely I suppose, and see what exactly he changed.