Server plugins -> reinventing the wheel again and again

[reaby]

Hi everybody,

We had an interesting talk at discord #development channel today. On topic re-inventing the wheel again and again for the community written serverplugins. What if we servercontroller people would actually come together, instead of one coding for this, one starting that project... all working for one big project, if it's even possible... and build ONE DAMN AWESOME CONTROLLER with a BIG TEAM for the game we all like. With a big team we can go faster forward, and if we can agree what to build it will be awesome since of so many different point of views come together.

What we thought at discord was, that if we agree, the that base controller would be writeen in ??? (other than php) and plugins possible in LUA, or native ??? (other than php). Plugins in LUA since it has wide variety of language support, and it's widely used for game scripts. but this is only initial idea, it sure can be thought over..


Is this idea in vain or would you be interested?
I think 8 to 10 persons would be enough actually to even think of starting this

And this is only some initial invite, everybody interested are welcome

Maybe first just reply what you think about, if we just could do collaboration together, and then if agree, we could actually build same project together, then invent what we want to build and then just building it together

Decide the base language for the controller, maybe JS, Python, Ruby, even Java, C# ?

So here comes the invites:
Tgyoshi
Toffe
Them
W1lla
Oliverde8
Undef
Guerro
Miss
Kremsy
Steffeen
Domino54
Xymph
Slig
Reaby
ManiaDesign

Banane9
PRGfx
zocka
thebigboo

Colors: pending, Not agree, Agree

Let's see how this comes together, or if this is just an idea that wasn't ment to happen
What you think ?

Edit:
Discussions for now, LUA for plugins is dropped, was not good idea.
Discussion about Co-routines vs threaded application, needs more of thinking
Also scripted language vs compiled language, looks like scripted one would be preferably better, but lets see how it comes together.

Be welcome to discuss at trackmania discord #development channel or just post here

[Dommy]

I'm in.
I would recommend building it in C#. Although I don't know much about "normal" programming languages, I learn quickly and might help with core itself.
You have me for design of client side manialink interfaces and their script communication.

[Guerro323]

I can help,
What I can do :
- Core structure
- Structured Class/NameSpace/Folders
- Communication between Players's Manialink/Server ( sending vars to player without refreshing a manialink )
- The design of the plugins ( how a plugin will be done, and the structure of it )
- A part of the GBX parser and xmlrpc thingies.

What I can't do :
- Creation of the plugins. ( I can, but not too much )
- Databases thingies. ( if this is a non-sql database, why not )
- How will look the controller ingame.

I can do in those language(s) :
- C#.Net/Mono
- other object-languages ( if the syntax look ok for me ).
- BASIC ( joke xddd )

Even if I'm creating a controller, I could drop my current controller and focus on this one.

[oliverde8]

Agreed, but not all the way.

For me definitively not LUA. The issue with LUA is that you will need a C++ system behind it. which means compilation, which means harder to install. Then of course you we can create the binaries and have then each plugin in Lua. Why not but a lot of work.

The main reason PHP is awsome is because it is easy to install. EAsy to code with, easy to extends... In my opinion the only issue with PHP vs another language is threading.

But then Java is east to install as well. Ruby is a pain on Windows with dependencies hard to get and stuff.

What I don't agree upon as well is,

. With a big team we can go faster forward, and if we can agree what to build it will be awesome since of so many different point of views come together.

Very simple to do a child you need 9 month ?
If you put 9 women together can they do a children in 1 month?
No

At first at least a big team, won't make things go faster, solid foundations needs to be put together by a few people after long deliberation and planning. Once that is done then multiple plugins can be coded at the same time but even then it's important to organise developments.

In any way such a project needs someone that knows exactly what needs to be done and have enough time.

Something that surely needs to be put in place is code reviews,

Last, I must say for the N'th time, I am tired of this game, eXpansion is the last thing keeping me here, as I don't want to drop the project while so many admins have trusted us. I won't be part of the project.

The game is dead guys, Nadeo has nothing left in their hocus pocus bag.

Edit : Look at what happened with MLEPP, we were to many coding stuff all our own way without enought supervision. Everyhing put together? it didn't work

The foundation,
=> Dedicated / Controller communucation
=> Chat & chat commands
=> admin system
=> configuration system
=> update system
=> multilanguage
=> window system
=> Basic elements to normalize all the displays

takes time to make properly.

And why not simply eXpansion? It works, the bases & more is there. It's far from perfect, code is not clean but that can be done. Why to go from scratch?

I am not convinced at all.

[Florenzius]

What I can do:

-Interface
-Translations
- logo
- intro (for a youtube channel with news maybe)

[w1lla]

As said by Oliverde MLEPP was too big of a team with no organisation.

The basis first would be to see what the team should develop with the type of coding, then look for the way it should be dealt with accordingly as too databases/structures/manialinks and plugin functions or features.

Then you should look for who will handle with what accordingly to the way it was worked out.

I am not saying no to this but it has to be a sublte structure, so a yes for me.

[reaby]

As said by Oliverde MLEPP was too big of a team with no organisation.

The basis first would be to see what the team should develop with the type of coding, then look for the way it should be dealt with accordingly as too databases/structures/manialinks and plugin functions or features.

True, we need stucture and somewhat organizing... MLEPP was great project at the time, I think we all who participated learned alot from it - even it didn't succeed at the end, but i think we learn from mistakes, if you don't try you'll never know.

From that experimence, I know, I'm not suitable to lead a project, I don't just have what it takes, and thats why who ever thinks of being capable of taking the lead, is more than welcome to say so. I agree on what Willa and Oliver said, it's not going to work without decisions made by leading team or leader, who keeps team organized... But then how to keep it fun for everybody participating most of the time

Anyway, it's important to discuss at first, and not too much, but enough.. on what each of us would like to build, i have just this idea, to gather our knowledge to build something working and useful. I don't know yet what it would be, but i think it should be mixture of current server controllers combined with the best parts.

Some ideas this far:
One controller instance should be capable of handling multiple dedicated server.
Personally I would like to see controller be easy to use, using graphical admin tools.
So what's your ideas, which would be fun to participate ?

[Xymph]

I stopped actively working on TMF/MP controllers a long time ago for multiple reasons, and will not return. Although I have my doubts (some already voiced by oliverde8) I wish you good luck.

[undef.de]

Unless Nadeo doesn't come with bug fixes, i discontinue all of my work i've done till now.

Also the amount of players need to increase strong till i come back and start working again (it makes no sence for me, to do all that work for the current size of the community).

Also i won't be a part of a project which uses an other programming language then PHP (because of all current controller are using that and the server admins knows how that work).

As well i don't join this project for the lack of time reason: i don't want to start from scratch for all my plugins, because i think the core structure will be require a lot of changes, which makes it easier to start from scratch (that was the main reason to start with UASECO).

Conclusion for me: No thanks, but good luck!

[reaby]

So it was just sunday mornings stupid idea...

I think we all have just different views, and passion for the scripts we've started already, maybe also too much of hours put in to the projects already started... Conclusion from my behalf after this is: it's better to have more choises than common controller... It brings diversity which I think, at the end is only good - more choises for different needs.

So lets forget this sunday morning stupid idea. Many Thanks guys for your insight and replys. I hope that some of the talks we've had with eachother and maybe someone else for this topic has been more positive than negative, and I hope we learned something new from this conversation. Thanks Xymph, Oliver, Undef, W1lla, ManiaDesign, Guerro, Domino and Toffe and Tgyoshi at #development channel!

My best wishes goes to everyone at server plugin community!
Good luck for the projects that is being worked on,
- Reaby