Spring developer crisis

[knorke]

I think before deciding to contribute anything, developers will want to learn about Spring by playing some games.
Here they face the same problems as normal players: "everything is complicated, nothing just works"
(lobby, game UI, the whole infrastructure)
It is easy to come to the conclusion that the project is not mature enough to be worths ones time. (players & devs)
Maybe the playable content is also not attractive enough to convience people to contribute?
All moddevs are former players (at least in OTA) maybe engine devs can be recruited same way.

On moddev side there are often plain stupid postings and questions, if potential new devs read the forum I can see how that can be repulsive.

The AI interface thing was already mendtioned..."the tutorial is outdated, and the original sources are... kind of lost?" does not show Spring in good light.

Are there other ways we can try to bring in new blood? Other avenues to advertise (like moddb)?

imo moddb is wrong site for spring (engines in general). Too many players, no developers.
The last news about engine is "Spring 0.82.7 bug fix release" and top news is some canceled $100 tourney. That is just silly.
In previous thread it was said that is the fault of engine devs/other content devs for not publishing news too, but obviously nobody has time for that. So just having a static profile page would look much more mature.

http://www.flipcode.com/archives/05-04-2005_rt.shtml
That was good display of spring to the outside. It hits the right audience and iirc comments were positive. (comments section is now down)
And unlike moddb, it is not the kind of site that players will spam with "game always crashes" comments.
I dont know if similiar pages exist?

[FLOZi]

I do worry from time to time but cleanrock is new and BrainDamage and Beherith have both contributed important changes.

an other big reason is the current bad state of the AI Interface(s), which is mostly my fault.
and as mentioned, of course, the not so nice treatment of new devs, that could potentially be engine devs.

Both are issues - especially as I agree that AI devs have been a good source of engine devs, however;

Another point is that Spring as it stands today is more or less "done". There’s not much low hanging fruit to take on that will have a big impact on the project.

This is the real issue, IMO. GSoC may have failed (tbh I never really felt we pushed it that hard) but the idea of having possible projects to do (as an intro to the engine, and as a guide to what actually is still to be done) is a good one.

one main reason is the uglyness of the general engine design (OO wise), and the inflexibility/unwillingness (in my eyes) of the current devs to change this. it is not possible to work on one part of the engine in a nice way, because of that.

This is a chicken-and-egg problem, you need established devs (and impossible amounts of time) who are familiar with the codebase to design and implement such a radical restructure. I think the incremental steps in this direction are realistically the best outcome we can hope for. I somewhat disagree with the conceit that 'it is not possible to work on one part of the engine' due to it, c.f. the implementation of ROAM and Assimp.

There are far better places to 'advertise' for engine devs than moddb - gamedev.net is an obvious one.

[PicassoCT]

Another point is that Spring as it stands today is more or less "done". There’s not much low hanging fruit to take on that will have a big impact on the project.

Violently disagree. There is stuff that could be added. And i think the bugfixxing goes easier if you have some share in the house on which you do maintainance duty.

[FLOZi]

Go on then - name it. Something that is

a. Easy to implement (particularly for someone new to the engine code)
b. Will have a large impact/benefit to the engine/game devs
c. Isn't done more easily or sensibly through lua

[malric]

a. Easy to implement (particularly for someone new to the engine code)

Why do you consider this a requirement? Do all of the previous devs started with something easy?

b. Will have a large impact/benefit to the engine/game devs

Looking at the gsoc page, I would say that limited play area, in-game lobby and host change during game will have an impact. Not to mention all the discussion about the AI interface on the other thread.

A page (like the gsoc page) listing all the feasible things to be implemented, could probably motivate new people to work. In my opinion it should be more advertised (like a top level link from Development, rather than a link from a subsection of Getting_Started).

[Neddie]

Perhaps I am wrong, but for years Spring has seemed base feature complete. A lot of the work that can be done is in the realm of radical expansion, optimization, and incremental feature improvement - this is not to say it isn't important or interesting, but that it is not low hanging fruit.

[FLOZi]

a. Easy to implement (particularly for someone new to the engine code)

Why do you consider this a requirement? Do all of the previous devs started with something easy?

http://en.wiktionary.org/wiki/low-hanging_fruit

In game play area can already be lua'd, with some caveats. In game lobby is also now in the lua realm. Host change is not low-hanging fruit.

[malric]

Host change is not low-hanging fruit.

Many people work on big projects. And you do not always have a "low-hanging fruit". It can be frustrating at first, but that is part of the challenge. Depends on your motivation. The great guys that started the project I bet they did not think about how easy it will be...

[Tim Blokdijk]

Another point is that Spring as it stands today is more or less "done". There’s not much low hanging fruit to take on that will have a big impact on the project.

Violently disagree. There is stuff that could be added. And i think the bugfixxing goes easier if you have some share in the house on which you do maintainance duty.

Ok, I could be wrong. Regardless of who is right, what I would like to see is some site infrastructure where developers can write down detailed design proposals with costs in time and money. Which can then be vetted by other developers and then publicly funded by donations and/or some other business model.
I know Licho is working on something with a more limited scope.

Am unsolved challenge is the "who is a developer" question. We would need some type of web of trust where existing devs vet another dev. Does anyone know of a existing system (software or theory) that can do web of trust stuff? A way to reliably and fairly discriminate developers (code&content) from the playerbase.

[hoijui]

i never suggested that better design can not be done in many small incremental steps. that is how i would have done it aswell. but it is not being done, and it could be done. it is also easy to do, it is just boring work. it is also against the will of jk and kloot, which is the real issue.
of course they are not against "nicer design", it's just that .. their idea of it is not the one of a software designer, and that for them, it has a very low priority. in my eyes, this is a problem, and it can not be overcome by new devs or payed devs.
ROAM and ASSIMP were added. that does not mean that they were working only on the relevant parts of the code. it means that it is possible to write code for the engine, even if you are not a long term engine dev. that is good of course, but it is a veeeery low measure of code design quality.

[Tim Blokdijk]

Having a me/you/us/them would derail this topic. Don't do this here now. I know you're frustrated but focus on solutions.

Having detailed technical designs written down with a requirement to get some of your peers to comment on it (vet it) before funding can be secured would help with getting better communication and function as a roadmap.

[Cheesecan]

I have played spring games since 2005, and learned some coding since then (enough that I can work as a programmer today). But coding on the engine never really occurred to me. Working on these kind of projects is a huge investment in time for very little reward, depending on what kind of person you are and what drives you to do things..

[Anarchid]

In game lobby is also now in the lua realm. Host change is not low-hanging fruit.

This strikes me as probably inaccurate. I've heard things about how luasockets work in spring, and those things weren't nice.

[FLOZi]

In game lobby is also now in the lua realm. Host change is not low-hanging fruit.

This strikes me as probably inaccurate. I've heard things about how luasockets work in spring, and those things weren't nice.

A realm is a pretty far reaching territory. In game lobby will never be implemented via anything other than the lua API, though of course the API itself will grow.

[smoth]

Spliff. You are dismissing cleanrock's work.

[SpliFF]

Just going by the commits in the last 4 months he and BrainDamage have only a few. The devs I singled out have a couple of commits or more each week in 2013. My definition of "active developers" is just more limited than yours. Obviously I mean no offense to anyone who has ever committed code to the engine (myself included) but it's the current "day-to-day" development that really keeps things afloat.

Also to be clear, cleanrock appears to be focused on flobby which is what I would call a tool, not the Spring engine itself. My concerns revolve around what I would consider the "core" or engine internals which is why I keep asking people not to derail with AI and game development complaints.

It is my view they while more is better when it comes to games, tools, lobbies, downloaders, maps and AI they are not things a new game *needs* (there are many existing working resources available). What a new game does need is a working engine that is compatible with current OS and hardware and handles essential things like pathfinding, targeting, graphics, sound, network and unit animation. Without the engine all other tools and resources are worthless. A building built on crumbling foundations is not one I would choose to live in, even if I had my choice of wallpaper and carpets.

I'm not that interested in whether people share this assessment but if anyone wants to discuss the pros and cons of different subsystems or contributors let's do it in a new thread. I want the focus of this one to be the lack of active "core" devs.

[PicassoCT]

Is there a way for people to volunteer for bugfixxing-aprenticeship? I mean, i report one, i get a cpp or h link to start the search, find it and show my fix to you, and get either approval or hit over the head?

If i fix 100 bugs i get a wish ...


:D

[abma]

@SpliFF:

did you compare 0ad with springrts?

they had much more attention in the news but imo they have much fewer devs because game devs commit into the repo, too.

hoijui is right, code could/should be made more modular but thats not an easy task. hardest part is imo rendering/sim, stuff like networking / filesystem should be easier to get modular.

maybe it wasn't done, because "more important" stuff was changed, not sure.


its hard to attract new developers, but sometimes very good pull requests just appear.

imo new engine devs need:
- a problem they can solve
- a flat learning curve
- easy to set up development environment
- a stable (working) codebase
- some sort of support

also, game devs are really important for spring, they offload many work from engine devs as they fix bugs / make bug reports / keeps gamers playing.

[smoth]

Cleanrock did some fixes to the engine. I think you are using a narrow net.

[gajop]

Abma points out it nicely. New devs are really unlikely to start doing mass commits.
What's more likely is that they're going to either find a feature they care about and try putting it in (which is probably going to be one big commit in the end), or fixing a bug they care about, which is also probably going to be a single commit.

And that's fine, we have the GSOC page already (probably should be renamed), so we can steal the ideas from there.