Forged Alliance Forever

Re: Future of FAF (Split from derailed topic)

2014-06-20T23:10:13+02:00

The best way forward (at a high level) that I can see for this is:
We improve communications to the community on what Zep's vision for FAF is - immediate and longer term.
Community discuses the above and provides feedback.
A roadmap is agreed based on the above (Zep has final say)

I think everyone should remember what Zep has contributed towards FAF and stop shooting their mouths off. I know the internet breeds contempt and taht people say things on-line, that they would never say to someone's face. However, Zep deserves our respect and sometimes (as a community) we should just try something new - even if we're sceptical - and see how it pans out. Then provide feedback.

I fear that FAF is dead (via stagnation), medium term if we don't stop pissing off THE MAN (not the dictator, or UberLord, or Nazi) - the man, who made this all possible.

Less drama more sense.

Lets prove to Zep that we're a community worth the effort that he has put in.

I have some concerns/queries about the new MM and how it will work. But I have every confidence that any issues will be sorted and it could be a great enhancement and be an evolutionary step for FAF.

gl hf

Statistics: Posted by Kof — 20 Jun 2014, 23:10

Re: Future of FAF (Split from derailed topic)

2014-06-19T23:26:28+02:00

This is my concern too. The way things have progressed the past 3 months, fixes are being made, but it remains uncertain they will go live soon, if ever. PilOt has said he wants to keep bugfixes in one go to ease integration and effort on his part, which is fair enough, but I know that if I were in charge I'd want to integrate things one at a time to make certain of stability and make bugfixing easier, and also because for the smaller changes there's really no reason not to improve the game ASAP.

Statistics: Posted by IceDreamer — 19 Jun 2014, 23:26

Re: Future of FAF (Split from derailed topic)

2014-06-19T22:12:39+02:00

The first thing we have to know is whether Pilot will let anyone continue development of FAF without his direct involvement. Someone would have to get full access to everything FAF related (and accept the responsibilities that come with it). If Pilot is ok with that we can think about setting up a team and a process that works for everyone (@RK: yes, IMO that means good testing aswell). If no... then at least our time is not wasted doing further discussions.

Statistics: Posted by Brute51 — 19 Jun 2014, 22:12

Re: Future of FAF (Split from derailed topic)

2014-06-19T20:04:29+02:00

I split the topics; this is new topic about FAF and can be used as such .

Statistics: Posted by Gorton — 19 Jun 2014, 20:04

Re: Recent server outages?

2014-06-19T20:03:59+02:00

RK4000 wrote:
I never said anything about me wanting to actually integrate them

Neither did I claim you did

Statistics: Posted by E8400-CV — 19 Jun 2014, 20:03

Re: Recent server outages?

2014-06-19T20:00:17+02:00

A mod should split the topic and close this one.

The OP was answered already.

Statistics: Posted by Sheeo — 19 Jun 2014, 20:00

Re: Recent server outages?

2014-06-19T19:32:33+02:00

RK4000 wrote:
But let's get back on topic here please. This thread is not about the explosions mod, but the process with which we should test (and maybe integrate) new things.

That also is not the original topic.

Statistics: Posted by ax0lotl — 19 Jun 2014, 19:32

Re: Recent server outages?

2014-06-19T16:41:42+02:00

I never said anything about me wanting to actually integrate them (it is originally not even my idea).

I'm just talking about the technical aspects and benefits of integrating mods for the ease their development and making it easier to understand by giving an example of my own mod (issues with destructive hooks, difficulty to test due to poor feedback regarding testing with mods, etc).

And as far as it being optional goes, it is impossible to make this optional, either everyone uses it, or nobody does, otherwise it will desync. This is stated in the FAQ in the mod's thread.

But let's get back on topic here please. This thread is not about the explosions mod, but the process with which we should test (and maybe integrate) new things.

Statistics: Posted by RK4000 — 19 Jun 2014, 16:41

Re: Recent server outages?

2014-06-19T16:35:02+02:00

Not everyone likes overblown graphics, so by all means: don't integrate it, and if you do, make it optional. Like an option in the video settings.

But none of this has anything to do with the original topic of this thread.

Statistics: Posted by E8400-CV — 19 Jun 2014, 16:35

Re: Recent server outages?

2014-06-19T16:39:55+02:00

As far as the topic on testing a mod before it is integrated goes, I believe that if a mod is proven to work 100%*, causes no bad after-effects, and most people agree on it's integration, it should go in. There is no logical explanation of why it shouldn't at that point. And if it's integrated, it also makes it easier to improve, code, debug, and improves compatibility.

*However, the testing needs to be extremely precise. And I mean EXTREMELY precise. You don't want to integrate with any bugs, no matter how tiny. And, of course, adding only one thing at a time certainly helps. That's why we need to take it slow with integrating things. Let's say two mods that fix things, and are planned to be integrated work with the standard game individually without problem. But if one of them was integrated, would the other one still work (if integrated)? If yes, integrate both aswell. If not? Then work more on debugging and fixing compatibility.

Then, on a little off-topic:

pip wrote:
rootbeer23 wrote:the process exists already. take for example rk's explosion mod. it has been played with numerous times.
i have never seen it misbehave. some might have. if it was to be made default, it would be a smooth transition.

No, because Rk4s explosions breaks one the featured mods (Nomads), so it would actually be rejected until compatibility issue is fixed. There is a reason why Pilot insists that all featured mods and coop missions are tested when new code is added : it's to make sure the FAF environment is coherent and fully functional.

Not strictly true. If it was integrated into FAF, not in mod format, I could easily fix compatibility with Nomads with as little as one or two lines. However, in mod format it breaks it because of a destructive hook which I am FORCED to do, due to the way loading hooks works. (Ask Sheeo about it) Because of this I cannot make it compatible with nomads as long as it remains a mod.

Compatibility with anything that doesn't use the code that is hooked destructively is guaranteed though, based on the feedback I have gotten, it's pretty much compatible with everything.

However, I very much want to see nomads using factionally-specific explosions aswell, but I am just not willing to do it since:

(EDIT: 0. The main designer behind the Nomads might already have a style of explosion planned for them, so I do not want to interfere with that (I know I wouldn't like that))

1. Barely anyone plays with Nomads

2. They're not integrated yet (and probably won't be for a while)

3. It would be a lot of work for little gain, because of reason #1. Making a whole factions' worth of factional explosions is by no means an easy task, and if barely anyone plays with them, what's the point of devoting so many hours into effects barely anyone will ever see. This mod is at most a 2 man job (me and Sheeo) as far as coding goes and a one man job as far as effects go (only me for effects), so it takes time to get from scratch to good (or better than standard) looking effects.

4. Making it bug-free would also take a while. The explosions mod was pretty buggy its first week, and it took me quite a while to get it to the stability it's at now (which, if I may say, is pretty much rock solid, even moreso with V7). Adding stuff for more factions would definitely compromise it, or at least make it take much longer to finish, because of the testing and debugging.

Statistics: Posted by RK4000 — 19 Jun 2014, 16:01

Re: Recent server outages?

2014-06-17T21:51:02+02:00

rootbeer23 wrote:
what i mean is this: there is this shieldfix mod, i have used it once or twice. what i want is for somebody to come forward
and say that development is finished and the mod is working; use it because it fixes issue X, and then i will use it regularly.
after a while, if it behaves nicely, you might as well make it default.

It is however not that simple.

Once you integrate with faf you don't want to keep tacking on the features and fixes as hooks--you want them integrated.

That code change requires testing, and preferably review.

Statistics: Posted by Sheeo — 17 Jun 2014, 21:51

Re: Recent server outages?

2014-06-17T21:43:39+02:00

what i mean is this: there is this shieldfix mod, i have used it once or twice. what i want is for somebody to come forward
and say that development is finished and the mod is working; use it because it fixes issue X, and then i will use it regularly.
after a while, if it behaves nicely, you might as well make it default.

Statistics: Posted by rootbeer23 — 17 Jun 2014, 21:43

Re: Recent server outages?

2014-06-17T21:39:45+02:00

what i meant to say is, if there is a mod and it doesnt get bugreports for a while and it is used regularly, then you have all your testing done. like rk's explosion mod if it worked with nomads like it works with standard.
gazui and hotbuild were incorporated years after everyone used them.
no need to make it so scientific.

Statistics: Posted by rootbeer23 — 17 Jun 2014, 21:39

Re: Recent server outages?

2014-06-17T21:40:49+02:00

Brute51 wrote:
What Pip mentions is a process that I've been discussing with Pip a few times now. We use it at the company I work at, and with great results. It ensures only quality code is added to the project (which in our case is vital - we're engineering automated robots for continuous production on the global market).
I think we could implement the same process here but it does require a team of a few active and dependable people (developers (implentation and code reviewing), testers (to test every aspect of a code change), a gate keeper (a senior developer who guards the trunk) and possibly a specifier (someone who specifies changes in detail)). We'll also have to develop automated testing because no one likes to do manual regression tests. I"m not sure about how to do "segment testing" here but we could explore that too.
(on FAF I think "segment testing" is a better word for "unit testing").

I can write a document about how the process works and share it, see what the reactions are...

@Pilot: Are you interested in keeping the server running? I can imagine you're not, if that's the case please let us (the community) know so we can find alternatives.

p.s. In case you're wondering, I work as software tester.

Thumbs up, and I'm willing to be a part of the team that does this for patches. Although I don't think it needs to be that rigorous.. We are just a gaming community.

I don't believe we can easily do automated testing, however, unless someone figures out how to run the sim in headless mode (which should be possible).

Statistics: Posted by Sheeo — 17 Jun 2014, 21:38

Re: Recent server outages?

2014-06-17T21:29:11+02:00

What Pip mentions is a process that I've been discussing with Pip a few times now. We use it at the company I work at, and with great results. It ensures only quality code is added to the project (which in our case is vital - we're engineering automated robots for continuous production on the global market).
I think we could implement the same process here but it does require a team of a few active and dependable people (developers (implentation and code reviewing), testers (to test every aspect of a code change), a gate keeper (a senior developer who guards the trunk) and possibly a specifier (someone who specifies changes in detail)). We'll also have to develop automated testing because no one likes to do manual regression tests. I"m not sure about how to do "segment testing" here but we could explore that too.
(on FAF I think "segment testing" is a better word for "unit testing").

I can write a document about how the process works and share it, see what the reactions are...

@Pilot: Are you interested in keeping the server running? I can imagine you're not, if that's the case please let us (the community) know so we can find alternatives.

p.s. In case you're wondering, I work as software tester.

Statistics: Posted by Brute51 — 17 Jun 2014, 21:29