Dev meeting minutes 2010-10-03

[KDR_11k]

But you've made it personal again, and you're being deliberately hurtful, and you can't point to anything I've said here or during the last month as provocation.

How about complaining that synced is synced between systems while unsynced is unsynced between systems?

[smoth]

Please do not defend me, I handled it in pm(I think) better that this thread actually be productive.

[lurker]

Zerver, what's the benefit of a single lua state version at all?

No benefit other than full backwards compatibility, and less rotten eggs thrown in my direction.

How so? The only way they see each other is through SYNCED, which is not hard to implement between two states.






I should clarify, I was talking about having a copy of the gadget handler file itself. Is it actually uncommon? Good.

[smoth]

seems that way. I could open all the mods I have but I do not see it in many non-ca based things. Evo and cursed probably needlessly include it.

[Argh]

How about complaining that synced is synced between systems while unsynced is unsynced between systems?

That isn't personal

That's just me exploring what might be possible and why.

It's OK to ask about whether we can change stuff. It was a productive conversation, and the reasoning behind, "no" was convincing.

I explained why I wanted to explore it (my interest in GPGPU is documented, as are working examples of my experiments), I listened to the arguments, and I agreed with the conclusions. And I was polite even when people got hot and bothered.

If people don't ask these things, nobody re-examines stuff, and we can't make progress. I don't mind being called an idiot for bringing this stuff up terribly much. I think it's healthy to discuss the status quo and see if there are any stones left unturned.

I've opened up a lot of conversation here. We've all learned things, decisions seem to be getting made, and it appears to me that zerver and jK are closer to burying the hatchet and figuring out how to make MT actually work.

That's productive and healthy, and if I have to put up with people who don't understand the real purpose of what I'm doing, that's OK. I'd rather be told that I've asked for the impossible because people have actually thought it through than to not get people thinking again and assuming that the conventional wisdom is all that's possible. It's not a fun role to play, but it's a necessary one.

[KDR_11k]

How about complaining that synced is synced between systems while unsynced is unsynced between systems?

That isn't personal

It's personal for anybody with a clue.

[zerver]

How so? The only way they see each other is through SYNCED, which is not hard to implement between two states.

OK, and how would this data access be synchronized?

[lurker]

OK, and how would this data access be synchronized?

With a mutex on the Synced state, much like the mutex used in the single-state method. I'm not saying it's going to multithread well to have direct communication between two states, I'm just saying that you might as well split them and keep the differences between versions smaller.

[Neddie]

1. I'd talk to all the people maintaining currently-available games, and help them help themselves in various ways- things like providing a solid descriptive page on the Wiki, having a press kit, providing us with effective banner representations, etc.

2. I'd do various things with the front page to improve its effectiveness and impact. I don't think I'll need to change anything structurally or change the frame.

3. Once we have things in place and people have been energized a bit, I'll outline some further things we can do that should help promote projects and the engine itself.

4. At this time, I'm not going to bother worrying about the overall maturity of the platform or the various wrinkles AF's brought up such a lack of a good stock UI. I suspect these are problems that will be solved when people feel motivated to solve them.

This is familiar. We worked for this in 2007/2008. I'm willing to work with you on it again, but people were hostile at that point and are unlikely to be less so now - you are forewarned.

If you'd like Neddie to do the job, fine. I honestly don't think that's going to work out, like it didn't work previously, but who knows, maybe he's learned some things. I'm willing to give him a month and see if he can get things done.

Just to clear up any ambiguity, I offered my uncompensated help with P.U.R.E. when I wasn't doing anything else, there was no follow up, and the point I was adamant on you didn't follow through with - working builds for various OS as advertised on release. I still regret laying groundwork for unused marketing support when I could have - as it turns out - helped much more simply by testing.

Side note, you are the same xenoargh from Mount & Blade? I recently got into Warbands and I noticed that mod for the original, but after conferring with Nemo - who also has some interest in the game, we decided it must be a different Argh from deviation in syntax. Of course, your voice has probably changed over the years. I've taken a break from the game in annoyance, as there really should be a cooperative campaign mode, the static settlement model isn't ideal, etcetera.

~

As for promotion now, it isn't a one person "job". It never was. I had results, we've all had results by doing this or that. Hell, I shoved a new player into Gundam two weeks back just by taking the time to talk to somebody. Making a difference is easy and it is something you're all qualified to do.

There are a lot of different skill sets which are useful in promotion. Video editing, audio mixing, image production, writing, survey research, focus groups, copy distribution, etcetera. People tend to discount the skill sets they do not excel in, primarily because their experiences are limited to the content they've provided. What I can do in the realm of advertising is often denigrated, but how many indie games are sabotaged by broken English or a complete lack of market research? Trust me, I understand the subject, I have practical experience in combination with formal education.

Last, the situation isn't really governed by committee. After all, what is there to govern? Nobody wants to contribute material. There are a few people who can weigh in but generally one person approves and everybody else sits back. If you have content to contribute, contribute it. Go ahead. Nobody is going to stop anything useful.

[zerver]

OK, and how would this data access be synchronized?

With a mutex on the Synced state, much like the mutex used in the single-state method. I'm not saying it's going to multithread well to have direct communication between two states, I'm just saying that you might as well split them and keep the differences between versions smaller.

Sorry, not sure I understand 100% what you mean anymore. Do you mean instead of a single state version, I should always split the synced lua and require _G.EXPORT to be used? In other words, throw away the backwards compatibility mode. From what I understood so far, other devs don't allow me to make such a change. I may be wrong though.

[lurker]

I'm saying that backwards compatibility mode should still have two states. It would make SYNCED in one state look at _G in the other state.

[Argh]

This is familiar. We worked for this in 2007/2008. I'm willing to work with you on it again, but people were hostile at that point and are unlikely to be less so now - you are forewarned.

Fair enough. Personally, I'd be fine, if you've got the ability to do stuff and we can execute. Last time, all we could was talk, and it didn't work out well. I think that there's less real hostility, though; we all know that what I said would happen is exactly what did; I think we all agree that change would be good, and I think that once people see something done well, that trust could be built.

Just to clear up any ambiguity, I offered my uncompensated help with P.U.R.E. when I wasn't doing anything else, there was no follow up, and the point I was adamant on you didn't follow through with - working builds for various OS as advertised on release. I still regret laying groundwork for unused marketing support when I could have - as it turns out - helped much more simply by testing.

I am very sorry that I was too drained and screwed-up by the launch to do anything with your generous offer, Neddie. Someday, I will have to tell the real story, but not here and not today.

Side note, you are the same xenoargh from Mount & Blade

That's me. xenoargh's the original handle, actually, I've had it since the dawn of ICQ, for ancient reasons. That darn mod was supposed to just be fun, but it rapidly mushroomed a lot more than I really wanted it to, and I'm sure a lot of people will be upset when I say I'm done with it. It doesn't sound like your cup o' tea, though- it's a mod about combat modeling and other stuff related to the hacking and slashing, and it features girls in crazy armor and stuff. Purely a fun project, but a fun experience and a welcome relief from Spring.

There are a lot of different skill sets which are useful in promotion. Video editing, audio mixing, image production, writing, survey research, focus groups, copy distribution, etcetera. People tend to discount the skill sets they do not excel in, primarily because their experiences are limited to the content they've provided. What I can do in the realm of advertising is often denigrated, but how many indie games are sabotaged by broken English or a complete lack of market research? Trust me, I understand the subject, I have practical experience in combination with formal education.

True enough- well-written ad copy or (more importantly, imo) press materials can contribute a lot. I'm pretty good at snappy phrasing, but I am fairly horrible at longer pieces.

Anyhow, yes, I agree with all of that. One of the biggest problems with P.U.R.E.'s marketing was that I was trying to wear all those hats at once and be the developer- it's just about impossible

Last, the situation isn't really governed by committee. After all, what is there to govern? Nobody wants to contribute material.

I wanted to get the issue of people constantly blocking stuff from happening on the front page so that I could do the things that would need to happen and get people to understand that yes, this time stuff will be done. The concept of "do stuff and show it here and maybe we'll trust you" is fucked; what will happen is that people will find reasons to complain and torpedo the whole thing before it can get rolling. It's no way to run a railroad. Hence why I've been insisting that that not be a problem. I love Spring, I want to help Spring, but by god, I am never, ever letting Spring's people screw me out of my time again. It's the one part of being here that always makes me angry; I hate having my time wasted.

So I was hoping to get it dealt with, simply if possible, but oh noes, people were wanting to debate first principle crap, like whether it should be "about the engine", etc. all over again, while throwing out aspersions that, from my perspective, are silly.

Meh. Here is the plan I was planning to pursue. It's not scary, it's not dictatorial. But it does require that people be able to act.

In terms of sequence, I was going to talk to S'44 or CA or BA (hopefully without pissing any of them off about preference- I think S'44 needs first mention, but in the end it comes down to who's organized and ready) and get material enough to do a front-page notice, then try to get them to tie in with their ModDB presence, if any, and release a press release talking about the game, focusing on whatever the developers wanted to get across.

Basic points there: getting a front-page kit done, getting a Wiki article done and clean, getting press-kit done, getting the project hooked into ModDB, if it wasn't already there, discussing what ModDB is likely to run as a front-page story, or (ideally) a top-banner project (which is hard to get!).

In terms of buildup and follow-through: this would be the guinea-pig; along the way, I planned to get some basic guidelines written (stuff like, "what you must have in press release" and "how to talk to journalists", etc.- stuff that people may have heard about, but don't understand the important points of) deal with traffic flow (S'44 and CA have a fairly sizable offsite existence, so how do we both serve Spring's overall interest and their narrower ones) and work with the team on the buildup plan, and then begin rollout. Call it a week of discussion, material prep, training, and other pre-rollout work, while the basics would go live on Spring's 'site immediately.

The general plan was to give it anywhere from 2 weeks to a month of play, post-rollout, and really try to work the kinks out, with the guinea-pig getting longer play in exchange for dealing with the inevitable learning curve. Hopefully along the way we'd get a break and media stuff would happen, but I think that everybody needs to understand that that doesn't start happening until it's not entirely vital- by the time the media gets involved, a PR project is properly viral anyhow- media's influence these days is a lot less than you'd expect.

But if the cycle kept going roundabout through the major releases, then it's going to change the picture and make it much more likely, and each cycle will prime the pump, while at the same time lighting up Spring as a destination.

Anyhow, that was the general plan I wanted to follow- contact the game developers, explain what was needed, execute promptly and try to coordinate with them and help them out.

In terms of content, especially with the teams with art crew, I'd let the game developers decide exactly what they wanted to do and how they wanted to present, and help with some guidance on stuff like copy editing and the process of promotion via media channels and general advice on what I've found to be effective. I'd provide help on artwork if it was needed, but basically the needs would be met in-house. The last thing I wanted to do was get into pissing contests about defining their look or whatever- opinion, maybe, but only if I was willing to dive in, the whole idea was going to be largely about letting them take charge of their process and be supportive.

At the end of the process, the game developers would have gone through a whole PR cycle and would (ideally) understand how to execute it, the Spring front page would have been used to actually Do Stuff, and they'd see a positive blip in traffic that could snowball if they kept feeding in minor news. The thing about PR is that unless you have a lot of resources, it's impractical to do big things often, but small stuff can still work nicely, if some care is taken.

Personally, I think that all three projects are properly positioned, but as I said, I would like to try with S'44 for the first go, as I think that's the easiest sale and the biggest potential for new growth. I haven't had a chance to look at their current release yet and see the nuts and bolts, like whether they've gotten some basic go-and-play SP trainer stuff done yet (I presume so, because they were heading that way), but if that's a need, that's something that I can help them with to some degree with some stuff I've done for P.U.R.E., and they've got C.R.A.I.G. and straightforward scripted (not a mission, just a simple Spring script and a launcher, if they don't have anything else yet). But their screens look really good and I think it could be doable.

Anyhow, that is what I intended to try out, more or less (please forgive the muddle, I am very tired) before people started throwing hissy-fits and yelling at me. This is why you simply can't do this stuff if you can't re-arrange the furniture on short notice, though- PR campaigns are a lot of work, and stuff can and will need to be changed along the way.

The first project would inevitably require changes and moving fast, and if that's the one where everybody wants to look for an excuse to jump down my throat... no.

I don't want to get involved if it's going to be half-assed talk-talk again, I've had enough of that from this place for one lifetime. I envy guys like hoijui, who probably doesn't even realize how every time that kind of "reasonable oversight" gets proposed, nothing happens because it means public review and a firefight, and the "oversight" person is inevitably a Dev who runs and hides instead of supporting the people who were moving things along

[zerver]

I'm saying that backwards compatibility mode should still have two states. It would make SYNCED in one state look at _G in the other state.

Ah, so two states, but mutexed like it was a single state. Yeah, that would work.

[Argh]

Sorry, other conversation.

What you guys are saying is that compatibility mode would basically not change things, i.e. in theory it'd be drop and go? What would be the speed advantage to full switch over?

[zerver]

What would be the speed advantage to full switch over?

You can use lots of Lua rendering stuff, without the CPU load going through the roof. To illustrate, if you enable all widgets in BA, and "give all" twice, current spring MT has a horrible 50% load on my system.

Using dual states and SendToUnsynced will make the CPU load drop to 10%. Using the EXPORT table would show a similar performance gain, if the table is kept small.

[Argh]

Wow, that is impressive.

Too bad I've just found out that I've hit the GPU wall at mach 5, lol. Oh well, not your problem.

[SpliFF]

Website
Argh, as others have pointed out the Spring site is for developers yet your focus appears to be on end-user marketing. Your time would be 100% better spent updating the wiki documentation in the areas you have direct experience in and marketing P.U.R.E through the Games page and other channels.

The front page is a jump-off point to other information. It needs to answer these questions:

1.) What is Spring?
2.) What can I do with Spring?
3.) Where can I find X information about Spring?
4.) What's happening with Spring?
5.) Where can I download Spring?
6.) Where can I find games made for Spring?

Does it answer these questions?

1.) Yes. "Spring is a powerful free RTS engine developed for Windows and Linux"
2.) Yes. The videos and screenshots provide a visual reference. Links to About/Learn More provides details.
3.) Yes. Top menu and large buttons for common tasks.
4.) Yes. The Latest News section.
5.) Yes. "Download and play" button
6.) Yes. Huge "See the games" button

So the front page serves its purpose. I'm not convinced you could do anything but fuck it up. That's not a personal attack, it's just an observation based on somebody wanting to fix something that isn't actually broken. People are naturally suspicious there's another agenda involved.

The only thing I'd like to see happen on the front page is support for HTML5 video with h264 and WebM. It's not essential, I'm just a big fan of open standards.

GPGPU
GPU computing is great but any suggestion of tying it to synced events should be rightfully laughed at. I don't see that ever happening since it wouldn't work with headless spring and would likely desync across GPU architecture. That pretty much means the only benefit you'll get from GPGPU is for unsynced AI and graphics. The second is already possible through lua shader support and the first would require skills 99.99% of developers don't have.

[Johannes]

Download page is still a mess though

[Argh]

I disagree with your main points, SpliFF. Meh, it's an ancient argument that predates you by 3 years, but I'll repeat it again.

1. Spring's just an empty shell without games to show off the capabilities.

2. Growth would be healthy. Community is small. Marketing is about growth. Games attract both end-users and would-be game developers. This is not a top-tier engine, it's devolving into hobbyist category. Given that this is also a very difficult engine to develop on, it is pretty silly not to try and find growth.

3. No game developer likes building things without an audience. BA's monopoly and engine developers' focus on MP means it is harder than for most engines. If you guys really want the DIY model to succeed, then do some work on SP functionality, quit making excuses and then blaming the game developers, most of whom simply do not have the time resources to do proper marketing, even if they know how.

4. This isn't "The Spring Developer's Project". I know I'm barking into the wind here, but the founding idea was that it was supposed to be something we're all part of. The actual result is people with the dev tag who are largely unsympathetic to their primary clients- the game developers- or are mainly doing dev for their pet projects.

If it wasn't for us, the engine could have been feature-frozen ages ago. A lot of the ideas have come from us. Yet we're second-class citizens in this dystopia, instead of partners who you want to succeed because it means you succeed. It's unhealthy, and it's one of the reasons why things have gotten messed up.

In short, welcome to the philosophical debate. I never win it, because you guys have the power and I just have the arguments.

[SpliFF]

You'll never win that argument because it flies in the face of all human experience. People do not work without reward and the rewards in any non-commercial project are entirely personal. That's why devs will always support their pet projects and interests. There's absolutely nothing wrong with that.

There is no monopoly here. If you feel excluded from the website (ignoring for a moment it's 90% wiki) then by all means start your own Spring community site (Like SpringInfo, jobjol, darkstars, etc...). If you don't like the way Spring is developed then fork it.

You need Spring more than Spring needs you. You'll never get anywhere with your attitude.

You spend too much time making suggestions on topics you don't understand instead of taking the time to understand them. Instead of asking questions and then disputing the answers like you so often do you would gain more by doing independent research. If you really want to make a difference here it's time you get out of your comfort zone and start learning C++ and the Spring source. In the long run you'll be less frustrated and less frustrating to others. The wiki will always lag behind the source code so understanding the source is critical to being productive instead of banging your head against walls.

That's why I went as far as learning C++ from SCRATCH to get what I want out of Spring. Until I started this project I'd never written so much as a "Hello World" in a non-interpreted language. I'm not even very good at maths. I'll never be as good as the other devs but I can at least understand enough to develop the parts that affect me (such as the model loader) and work out what's happening behind the scenes when my game doesn't behave as expected.

Yes it would be nice if everything "just worked" in Lua and had adequate documentation and was everything to everyone but that isn't going to happen ever. So you do a reality check and take the necessary steps to help yourself. Along the way you develop skills that will better place you in the games industry (you'll never get a job in games development if you don't know some C++).