Commandline poll

[Crymson]

How about a poll on if people like a blue sky?

[Olrox]

How about a poll on if people like a blue sky?

That's a double-post, and it's... hmm... nevermind

~Olrox

[stiv]

But I'm still intrigued, because I'm still figuring out how the folder thing is related to launch WZ

The root of the discussion is what we need to implement a working and usable system for managing mods. Right now we have two mechanisms: the command line switches and the autoload folder. One is complicated for anyone who clicks rather than types, and the other can cause some difficult to understand problems.

A complete solution goes beyond an in-game interface or cross-platform launcher. We also need a mechanism for saved games to remember and load the mods they were running under as well as a way for people joining multiplayer games to all get on the right page.

[Olrox]

The root of the discussion is what we need to implement a working and usable system for managing mods. Right now we have two mechanisms: the command line switches and the autoload folder. One is complicated for anyone who clicks rather than types, and the other can cause some difficult to understand problems.

A complete solution goes beyond an in-game interface or cross-platform launcher. We also need a mechanism for saved games to remember and load the mods they were running under as well as a way for people joining multiplayer games to all get on the right page.

Well, the savegame format will already be changed due to a series of reasons, so implementing a "mod name" line to it shouldn't be hard, I guess - possible problems with people changing between coincidently identically named mods and trying to load a savegame from the previous mod (assuming that someone happens to name a mod identically to another one and publish it ) would be really negligible. As I understand it, the problem is making the game load the mods during execution rather than being launched with the mods, but then, this is already planned, hence my confusion.

But then, if suspending support for the smart folder system during alpha testing is necessary, I think that you should do that - if it is necessary to implement a better system after that, befor it reaches beta phase, that is

Or at least this is what I think, with all my ignorance about the work that this represents - The decision and greater knowledge over the facts is all yours of course! The lesser folly of commiting occasional gaffes in the proccess of trying to help is my job

Someone once said that "What really matters is the intention", but that's a hypocritical lie IMO, what really matters the most is the result
But lo, we'll have no good results in the end if we have no good intentions.

OK, enough of Fabian Phylosophy for now

~Olrox

[Zarel]

The root of the discussion is what we need to implement a working and usable system for managing mods. Right now we have two mechanisms: the command line switches and the autoload folder. One is complicated for anyone who clicks rather than types, and the other can cause some difficult to understand problems.

...not anymore, now that those problems have been fixed.

A complete solution goes beyond an in-game interface or cross-platform launcher. We also need a mechanism for saved games to remember and load the mods they were running under as well as a way for people joining multiplayer games to all get on the right page.

Well, then, good thing we already have "a mechanism for saved games to remember and load the mods they were running under as well as a way for people joining multiplayer games to all get on the right page", isn't it?

Well, the savegame format will already be changed due to a series of reasons, so implementing a "mod name" line to it shouldn't be hard, I guess - possible problems with people changing between coincidently identically named mods and trying to load a savegame from the previous mod (assuming that someone happens to name a mod identically to another one and publish it ) would be really negligible. As I understand it, the problem is making the game load the mods during execution rather than being launched with the mods, but then, this is already planned, hence my confusion.

Not only is it planned, but it's already done.

But then, if suspending support for the smart folder system during alpha testing is necessary, I think that you should do that - if it is necessary to implement a better system after that, befor it reaches beta phase, that is

The problem isn't that, it's that some people want to remove the folder system for 2.3.0.

Anyway, it's not really up for forum discussion at this point, so this conversation is moot.

[Crymson]

How many times has the savegame format changed anyway?
It is being broken on all the beta releases, like the netcode right?

[Crymson]

Anyway, it's not really up for forum discussion at this point, so this conversation is moot.

What does that mean?
When isn't this stuff up for discussion?

[Olrox]

Anyway, it's not really up for forum discussion at this point, so this conversation is moot.

What does that mean?
When isn't this stuff up for discussion?

Because this is stuff for the devs to discuss, I think, not us regular members. It's not anything personal, certainly

~Olrox

[stiv]

I don't see any reason not to be open about what is being discussed. We are not talking about state secrets. We are talking about an Open Source software project.

[Olrox]

I don't see any reason not to be open about what is being discussed. We are not talking about state secrets. We are talking about an Open Source software project.

Yeah, that's right - but still, important decisions should be made by the staff, otherwise thing will never be done (I believe that it's already hard to reach a consensus within the staff, imagine with everyone giving their "2 cents", we would have some mighty big trouble for deciding ).

You were already kind enough to explain what the thing was all about, and I appreciate that very much, but I'm really not proficient enough in coding to really say what I think you should do - therefore, I would take only a small part in any discussion about that, even if I would take any part at all. I think I can make some ridiculous mistakes with my many guesses, and I prefer to avoid that once in a while not to risk too much

But really, polls are a nice way to know what we think without having to hear everybody and read occasional uninformed or reduce potentially off-topic talk. If the staff can come up with the best options and present them to hear which ones the folks prefer, and sometimes, why, then I think it's already a great deal.

All in all, I think that reall there is no reason not to be open about what is being discussed, but I do believe that if you leave the discussion itself opened to everyone, you'll have a hard time with lots of arguments and misunderstandings, much more than if you had previously decided which ones are the best options. I think that is what Zarel meant with "it's not really up for forum discussion at this point", but I may very well be wrong - at least not entirely, for sure.

~Olrox

[Zarel]

How many times has the savegame format changed anyway?
It is being broken on all the beta releases, like the netcode right?

The last time savegames were broken were 2.3 beta 7 for skirmish, and 2.1 beta 4 for campaign.

My change won't break the savegame format at all. Stop worrying.

I don't see any reason not to be open about what is being discussed. We are not talking about state secrets. We are talking about an Open Source software project.

It looks bad if we put it up for discussion, and then completely ignore user feedback. It also looks bad if we take developer arguments public.

I can summarize what happened:

- Buggy proposes removing the autoload feature on the ML.

- stiv and Per agree, I vehemently disagree (pretty much what usually happens when feature removals are proposed).

- Their main reasons for removing the autoload feature is because if a game is saved while a mod is active, and then loaded while the mod isn't active (or vice versa), the game can crash. This ignores the incredibly important point that this is still true if the mod is activated using the command line or an external launcher. The whole reasoning boils down to "Less people know how to use the command line, so we'll get less complaints" which I find a silly reason to remove a useful feature (especially since there were practically no complains either way).

- My reasoning is that Warzone's had an autoload system since 1.00 (notwithstanding its removal between the 2.2 and 2.3 betas). One of the reasons I'm a dev is because of the autoload system in the 2.1 series enabling anyone to install mods, enabling my Rebalance Mod to be popular, enabling the devs to notice and recruit me. Many other commercial games also have an autoload system with the same flaws 1.00-2.1 had, and yet, they thought it was worth it. Anyone who has trouble with the autoload system can simply stop using.

- Anyway, I never win these sorts of arguments, so I give up and just fixed the savegame problem.

- As I understand it, currently stiv and Per want to postpone my fix until 2.3.1, and disable autoload in the meantime (all subsequent 2.3 betas and 2.3.0). I disagree with disabling autoload for the reasons mentioned above. I disagree with postponing the fix because I've never understood the reasoning for refraining from committing code that is guaranteed to fix crashes, because it might possibly introduce bugs. It reminds me of refusing a cure for a fatal disease, because it might cause drowsiness as a side effect.

- Note that no actual bugs have been found in my code; they're just really paranoid that there might be one (which is possible, but I doubt it will be worse than the outright crashing that currently happens, nor will it be difficult to fix).

[Crymson]

How many times has the savegame format changed anyway?
It is being broken on all the beta releases, like the netcode right?

The last time savegames were broken were 2.3 beta 7 for skirmish, and 2.1 beta 4 for campaign.

My change won't break the savegame format at all. Stop worrying.

I don't see any reason not to be open about what is being discussed. We are not talking about state secrets. We are talking about an Open Source software project.

It looks bad if we put it up for discussion, and then completely ignore user feedback. It also looks bad if we take developer arguments public.

I can summarize what happened:

- Buggy proposes removing the autoload feature on the ML.

- stiv and Per agree, I vehemently disagree (pretty much what usually happens when feature removals are proposed).

There are 20 members though. http://sourceforge.net/project/memberli ... _id=142032 What do they say, and why are there only a handful that ever use the forums?

- As I understand it, currently stiv and Per want to postpone my fix until 2.3.1, and disable autoload in the meantime (all subsequent 2.3 betas and 2.3.0). I disagree with disabling autoload for the reasons mentioned above. I disagree with postponing the fix because I've never understood the reasoning for refraining from committing code that is guaranteed to fix crashes, because it might possibly introduce bugs. It reminds me of refusing a cure for a fatal disease, because it might cause drowsiness as a side effect.

- Note that no actual bugs have been found in my code; they're just really paranoid that there might be one (which is possible, but I doubt it will be worse than the outright crashing that currently happens, nor will it be difficult to fix).

I rather have a fix that stops the crashes, but you can't possibly guarantee that it will have no side effects, since you just made the patch!

[Olrox]

I'll dare to say what is my interpretation.

There are 20 members though. http://sourceforge.net/project/memberli ... _id=142032 What do they say, and why are there only a handful that ever use the forums?

Most of the devs communicate through IRC, and see, there's a staff forum. Most of the devs are split between RL stuff and fixing bugs, they don't usually have spare time to post in every topic, so Zarel is the most active member of the staff in the forums, actually he's the most active member in any aspect - hence there's frequently the impression that "Zarel must be speaking for himself", but that's not truth and I must say that, in the beginning I used to think that as well.

I rather have a fix that stops the crashes, but you can't possibly guarantee that it will have no side effects, since you just made the patch!

You can be sure that he doesn't think that, and you must see that he didn't actually write that. Nothing is perfect, and that's as obvious as nothing else, and I guarantee you that Zarel doesn't commit the folly of thinking that he can make perfect things. He just tries to do his best, and I say, I can admire that - and he knows that even though he might try his best, some things will still need to be improved.

That's why the humans came to work as teams... ever since they went bashing clubs in each other's heads and wore beast furs.

What Zarel does is a job of mantaining the forum running, and to some great extent, keep everyone informed about what the staff overall thinks. This is great because we don't usually have to wait more than 1 day to get an official response for a question, and in the overwhelming majority of those cases, he is right. The problem is that, sometimes the staff didn't reach a conclusion to a question, and in this moments, it seems like they aren't working as a single team, which is a false impression. I believe that having this kind of polemical decisions, about which path the project should take, public - it isn't worthy all the misunderstanding it can cause to regular members.

Not to mention the problem that many folks read half of the posts and go regurgitating their incomplete thoughts.

This is the way I see it, and there's no reason to raise tumult because of such a small misunderstanding.

~Olrox

[Zarel]

There are 20 members though. http://sourceforge.net/project/memberli ... _id=142032 What do they say, and why are there only a handful that ever use the forums?

There are 20 entities with commit access. That's not the same as 20 members. Some of them aren't officially staff (DyDo, Delphinio), some of them are bots and not humans (gitsvngateway), and many of them are inactive.

The active members are me, Per, Cybersphinx, dak180, stiv, and Buginator. Safety0ff, Giel, i-NoD, and Cyp are semi-active, but I haven't heard anything from them recently. Of the rest, Buggy and Cybersphinx haven't weighed in (Buggy is away), and the rest usually don't comment on these kinds of things.

I rather have a fix that stops the crashes, but you can't possibly guarantee that it will have no side effects, since you just made the patch!

I can't, but I can say that I personally believe that it is near certain that the benefits will outweigh any possible side effects, and I would go so far as to call any side effects unlikely. While no one can be sure about anything, I do wish to point out that having written the patch, I am probably better suited to evaluating its effects than anyone else.

Most of the devs communicate through IRC, and see, there's a staff forum. Most of the devs are split between RL stuff and fixing bugs, they don't usually have spare time to post in every topic, so Zarel is the most active member of the staff in the forums, actually he's the most active member in any aspect - hence there's frequently the impression that "Zarel must be speaking for himself", but that's not truth and I must say that, in the beginning I used to think that as well.

Most discussion takes place on either the dev IRC channel, the ML, and the tracker (i.e. publicly). The staff forums are practically deserted, and the only thing we keep truly private are discussions of whether or not to give people commit access, or of forum administration, and similar.

[Crymson]

Olrox, but it is causing misunderstanding when one person says one thing, then another one says something else.
Reading viewtopic.php?f=6&t=4706 it shows

In future releases (after 2.3--it has been a learning experience), we are hoping to get more stricter on what we allow in these 'beta' testing periods, and hopefully we won't have a repeat of how 2.3 came about.

That shows that they acknowledge there is a problem but then it looks like people can't agree what should be allowed, and what shouldn't be.
Then when someone can't agree on the ML or IRC or staff forums, then they make the issue more public to get more support in hopes of convincing everyone else to allow this fix, and I do want this fix.