The Betas are slowly getting worse for some reason....

[Imprisoned]

Full versions (ZC 2.10 etc.) : For making and playing quests
Betas : For bugtesting. You can make or play in some of them but expect bugs.
Alphas / Development builds : For bugtesting and reporting only. If you're making a quest don't download these. (Perhaps I should disable saving just to make it clearer. No, of course I won't, but you get the idea)

There's some really good point there... but still, people should use the betas so
we can sort out the bugs faster, and get the stable builds faster too :)

[Petoe]

Very good post Koopa, that's a beautiful idea you got there, but it's just an idea.

I mean, you compare ZC to a house. Well to me and many others ZC is like a house that never gets finished. It has been under construction for such a long time, and even though the blueprints indicate that the house will be beautiful and comfortable, a really appealing house when it is finished, people interested in the house are getting frustrated and lose faith that it will never be a house where you can live in. Why?
Well the designers of the building aren't co-operating and some of them keep adding stuff that isn't necessary but something very cool to bring up the price of the house. After all, it must be the best house ever because so many people want it to have everything a house can have, and even more. It would be a fine and livable house without all that extra stuff, but now it seems like the house just keeps getting more unstable and dangerous even though the builders are trying to feverishly fix all the problems. But it may be impossible to get the house perfectly stable and safe anymore without taking the whole bulding down and re-constructing it after all those additions that made the structure very unsound.

Yeah, I know, what's done is done and there's nothing we can do about it and I really hope ZC hasn't gotten so complex that it is beyond fixable. I really want to believe that all you developers have learned from all these mistakes, and since currently the main focus SEEMS to be on bug busting, I am still a faithful tester because damn, I want to see that stable release and I want to be able to release a quest that has encountered too many problems already and has been delayed way too much.

[koopa]

The second point this thread raised is entirely valid. When I joined, it was just DN and jman and they coordinated it all via chat or AIM. As we grew to a larger team we needed different tools - a SVN server for one, or coordinating would be hopeless - and it's possible that things could be, and may be, planned and coordinated even better in future.

Towards the end of last year the focus was really on putting lots of new stuff in - scripting, custom enemies and items, ice combos and whatnot. Petoe's right, there's more bugfixing going on right now. DarkDragon being around again can also only help there.

As to bugs appearing and disappearing (Magic Shield etc.) ZC has gone a considerable rewrite recently. Central things like the item system or the enemy system had to be changed - and moreover changed bit by bit - but all in all I think when the newer system is in place it'll actually improve the stability, as it's built on a more solid foundation. For instance, some code that was a patchwork of "a bit of this here, a bit of that there" is now much more clean and organized. Some of this "tampering" was necessary to allow custom items and subscreens, and some of it was more "let's redo it thouroughly while we're at it".

Like DarkDragon said, it's an opportunity to fix things that have been lurking for years and now the new features bring them to light. Another point is we don't always get it right the first time either. Sometimes one person changes a line of code to fix one bug, but it causes another one somewhere else. So that line has to go back out and we have to deal with the first bug somewhere else. Sometimes we have to rewrite half the item system to allow some new feature (think true arrows) to work. One of the consequences of seeing the work in progress is that you see all our trials and errors. I hope that goes some way towards explaining how we "put bugs back in". If I change something and it causes a new bug, I'll take it back out (as happened with the map view in ZC). And so the old line and the bug it caused is back in (DD fixed that one finally, though. thanks). I'm not denying that at other times we do put actual bugs back due to lack of coordination, though.

On the point of whether we should all use one compiler, I don't think that's that much of an issue. Except for the recent dll issue it hasn't caused too much trouble - the bugs we're talking about here are bugs in the code and don't have much to do with whether we use gcc or msvc. If we want a ZC that runs on Windows, Linux and Mac OS we pretty much have to use different OS - but as it all boils down to the allegro library in the end, the actual ZC code is pretty much platform- and compiler-independent.

The community has expressed many times that they would by and large prefer something that works and that in a reasonable amount of time over something with many more features than we have now some time in the distant future. I - and the rest of the development team, as far as I can tell - have noticed that and agree with it. I know we promised a stable ZC by last Christmas. Getting annoyed with the delays since and feeling let down is somthing I can understand.

[Limzo]

Well, let's take one of the best examples, the buffer overrun. Suppose ZC is coded to reserve, say, 1000 bytes per screen for combos, but because of a coding oversight, some quests try to save 1100 bytes worth of data. The extra 100 bytes "spill over" and are written somewhere else in memory; this is the buffer overrun. What happens when these 100 bytes are written is completely unpredictable. If you're lucky, it gets written to useless memory and nothing bad happens. If you're unlucky, it overwrites other useful data, such as the next screen's combo, without ZC or the user detecting anything bad has happened. Sometimes the OS can detect a buffer overrun and crash the program, but a lot of the time the corruption happens silently.

Thus, what can happen is that a buffer overrun can lurk undetected in the program for years, silently corrupting data and causing all kinds of weird behavior. Then, all of a sudden, a harmless change elsewhere in the program, such as a fix for another bug, changes the layout of memory slightly, and what use to just silently corrupt data now crashes ZC. It's not that a new bug has been created - the old bug's behavior just changed.

Actually, a bug crashing ZC is sometimes very good news; a buffer overrun silently corrupting data is very, very difficult to track down, and might not be fixed for years (take for instance the combo corruption bug; we've never been able to find that one.) Once it has "mutated" to crash ZC, on the other hand, it becomes much easier to pinpoint the bad code that causes the buffer overrun, meaning it is very likely the bug will be finally and promptly fixed.
Obviously when ZC crashes, it sometimes is genuinely the fault of a recent change, and not of a lurking buffer overrun, but you should keep in mind that more crashes doesn't necessarily mean everything is going to hell - maybe the crash will give us an extra angle of attack allowing us to finally nail elusive bugs.

So, its kinda like this then. (Bare with me, I get a bit farfetched sometimes, so this is a bit unrealistic)
There's a dinosaur that dies, and then all of it's bones are crushed onto a big rock. And this fossil has bits of the Dino's DNA left over on it, so anything that touches the fossil dies, and it stays in the earth for years and years, but noone knows the fossil is there. Then, the people dig down into the earth to get oil, and this rock is thrown out onto the ground. Then, for some bizzare reason, oil is spilled onto the fossil. Suddenly, the DNA reacts with the oil and a huge dinosaur is formed in seconds, and attacks the people.

Does my metaphor make sense, or just ruin your explination completely?

[The_Amaster]

Ummmm.....Hmmm, maybe I'm interperating your metaphor wrong. I was thinking of filling a huge jug of water over a bunch of delicate equipment. If you overfill, chances are that it's going to mess up somthing.

[Pineconn]

Hey Petoe, your PM box is full. I wanted to send you a PM but couldn't.

You know, 2.11b10 was a pretty solid beta. When in doubt, if the following betas become too buggy, we could just scrap them and go back to the drawing board with b10. This would be last-resort only, though.

[Petoe]

You know, 2.11b10 was a pretty solid beta. When in doubt, if the following betas become too buggy, we could just scrap them and go back to the drawing board with b10. This would be last-resort only, though.

Whoa there Pineconn... we? When did you become a developer? O_o

Anyways, b10 sure was good, but so is b15. It's just that both of them have those really annoying bugs. It seems like it will take forever for 2.5 to be stabilized and released but I'd really love to release the group quest Lost Isle before that. The problem is that there's always one serious bug preventing the quest to be released for one of the betas. First there was the bomb crash bug, and when it was fixed there was the subscreen/HCP bug, and now that they're fixed there's this Ring cset bug... agh. Otherwise the betas and even the builds are great, so I shoudln't be complaining as much as I am currently.


Anyways, I have to admit I have regained my faith in ZC 2.5 and in the developers now that DD has come back and is doing tons of bugfixing. Oh and I wonder where _L_ disappeared... did he officially take a break or did he just vanish somewhere? No need to answer that question, but even though he has some awesome skills, I kind of like it now that he isn't here adding those features behind everyone's back... everything just seems to go so nice and smooth right now.

[Majora]

Oh and I wonder where _L_ disappeared... did he officially take a break or did he just vanish somehwere? No need to answer that question, but even though he has some awesome skills, I kind of like it now that he isn't here adding those features behind everyone's back...

I was about to ask that! Well, _L_, if you are reading this thread, any input?

EDIT: Is their any chance I can become a tester?

[koopa]

majoras_wrath, everyone can become a beta tester now ... just download the latest builds from AGN and don't forget to report bugs! See the box at the top of the beta forums.

[Eliwood]

...
I'm perplexed as to how you've reached that conclusion. The only NEW bugs introduced since the start of the public alpha builds was the problem with ZC requiring random .dll files, which I've fixed and take full responsibility for. Every other bug was there before. Yes, there is a problem with MIDI loading, which wasn't as visible before as now, but I'm working on that one. On the other hand, excluding the ZCL (which is being worked on exclusively by Sean and thus is not taking any time away from the other developers), I count
- 21 bug fixes, nearly half of which are for critical, crash bugs
- only 2 new features (integrity checker and romview)
- 2 minor cosmetic tweaks (scripts menu and logo)
That's an average of around 2 less bugs per day, with 90% bugfixing and 10% new features. Seriously folks, what more do you want?

I have no problem with people voicing their opinions, and welcome constructive criticism, but I don't see the point in attacks such as "omg ZC is getting buggier by the day" which clearly contradict the facts, and make me wonder why I'm donating my time to this project.

Although Koopa's already done a great job mediating the situation, I'll step in and say that you guys really need to hire a Community Manager because what we have here is a big gap in communication between the Developers and the general public. It's nobody's fault really. It just happens when the Developers are one level and the rest of us are on a different level, and the Developers assume that the public knows what's going on, and vice versa. Nobody's entirely right, or entirely wrong. It's a situation that's remedied by someone who can act as a liaison between the development team and the public.

For real companies with real products (Blizzard, Valve, etc.), this position is called the Community Manager, and the community manager bridges the gap between these two groups. The Community Manager takes cues from the development team and presents them to the public in a way that is understandable, and goes out of his/her way to make sure the public knows what's going on development-wise. Likewise, the CM takes all important open issues and suggestions and presents the to the developers in a way that they will understand.

In this way, we have someone who can communicate with both sides, and prevent a thread like this from ever existing. For months, there's been a disconnect between what the Developers think and what the public thinks, and it's about time to close that gap for good.

Just a suggestion. There are 1-2 people here who would be qualified for the spot, and you know who those people are.