Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[darkhog]

Basic idea is to take what Minecraft is doing and release VERY work in progress builds of DF every week or so (aside of stable updates), even (and possibly even especially) if it's a buggy, crashy mess.

Why? Few reasons:


- People who really can't wait for the new update can try out new features (with the recognition game may be/is broken until a full release)
- Toady can get feedback both on design and bugs so the stable release can be really stable and not require .many fix releases.

This would especially help during the Big Wait, which in this case would be Big Wait for a full, stable release, but people who aren't scared of crashes and bugs that may end up potentially destroying a world (if something nasty such as data corruption would be discovered) could try the new stuff earlier and give feedback as it's being developed to help steer the development that will satisfy both Toady and the community.

[Shonai_Dweller]

While nice fan service, what's the benefit for Toady to do this?
First of all it makes the bug tracker way harder to deal with though.

Toady's only one guy. Releasing crashy, buggy alpha alphas throughout the long-term development periods means spending a huge amount of time analyzing everyone's issues. He already releases 15-20 or so versions for each release and that seems to keep him pretty busy. And if he's not going to do the analysis at all, then it's just time taken to put out a release with no obvious benefit.

If he had a QA team (as Minecraft presumably does) it would be feasible. But as a one-person project, probably not.

That's just my initial thoughts. If you have experience with one person projects doing this and it benefitting the project then please share.

[darkhog]

As I've said, the benefits are obvious. Less bugs such as drunken cats sneaking into the (allegedly) stable releases of the game. Also, once the game starts being sold on Steam (and believe me, it will sell like hot cakes), Toady will have more than enough money to hire someone to just fix the bugs - hell, I'd be glad to do it as I'm quite a good programmer myself while he works on stuff like new features.

[Shonai_Dweller]

As I've said, the benefits are obvious. Less bugs such as drunken cats sneaking into the (allegedly) stable releases of the game. Also, once the game starts being sold on Steam (and believe me, it will sell like hot cakes), Toady will have more than enough money to hire someone to just fix the bugs - hell, I'd be glad to do it as I'm quite a good programmer myself while he works on stuff like new features.

Ok.
But, just supposing his pixel art indie game only available in English somehow doesn't become a global hit outselling Minecraft and he continues working on it with Zach...

All the features he adds are related to other features including all the new features he's about to add. A report that a feature is buggy means nothing if it's meant to interact everything else in the release. That renders the bug tracker near to useless unless he follows up issues in real time with each release (which is exactly what he does right now with 15-20 bug fixing updates. But if you add weekly releases during long-term development too he would have to do that all the time. Which I don't think is something he's likely choose to do).

There were no dead drunk cats in the "stable" release of the game. He released and during the cleaning-up period when he puts out fixes based on bug feedback both on the tracker and in real time on the forum he investigated and fixed the cats. Final release of that arc had no dead cats.

He's already said he'll be fixing major bugs while working on long-term development from now on, so stuff discovered after the final release (weapon traps) can still get looked at.

[voliol]

Steam has support for unstable branches. Considering the very first iterations of a major version almost always have some kind of game-breaking bug, just releasing those as unstable and then maybe the third or fourth version as stable should suffice.

Snapshots could also work, but I imagine in-progress DF versions might be too broken for even this. There also is the aspect of Toady wanting not to leak too much of the game’s code, and having to deal with that.

Let’s also remember they are already in collaboration with Kitfox. IIRC they already considered having them fix some play-testers as the premium release (and presumably others as well) approaches, though I might be misremembering here.

[Shonai_Dweller]

Yeah, he mentioned needing to think about closed testing before the Steam release rather than the usual release-then-fix method.

[darkhog]

You really think it won't outsell Minecraft? The only reason DF made so little money so far is that it was donation ware and ASCII game (unless you get a tileset). Both of which are getting fixed for the Steam release.

Trust me, both Adams brothers and Kitfox will make millions out of DF, if Kitfox won't screw up marketing and won't have Toady make arbitrary changes that fly in the face of what DF is.

[Shonai_Dweller]

You really think it won't outsell Minecraft? The only reason DF made so little money so far is that it was donation ware and ASCII game (unless you get a tileset). Both of which are getting fixed for the Steam release.

Trust me, both Adams brothers and Kitfox will make millions out of DF, if Kitfox won't screw up marketing and won't have Toady make arbitrary changes that fly in the face of what DF is.

Minecraft sells millions a week. It's used in schools throughout the world as an educational tool. It's available on multiple formats in many different languages. It's backed by Microsoft.

Yes, Kitfox are nice folk, Toady's game is the best game ever to us scant few. But, no, it's not going to outsell Minecraft. Or even Grand Theft Auto 5. That's just not how the world works.

[darkhog]

And it was selling like that even before M$ purchased Mojang. It will sell and Toady will become a very rich person (which he totally deserves to).

[PatrikLundell]

Making a release takes time. First you have to have a somewhat stable build. Then you have to compile it on the other platforms. Then you have to perform regression testing and at least some testing of the new stuff.  This probably takes 2-3 days purely for the motions that have to be gone through, and the commercial platforms probably add some additional stuff on top of that. An emergency release to fix a single crash bug might get away with cutting down the regression testing to a bare minimum, but it's still a fair number of purely bureaucratic steps to get there.

My suggestion (which has been made earlier) is to aim for a "yes, we're still here and supporting the game" release about once per month during the Big Wait (the start of which currently seems to be 2-4 years away), unless there's a need for an emergency crash bug fix. These releases would include bug fixes of bugs in systems that aren't expected to change in the next arc branch (i.e. the M&M one), and so can be implemented in the current arc (which would then be the Premium + Villains catch up + Army improvements one) and just transferred to the next branch more or less unchanged. Similarly, minor improvements to localized systems that aren't expected to change might fit in here as well (say improvements to hospitals, or introduction of multi hauling for light items [goblin teeth, socks, etc.] beyond plant gathering, to take some possible examples). I'd aim for spending about a week per month on these releases, including the time taken to perform the releases.

I'd like this approach to be taken earlier than the Big Wait, but given that there seems to be a hard time limit for the Premium release with contents candidates that seem to far outmatch the available time, I can see the wisdom in only fixing serious bugs for the "current" branch (Villains) during the Premium arc (and the tail end of that doesn't have any need for extra releases, as bug fixes will have to come frequently anyway). If Villains catch up and/or Army Improvements sections need multiple months of work that won't be fit for release in chunks, I'd make "yes, we're still here and supporting the game" releases during those sections as well, but part of the adaption to the commercial arena will probably be to try to break up things into chunks that can be released: if the "cost" of uninterrupted development is to spend effort on "yes, we're still here and supporting the game" releases, it might be worth the effort to split things up so the chunk releases provide the progress indication (although there's probably no getting around some fixing of non critical bugs to ward off some of the criticism of bugs never getting fixed).

And like Shonai_Dweller, I don't trust darkhog when it comes to the enormous sales that will just roll in. The course is set, though, and we'll see what an entry into the commercial market brings in about a year, hoping (and working) for a good outcome. It's a game by geeks for geeks, and the Premium release won't change that, even though the tile set might lower the geekiness threshold a little. However, it might bring more geeks to find the game, without getting too many casuals/RTS:ers who'll feel cheated when they realize it's a geek game (hopefully with some of them discovering their inner geek and liking it).

[voliol]

@PatrikLundell Monthly snapshots during the Big Wait(s) won’t be plausible as e.g. the map rewrite will take several months by itself. Intermediary releases of it would be broken by nature.

Or do you mean bug-fixing versions releasing in parallell to the main new feature development, with them being merged at the time of the big release? I am more positive to this model for sure. That way we can get some bug-fixing/QoL without being bothered by never arriving at the Big Wait.

[darkhog]

First you have to have a somewhat stable build. Then you have to compile it on the other platforms. Then you have to perform regression testing and at least some testing of the new stuff.

Except that's the thing: When releasing a snapshot you don't care about any of these things - that's the magic of it! Players brave enough to download a build clearly marked as potentially buggy/crashy (more than a regular DF) will do all the testing FOR you, free of charge.

[PatrikLundell]

@voliol: The latter, i.e. releases in the current branch that are merged into the next branch, mostly selected to be easily merged. The current branch is the one that's actually used by players, while the next branch is the arc Toady is working on, potentially for years. The next branch won't replace the current one unless it's possible to break the arc into chunks (and Toady didn't see any point in doing the map rewrites first, release what's roughly the same functionality [although some spheres would replace Evilness and Savagery, but probably to about the same effect] and embark on the actual M & M stuff after that, in parallel with fixing bugs players find in the map stuff).

@darkhog: The first release of an arc is typically buggy to the hell and back, and so are the bug fixing releases following it for a month or so, before DF becomes somewhat playable again (but still very buggy). You can't really lower the standard further, and you're sorely mistaken if you think there's a clear net win to exchange testing things yourself in a somewhat controlled manner for hundreds of poorly written bug reports (and a few, but far too few, clear, well written ones) where you have to try to guess half of what the player did to try to replicate it so you can analyze and understand it. Sure, waiting further to have the DF community do some of the analysis can save some work for Toady on that part, but even though there are some really bright people here, analyzing things without access to the code is hard.