Shouldn't trunk be a wild cesspool of experimentation?

[Pollen_Golem]

That's what I would naturally expect since devs have a good set of players willing to try out anything. But devs seem very conservative about making changes in trunk. I get the impression that anything devs push into trunk, devs expect to stick. Not sure why. I would understand if 99% of players played stable versions only, or if Crawl was a massively multiplayer RPG where players grinded their character for years, but neither applies to Crawl. Trunk is unstable by definition so players expect oddities anyway. This can't even be a coding problem - if it's hard to retract a commit after several others have been piled on top of it - because devs could just freeze trunk for a few days while the experiment goes on. Instead we have these "experimental branches" easily available only on berotato webtiles, as if to avoid ruining the common trunk crawler's experience.

[Sar]

are you sure you know what "cesspool" means

[greedo]

Because people don't really want changes I think.

For instance, I made a mod that gave people pretty much everything they wanted, foodless, auto-id, shorter less drawn out game, more difficult etc. I think.. 2 people have played it? Maybe three or four in case some people didn't post. If people didn't like it that would be one thing, but it was kind of bizarre finding people didn't even want to play something they said they wanted.

Secondly, it's kind of funny that MOST people play trunk. Crawl has a weird player base in that the players expect the developent branch to be fair and balanced and throws a shitstorm when it isn't because it might mess up their winrate oh no!

[KoboldLord]

Wild experimentation usually gets pushed into its own little branch. Once it gets bashed into being actually mostly functional, then it gets turned over to the trunk players for playtesting. There's no sense giving playtesters your code if it doesn't actually work yet when they could be testing the stuff that you are reasonably sure you will be releasing.

[archaeo]

Don't forget that coding big features and changes isn't at all trivial, and nobody's developing Crawl full-time. I imagine that most of the devs prefer to work on stuff that has broad support among their colleagues to avoid coding a bunch of stuff that MPA will just ruin remove from the game, so most of the "experimentation" comes in the form of crawl-dev chatter and other stuff. Furthermore, Trunk is everybody's sandbox, and a "wild cesspool of experimentation" is a good way to have a messy sandbox that makes it tough to untangle problems as they arise.

It's not like there's a deadline anyway. "Wild experimentation" is a fine idea when you want to get something done fast, but this is an open source game developed by hobbyists. Why not take it easy? Criticize it as "conservatism" if you want, I suppose, but I'm not sure what advantage there is for Crawl if there's a bunch of hurry; anyone, developer or player, who finds themselves bored with the current Crawl can dip out for 6 months or a year and come back to find a fresh set of changes and features to play with.

For instance, I made a mod that gave people pretty much everything they wanted, foodless, auto-id, shorter less drawn out game, more difficult etc. I think.. 2 people have played it?

You should collaborate with a server admin/dev to get the game onto a server, which is a much more reliable way to get people to play it. For example, I haven't tried out either of the foodless branches people have put together because I'm on OS X and can't be bothered to learn to build weird Crawl variants.

[tasonir]

Agreed; I'll usually try out experimental branches, but they have to 1) be on a webtiles server, and 2) I have to hear about them, usually on the tavern. A tavern post isn't required but it's certainly helpful to direct attention. I appreciate what greedo's going for, but I am also a lazy person who doesn't have enough time in the day...