Mozilla Firefox had a fairly slow release schedule up to Firefox 4; roughly once a year. New releases were fairly significant in terms of features, although there were lots of performance and security updates along the way. Especially pre-Chrome, Firefox had a strong community and the addon community really pushed Firefox’s limits; Firebug in particular comes to mind. Firefox addons are required to say which versions of Firefox they are compatible with; which is good, as most Firefox addons are tied to a particular version.
Google Chrome came out and it really raised the bar in terms of performance. Chrome follows a very rapid release schedule getting 3-4 versions out a year. This means that extensions written for Chrome are nearly forced into following the same release schedules.
In practice, extensions are almost always open source side projects and very rarely get the kind of dedicated development time to match Chrome’s schedule. Google chose to allow extensions designed for older versions of Chrome to be installed on newer versions, which leads to a lot of ‘Will it work?’ moments, but since Chrome’s extensions are closer to glorified bookmarklets, often they still will.
Mozilla Firefox is still struggling to get near Chrome’s performance, and one of the decisions Mozilla made was to adopt Chrome’s rapid release schedule. This is terrible to the Firefox addon community. At this point, every Firefox addon requires an update every few months, which is just out of scope for what most of these side projects get.
Enter Mojang. First off; I know Minecraft is in beta, Firefox is not. Minecraft seems to be switching towards an (even more) rapid release schedule. This is dangerous territory, as the game changes so much with every release that even texture packs break every release, let alone the websites (texture pack customizers), mods, server wrappers and variants that the community has built around Minecraft. This really drove home the point to me that community software really can’t keep up with corporate development, as the time investments are just too high for most community members.
I’m not advocating that Minecraft be released more slowly. “Release Early, Release Often” is an excellent decision for Mojang, as it allows the community to be involved with finding bugs and suggestions for feature design. Creating a modding API will also help. Making sure to document or distinguish between which releases will affect the server commands, for example, would help server wrapping software. Lastly, perhaps adding the ability for textures to ‘fallback’ to the default textures when they’re missing from a texture pack would reduce the pressure on texture pack artists to crank out new textures as fast as possible after releases.
