Mozilla Considering 5 Week Release Cycle

[DKT27]

Mozilla, not content with its monumental shift from four major builds in five years down to a new stable build every six weeks, is looking at outputting a new release every five weeks, or perhaps even less.

Christian Legnitto, a project manager at Mozilla (and currently the "release manager" of Firefox), announced the intention to shift to a shorter release cycle on Mozilla's planning mailing list. In response to one developer citing the success of the six-week release cycle, and asking whether it would be feasible to speed it up even further, Legnitto said: "Yes, I absolutely think in the future we will shorten the cycle," but recognizing the pains caused by the sped-up process, he added "But it won't be soon. We have some work to do to make 6 weeks smooth from a process, tool, and product side."

The shift to the six-week release cycle has actually been less painful than anticipated, but there are still a few aspects of the Firefox ecosystem that have dragged their heels — namely communication from above, localization, and add-on maintenance — but Mozilla intends to tackle these problems in the next few weeks with blog posts and videos.

As it stands, with the four-channel system, it takes 18 weeks for changes to the Nightly (mozilla-central) channel to percolate down to end-users. A five-week release cycle would reduce that period down to 15 weeks; a significant improvement. The main question, though, is whether each of the channels (especially Aurora and Beta) can keep up; a lot of usability and stability testing, and localization has to be carried out before a build can be labeled Stable, and five weeks (or less, once you factor in the time it takes for builds and installations to roll out) might not be enough.

There's also the matter of whether Firefox (or Chrome, or IE) needs to move faster than it already does. The initial shift to a six-week release cadence was almost certainly Mozilla's reaction to Chrome and Internet Explorer's crazy pace in 2010 — but now, in 2011, the browsers are at standards and performance parity, and the only big differences being stand-out features like SmartScreen, Sync, and Cloud Print. The only real reason for a mega-fast release schedule is to plug security holes; but even before 2010, all three browser vendors already had the mechanisms in place to deal with that.

You could argue that the internet is a fast-moving place and that browsers need to be agile creatures — but if anything, the web is actually maturing at the moment, and web browsers are following suit, becoming more and more like dependable operating systems by the day. You can't see Microsoft or Apple releasing a new version of Windows or OS X every five weeks… so why should consumers, developers, and enterprises buckle down and brace for impact every month and a bit?

It might work if Mozilla rolled out a service pack every five weeks, but kept the major version number bumps down to once every six months or a year — that would be more in-line with how operating systems currently work, and it would probably go a long way towards appeasing add-on developers, and admins who are in charge of rolling out a new version of Firefox onto thousands of computers every few weeks.

View: Original Article

[DKT27]

Please note that this is NOT going to happen till they fix the addon compatibility issues. The addon issues solution is already underway and the above will only apply after they have fixed that and other issues.

[ck_kent]

With these fast release cycles, I wonder how long will it take for the developers to run out of ideas for improvement? :huh:

[marshall39]

With these fast release cycles, I wonder how long will it take for the developers to run out of ideas for improvement? :huh:

I think they will never run out of ideas , when it will be the case they will only ask themselves "WWGD" What Would Google Do? :P

[Gabben]

This is ridiculous... :frusty:

[HX1]

I firmly believe that if they address that issue facing them and their users right now with the current cycle first.. Then successfully this could be done.. but otherwise.. I see it only amping the situation up even farther.. I also see vulnerabilities and procurement of product safety.. fixes/updates/ others.. also being an issue that should not be left behind; as a side effect of these problems... ( in extensions and code as well in the core of the program.. )