This is a continuation of The Three Pillars post as the article referenced in that post also discusses this as well.
In another email, Dave Camp is “Revisiting how we build Firefox”. This one is a lot more technical as it focuses a lot on getting Firefox away from XUL and XBL. However, in the beginning of the email he does talk about speeding up deployment:
Since Firefox began, the industry has continually evolved how
it deploys code to users, and today it isn’t done on an 18-week cycle. We think there are big wins to be had in shortening the time that new features reaches users. Critical fixes should ship to users in minutes, not days. Individual features rolling out to small audiences for focused and multi-variate testing. As Laura Thomson put it in her Whistler presentation – “The trains have served us well, but it’s time to build a hyperloop.”
I agree that critical fixes should ship to users sooner. But, at the same time it is important to make sure these critical fixes not only fix the issue, but don’t also break something else. A good example of this going wrong was Firefox 33 which in a span of one-month from the first release had five updates. Further users are still not happy with getting major updates every 6-weeks, even though this has been going on now for 4-years (Firefox 5 June 2011 was the start of Rapid Release). Some users may view multiple updates of the same branch as Mozilla not being able to get it fixed right the first time. In reality though it is the nature of the beast (in this case the web) in that as soon as one exploit is fixed several more are discovered. This why there are so many frequent updates to Flash and for that matter Microsoft Windows.
“Individual features rolling out to small audiences for focused and multi-variate testing”. Haven’t a clue on what exactly they mean by this, much less how it is going to be implemented. But at the same time, isn’t that already in place with the Nightly, Aurora and Beta branches? To me it sounds like they want to make select end user guinea pigs. I suspect there will be more about this in the future or it will be scrapped.
Now getting back to the main focus of Dave Camp’s email, getting Firefox away from XUL and XBL. This is one of those ideas that looks great on a paper but the reality is this would be a major task and a lot more trouble than it is worth. Sure this was done with Firefox for Android, but that is an entirely different platform. Plus when they did this Firefox for Android had not been in production as long as the desktop version of Firefox. This discussion in mozillaZine points what the consequences are for removing XUL from Firefox.
Sure Dave Camp has some great ideas on how to build Firefox, but they seem unrealistic and unreasonable at this time. There is so much on Mozilla’s plate with bug fixes and other things that have a higher priority at this time (such as e10s).
Source: ars Technia UK