Firefox version fragmentation increasing – is Mozilla in trouble?

FirefoxMozilla’s development pace for Firefox went into overdrive this year, as they adopted a strategy similar to that which Google uses for the Chrome web browser. Mozilla’s new, rapid release schedule for Firefox calls for a new version every six weeks. On Tuesday, November 8, it’s already time for the release of Firefox 8.

But there are clouds on the horizon. For every new version of Firefox that Mozilla releases, a fraction of users are for whatever reason not being upgraded. There’s a long tail of older versions starting to form, and over time this may accumulate enough version fragmentation that it could become a real problem.

To give you an example, Firefox 5 was the first version in this rapid release schedule and should be history by now, but it still clings on to almost 4% of the Firefox user base.

So imagine the effect of this kind of “left-behind” retention in the long run. Remember, at the current pace we’ll see at least eight new Firefox versions per year. Next November, the latest version will be Firefox 16.

The current fragmentation of Firefox

Here’s the distribution of Firefox versions for the final week of October 2011, as measured by StatCounter. Note that this is at the later part of Firefox 7’s six-week cycle as the latest release version, so browser users should have had plenty of time to upgrade – or rather, be upgraded.

Firefox version division
In case you’re wondering, the vast majority of the “3.6 or older” bracket is made up of version 3.6.

The good news is that the “legacy” versions of Firefox – those being version 3.6 or earlier – represent a rapidly diminishing portion. Early in June they accounted for 43% of all Firefox usage. Now, five months later, those older versions are down to representing “only” 28%. At the current rate of change, they will be down to around 8% in a year.

The bad news is that there is now a good-sized portion of users who are not following along from one version to the next in this new release schedule that Mozilla started with Firefox 5 in June.

Why is this happening?

It could possibly be an effect of the upgrade process not being automated effectively, or perhaps enterprise users are to blame. The typical reluctance of enterprises to upgrade software without ages of testing has been the bane of software progress for quite some time (part of the reason why IE 6 is still alive).

Perhaps those are both factors. We honestly don’t know. Firefox users do have the option to disable automatic updates, but they’re on by default. It’s hard imagining that many people opting out of upgrades.

A silver lining, the saving grace, the Good Thing

The saving grace in this whole situation is that at least the share of Firefox users running the very latest versions is not going to drop for a while. It might actually grow.

Why? Because people keep upgrading from older versions, like 3.6, filling up the gap that would otherwise be created. But once the pool of older Firefox versions starts to dry out, Mozilla could be in trouble.

Are we facing fragmentation hell?

As we mentioned in the introduction, with a new version coming every six weeks, we’re going to get eight new versions of Firefox every year (well, 8.7). If Firefox is going to leave behind a couple of percent of its users or so in every version update, things will quickly become messy even if those shares individually are rather small and will decrease over time.

This is what things could possibly look like just one year from now, unless Mozilla comes up with a more effective upgrade process that doesn’t have users staying behind for every new version that arrives:

Firefox version division, a hypothetical future

Now imagine web developers around the world collectively going bold from tearing their hair trying to test and support that many versions of Firefox. Then add another year on top of that, when we’ll be up to Firefox 24.

Final words

The bottom line is that leaving a chunk of the Firefox user base behind with each new version is not a sustainable situation in the long term, at least not with such a rapid release schedule.

It may well be that the good people at Mozilla already have some awesome plan in place to counter this trend, but if they don’t, it may be time to start thinking of one.

13 comments

  1. It’s partly enterprise and partly add-ons. Mozilla is killing their addon community by requiring them to update their addons every 6 weeks for the next major version relase. Most addons are done by developers as part time learning projects. That’s ok when you only have to update it every 6-9 months. When you have to update it every 6 weeks, it’s too hard to keep up. And if you’re a user relying on these addons, you either switch to Chrome, or stop upgrading until your addons are updated to work in the latest version.

  2. They keep breaking some of my favorite plug ins. I stopped upgrading because of that. Plus, the newer versions were *flaky*.

  3. The constant upgrade cycle and breaking all the plugins have forced me to stop using Firefox. Why should I go all the trouble every six weeks – each time afraid that plugins will stop working.

    I have switched to Safari / Chrome combo and it’s much less stresss.

  4. I am writing this in FF 3.6. I tried upgrading to FF4 when it had been released, but had to downgrade, because some addons were not available for the new version. Since then, every time a new version is released, I check my favorite addons only to find out that at least one of them won’t work with the new version. So, I stay with 3.6 — the version packaged with Ubuntu LTS 10.04.

    You probably already understand when I will update my Firefox. Within the first years of life of the next Ubuntu LTS version.

  5. A side from the add on issues a good chunk of the legacy percent could be attributed to the linux community, for example the last lts version of ubuntu (lucid lynx) is still running 3.6. so I’m sure that once the next lts is released those numbers will probably drop even more.

  6. Firefox is relatively in better state than Internet Explorer. IE 9, (the latest stable version) has just 10% market share. Possibly because it is not supported on WinXP.

  7. Mozilla is probably bored and don’t want to be leader anymore. Fast major version changing is good way to say CU to the market. In a year max two ppl will swich to chrome, same crazy version changes but … still working.

    It’s a shame for mozilla as open source and open minded project to do idiotisms like this. 🙁

  8. Some people are already mentioning that add-ons actually break, but maybe more important is how we users perceive version number changes: a major version number change is associated with breaking functionality, radically changed interfaces, licenses or automatic upgrades that are no longer valid (well, not with FF of course). This perception could very well be a reason that holds people off.

  9. This activity by the developers is destructive. It really is to no advantage of anyone involved, because there are no resources to take care of the plugins. If things need to change that often, then there is not much thought going on for each change. The web technologies and the need for new features, in general is not changing that fast.

    Lately, I’ve stopped using Firefox too, just because it seems so slow and clumsy on my machine.

  10. One moment ago I updated Firefox from version 7.0.1 to version 8.0. It paused to check my add-ons, themes, and plugins for compatibility, accepted them all, finished the install and started up. So, why are you laggards holding back?

  11. Theory 1. Clumping together 3.6 actually includes multiple 3.6 versions. Some of v4+ are effectively micro version updates, so they just look like more versions.

    Theory 2. FF users are more techy, thus are less likely to let it auto-update. Major version changes sound like trouble, so people avoid it.

    Maybe it works better for Chrome because people got used to it since the get go, but I think this whole “I got a bigger number” race is childish or at least counterproductive. It’s almost as bad as having an SVN revision your only version identifier.

    Classic version numbering makes sense and allows you to get a quick assessment on the nature of changes. Micro version = bugfixes and minor feature. Always update. Minor version (in the case of FF at least) = bigger changes, better read changelog and maybe wait 2-3 microversions. Major version = drastic changes. Dual install, assess situation longer term.

    Also from the web developers’ perspective too many versions complicate matters. A stable platform is better in general (the same goes for Ubuntu, BTW, but that’s another matter).

Leave a Reply

Comments are moderated and not published in real time. All comments that are not related to the post will be removed.