sungate.co.uk

sungate.co.uk

Ramblings about stuff

Firefox versions: someone’s not following The Rules

In March 2011, Mozilla released Firefox 4.0: this was the first new ‘major’ version for some time. Prior to that date most Firefox users were running Firefox 3.6.16, which was the previous latest version.

When a new major release such as this appears, users are typically faced with two choices: (a) stick with the current ‘branch’ (in this case 3.6), receiving further minor updates (3.6.17, 3.6.18,…), or (b) to plunge into the new major version.

Typically, new major versions bring with them: (i) new features and (ii) new bugs, so the choice someone makes (or your IT department makes for you) is normally made based on fairly simple criteria: in a conservative environment where a stable platform is highly valued (e.g. office, academic, companies), one would normally stay with the current version, to avoid exposure to the new bugs and where the new features are not considered very important; the latest version would be carefully tested before being made available to staff. Alternatively, adventurous home users and software developers who like having The Latest And Greatest (even if that’s often far from the most accurate description) will usually go for the new version.

When making such a decision, one can read the release notes (“What’s new? What’s fixed? What’s different?”), but people are increasingly making such a decision based solely on the Software Version Number. This may sound strange, but it’s a surprisingly reliable approach: if you’re running QuirkyApp 1.1.3 and you see an announcement for QuirkyApp 1.1.4, one won’t expect there to be much of a change. A few bugs or security flaws fixed, perhaps. However, if you see QuirkyApp 1.2.0, you might think “Hmm, something a bit new, there.” Further, an announcement for QuirkyApp 2.0.0 would be greeted with “Whoa! That sounds like a completely new version, lots of changes.”

One’s response to these changes will vary depending on your point of view. “Just minor fixes” could be seen positively (“Excellent, they’re fixing those annoying little niggly problems with the software!”) or negatively (“Dull, nothing new or exciting…”); equally, the “new major version” could be seen positively (“Lots of new features, looks shinier etc.”) or negatively (“Oh dear, they’ll be introducing lots of new bugs and rewriting the graphical interface for no good reason…”). And so on.

Of course all of this relies on the software developers following The Rules. When I refer to The Rules, I mean something like the following link. For a full discussion of how good software version numbers can be useful and informative, read Wikipedia page about software versions

In late April, Firefox released 4.0.1. All those 4.0.0 users think “Ah, right, security fixes. We understand this…” and everyone installs the update. The usual drill.

Then, a few days ago Mozilla released Firefox 5.0. “Erm, hang on, what’s this? We’ve only just got 4.0, now there’s 5.0?” Quite. This was my reaction. In the past, there has typically been about two years between major releases, so how come they’ve made a whole new version so quickly? The answer is, of course, that they haven’t. According to the developers and the people who know about these things, version 5.0 only includes minor bug and security fixes compared to 4.0.1: therefore, the versioning rules indicate that it should be called 4.0.2, not 5.0. Yet this change will have confused vast numbers of users, especially corporate IT departments who have to plan carefully for major version changes, but for whom simple security fixes (4.0.1->4.0.2) are more routine.

And it gets worse: there are preview releases already available for Firefox 6.0 and Firefox 7.0, both expected to be out this year. Yet the changes expected to be included in these updates are quite small.

Someone has broken The Rules. Who?

Mozilla’s Marketing department, it seems. Apparently, they seem to think that Firefox will be more popular if it gets a bigger version number. After all, its ‘competitors’ Microsoft Internet Explorer and Google Chrome are up to versions 9 and 12 respectively, so they must be better, right? I’m not going to go into a lengthy history of those two products, but suffice to say that both Microsoft and Google have been guilty of playing fast and loose with software version numbers for their browsers; that is, Not Following The Rules. And I don’t think anyone really believes that a bigger version number means that the software is better in any meaningful sense.

That Mozilla are taking Firefox from version 3.6.13 in January 2011 to version 7.0 by the end of 2011 makes one believe that – based on a traditional interpretation of version numbers – more development work has gone into the application this year than since development began more than a decade ago. Of course that’s not the case. I don’t think this helps anyone; it certainly doesn’t help end users, it must confuse the developers and it certainly complicates corporate environments where an IT department centrally manages the installations on a number of machines.

So can anything be done about this? Not really, the decision is not made by the users. However, I hope that at some point – perhaps when Mozilla Marketing are satisfied that the Firefox version number is Big Enough – they return to the more familiar and informative versioning scheme. Perhaps they can call Firefox 9.0’s security releases 9.0.1 and 9.0.2, rather than 10.0 and 11.0?