30 comments

[ 3.2 ms ] story [ 84.8 ms ] thread
I can understand why Mozilla feel threatened by Chrome's growth but why the use of major version numbers for what could have very easily been 4.1 ?
Marketing reasons, tech blogs don't write about minor version increments.
That's the best explanation I've heard so far, makes a lot of sense!
Sadly, there isn't much to write about because there isn't anything new here. As far as I can tell, tech blogs don't write about Chrome versions anymore -- I don't even know what number they are up to.
12 is the current release, and 13 should be ready soon. One thing chrome has done really well is smooth and seamless upgrades, this helps them move things forward and people don't notice the version numbers crawl up.
The dev channel is already up to 14.
Tech blogs most certainly do mention Chrome versions. For example, Ars Technica and TechCrunch did so for the last Chrome release.

The theory that people will start ignoring versions has not yet been proven. It still might though.

I wish the lesson learned from Chrome was that version numbers don't matter. I think it's unfortunate that some people have drawn the exact opposite conclusion.
Exactly. I hope in the future we don't think about whether you are running version X or Y but whether you run the stable or unstable update channel.

It's not like your users really give a fuck about version numbers.

They're going to make four releases a year, so there will never be another major release. Sticking with the old numbering scheme would result in "deflated" version numbers like 4.39.
It's the same release model as Chrome. Chrome doesn't go from 10.1 to 10.2 to 10.3... It's a time boxed release. 4-6 times a year you get an automatic upgrade and the version number is increased. They could have chosen the smaller increment like 10.01, 10.02, ... but then we would never have a major upgrade, and I think humans prefer the major upgrades numbers. In reality, it's just a number. They could have started with 3.14 and added the numbers of Pi, but that's already been done.
The expanding-decimal notation is very appropriate for TeX, because its implementation is just converging to the behavior specified in the TeXbook (i.e., pi), not adding features.
Because every release of Firefox could potentially break APIs.
Really cheapens the numbering of versions. We're going to have Firefox 1243053 soon.
Uh, no, not really. They are shipping 4 versions a year, so if they keep the current pace, we in fact won't see version 1243053 until the year 312774

Granted I hope we have transcended browsers by that point.

https://wiki.mozilla.org/Firefox/Roadmap

It's almost 9 releases per year (every 6 weeks). So by year 2015 we will be looking at Firefox 30. Makes the Firefox 4 look kind of obsolete...
I sure hope that we have enough advances in 4 years to make todays browsers obsolete...
I think we'll see a phase out of version numbers on public facing download sites and just have "Firefox" instead of "Firefox v4, v5" etc. Internally, devs might have Firefox 1243053 but we might never see that as the version number, sadly.
I'm not sure what the article means by "ahead of schedule". The release date was set months ago, and new versions are always released a few days early via the Mozilla FTP servers.

Firefox 5 isn't a huge release, and there won't be nearly as much fanfare as there was with 4. The point of Firefox 5 is to switch over to the "train" model. The train leaves four times a year, no matter what-- and if a feature doesn't make the train, it has to catch the next train. There is no waiting for a feature.

Incrementing by one each release (rather than 4.1, etc) may seem like Mozilla is making a big deal out of nothing, for the sake of publicity. However, it just means that all releases are equal. There will never be a huge 3.6 => 4 style release ever again, and eventually people will lose track of what Firefox version they have (much like how people have lost track of Chrome versions).

tl;dr: While it has a few cool new features, Firefox 5 is more to get Mozilla in the release-often groove.

1. Are they moving to a seamless upgrade system like Chrome?

2. If not, will it be auto-update or opt-in update?

3. Are plugin writers going to have to update their plugins every time there's a new release?

1. Upgrades will be downloaded in the background and installed automatically by default. It's not yet as seamless as Chrome, though it's moving that direction. (For example, a while after the upgrade is downloaded, current versions of Firefox will prompt you to restart, with a dialog which is a bit more intrusive than the tiny "Christmas tree" toolbar icon that Chrome uses.)

2. Updates will be automatic by default. They'll be installed without prompting, unless users disable auto-updating. This is important because, just like Chrome, the updates will include security fixes.

3. Add-ons hosted on addons.mozilla.org will be automatically updated: http://blog.mozilla.com/addons/2011/04/19/add-on-compatibili...

Lots more details in the links at http://mozilla.github.com/process-releases/ and https://wiki.mozilla.org/RapidRelease

Small correction, releases are planned each 6 weeks, same as Chrome. Which is almost 9 releases a year.

The reason this first release took longer, is it is the first one. But FF6 is planned for release 6 weeks after FF5 and so forth.

Is this the direct link?

ftp://ftp.mozilla.org/pub/mozilla.org/firefox/releases/5.0/

Joe Hewitt (early firefox) and crew are amazing for making a kick ass open source web browser that was the first to gain significant mind share in the browser battle. However Chrome has been so much better in many ways and evolving so quickly to cover its deficiencies that Firefox seems like old news. If you haven't at least tried Chrome you are behind the curve here.
I feel like everyone saying how we'll have Firefox 346 is missing the point.

The point is really to set up something like a rolling release system where you don't really care which version you are running. The only thing you choose according to your risk appetite is how stable your channel of updates is.

Maybe they'll end up with something like Processing's[1] versioning scheme, where there are fairly regular numbered releases (e.g., release 0115) and some are granted a special status (e.g., release 0196 = version 1.5.1). I suppose it's really just another 'layer' of versioning: VCS commit -> release number -> version number instead of VCS commit -> version number.

[1] http://processing.org