I would have much less of a problem with rapid release if they'd stop changing major version numbers.
There is a three part version number for a reason, use it.
Firefox 5 should have been 4.5 and 6 should have been 4.6
If you want to release every other month to imply there are improvements, go for it - just don't imply there are major changes for "political" reasons.
The reason Firefox keeps changing major version numbers is that they no longer guarantee binary or source compatibility for extensions from one release to the next. Since the releases are much closer together, there's less chance that any particular addon will be affected by the changes between version X and version X+1, but there's a much higher chance of addons being effected than, say, in any given six-week period during the reign of Firefox 3.6.
This in itself is kinda crazy. Maybe Firefox has hit some new stage in its lifecycle, but it seems to have completely turned face on what originally helped it catch on. I can't imagine any add-on author targeting a platform that makes no guarantee on compatibility after 3 months.
Why would anyone support old releases of Firefox when it is so easy to keep it updated? I'd rather have a break-early policy for my code than a break-badly-later one. By making changes earlier, the likelyhood of bad breakage is much smaller. This is actually good for plugin writers.
If it's so easy to keep it updated, then it by definition must be easy to support older versions. I.e., either the changes are big or they aren't. Supporting your existing userbase is never about how easy you think it is to upgrade.
Maybe all the big guys are doing it wrong, but with every new Firefox release since 3.6.x I've had problems getting Firebug and 1Password working. And I have a hard time faulting them for not being able to keep up. And I've just had to give up on using the ElasticFox add-on at all.
It's not quite as crazy as it sounds; many of the most popular addons have a significant user-base on the nightly or beta channels (and now Aurora), so there's always been some pressure to update add-ons as changes were made to the underlying platform, even before those changes made it into an official release.
Also, these days Mozilla has the "Add-On SDK" which provides an abstraction around the core Gecko and Firefox APIs that is more Chrome-like - more limited than what a traditional Firefox extension can do, but guaranteed to stay compatible for a long time into the future.
Please tell me you are making the exact same post about Chrome's version inflation? "Political Reasons"? please elaborate. Mozilla announced months ago that it was not using version numbers for marketing purposes and they announced recently that they are doing away with version numbers completely, so I fail to see what the "political reasons" might be.
I thought the goal was to motivate users and IT organizations to stay current. This frees up Mozilla developers from fixing bugs in N many different old versions of Firefox.
To me, the reasoning behind the change in version is pretty clear. Especially looking at the disaster that was IE6. As soon as there is any support for old versions some IT department somewhere will immediately decide it's time to freeze their version. Moving forward after making a freeze like this becomes very difficult (that has been my experience). This presents two options 1) It's their own problem and move on or 2) Continue to patch old versions. The issue with patching old versions is bad IT decisions end up costing Mozilla time. Meanwhile chrome just continues chugging along at full speed only supporting current.
The current linux distro versioning and support structure seems to map well to this problem. Ubuntu, for instance, has LTS versions that IT can stick with and are patched for X many years.
Windows users update their anti-virus software and definitions in near real-time. Operating Systems update near real-time... why not browsers. I now use chrome most of the time and I am (surprisingly) happy with their versioning. I was a naysayer when they first started too.
Isn't this self-defeating when there are eight major version number changes per year? Won't people start ignoring the need upgrade because there will be another new version in a few weeks?
In all seriousness - Firefox version 25 in two years, think about that and how no-one is going to take major version numbers for real anymore.
If they insist on this, I'd rather have a YEAR.MONTH version number. Firefox Version 2012.4
I've stuck with Firefox over Chrome because I don't want Google to win the browser. But with most of the signs and some news out of the Mozilla camp having not been good lately, each piece of news nibbles at my ear just a little bit more.
I was considering moving from Firefox to Chrome, but with Firefox's recent performance improvements, I no longer see any reason to. The one benefit which Chrome had (performance), I no longer see. I like that Chrome exists, because it pushes Firefox to be better.
I've been using Chrome for Mac ever since it was in early, buggy developer builds. I was using Safari before that, but only because Firefox on OS X sucks.
For Windows, on the otherhand, I love Firefox so much. I do fine it's performance and UI so much better. Chrome sucks with lots of tabs open and the UI just bores me.
Not really. The actual number of bugs (not enhancements) confirmed as legitimate (NEW, ASSIGNED, REOPENED) but not fixed across both the Firefox (interface) and core (backend) projects in Bugzilla is 31,992.[+]
This doesn’t matter. At the end of the day it’s about severity. If Firefox ships with 31,992 critical end-user visible bugs then it’ll be rightly panned. It hasn’t been, therefore the large majority of the issues aren’t that severe, therefore they can be dealt with at a slower pace.
It took me some time, but I have come to terms with the rapid release process. I believe that it is actually a very good idea, as it makes the web much more competitive against, say, native iPad development – I may be a tad sectarian, but I tend to prefer the (mostly) open web to the (definitely) closed iPad model.
Now, the problem mentioned in the OA is quite real. Some of the processes of Mozilla have not scaled up yet. I suppose that they will.
In all large software projects I've worked, amount of open bugs always increased over time, only way to decrease them was to aggressively mark reports as ignored or wontfix.
That's great! That means we will get frequent releases, each with some bug fixes, instead of being stuck with an old, bug-ridden version, forever waiting for the perfect release.
Right, and not all of the "in-progress" features get shipped in each release. Each release only ships the parts that are ready. That way, users get the benefit of new, finished features without having to wait for thousands of bugs in unrelated parts of the code to be fixed first.
32 comments
[ 2.8 ms ] story [ 51.3 ms ] threadThere is a three part version number for a reason, use it.
Firefox 5 should have been 4.5 and 6 should have been 4.6
If you want to release every other month to imply there are improvements, go for it - just don't imply there are major changes for "political" reasons.
Maybe all the big guys are doing it wrong, but with every new Firefox release since 3.6.x I've had problems getting Firebug and 1Password working. And I have a hard time faulting them for not being able to keep up. And I've just had to give up on using the ElasticFox add-on at all.
Also, these days Mozilla has the "Add-On SDK" which provides an abstraction around the core Gecko and Firefox APIs that is more Chrome-like - more limited than what a traditional Firefox extension can do, but guaranteed to stay compatible for a long time into the future.
To me, the reasoning behind the change in version is pretty clear. Especially looking at the disaster that was IE6. As soon as there is any support for old versions some IT department somewhere will immediately decide it's time to freeze their version. Moving forward after making a freeze like this becomes very difficult (that has been my experience). This presents two options 1) It's their own problem and move on or 2) Continue to patch old versions. The issue with patching old versions is bad IT decisions end up costing Mozilla time. Meanwhile chrome just continues chugging along at full speed only supporting current.
The current linux distro versioning and support structure seems to map well to this problem. Ubuntu, for instance, has LTS versions that IT can stick with and are patched for X many years.
Windows users update their anti-virus software and definitions in near real-time. Operating Systems update near real-time... why not browsers. I now use chrome most of the time and I am (surprisingly) happy with their versioning. I was a naysayer when they first started too.
In all seriousness - Firefox version 25 in two years, think about that and how no-one is going to take major version numbers for real anymore.
If they insist on this, I'd rather have a YEAR.MONTH version number. Firefox Version 2012.4
They're going to stop having any version numbers.
I guess it's a matter of perspective (and of which HN headlines attract your attention :) )
For Windows, on the otherhand, I love Firefox so much. I do fine it's performance and UI so much better. Chrome sucks with lots of tabs open and the UI just bores me.
[1]: http://tylerdowner.wordpress.com/2011/08/29/even-more-clarif...
Now realize that's just semantics.
For the record I am a diehard firefox user, 6000 bugs or not.
This doesn’t matter. At the end of the day it’s about severity. If Firefox ships with 31,992 critical end-user visible bugs then it’ll be rightly panned. It hasn’t been, therefore the large majority of the issues aren’t that severe, therefore they can be dealt with at a slower pace.
[+] https://bugzilla.mozilla.org/buglist.cgi?list_id=1199753&...
Now, the problem mentioned in the OA is quite real. Some of the processes of Mozilla have not scaled up yet. I suppose that they will.
http://catb.org/~esr/writings/homesteading/cathedral-bazaar/...