Guest post by A remarkably well reasoned editorial by Peter Bright at Ars discusses the implications of the new Firefox release schedule. The Crux of the argument is that by complaining about the “new” Firefox release, corporate customers are fundamentally misunderstanding the web and their place in it. He also reflects on historical reasons for their attitude and what they should do in the future to maintain parity with the evolution of the web.
I don’t get it, seems to me that the conclusions and suggestions for the enterprise somehow amount to “Yeah, just use IE”. Which is fine, that is what seemed to be the way to go anyway if Firefox insists on this rather loose style of release management.
I don’t blame Mozilla in any way (they are certainly not wrong in what they are doing), focusing on the consumer market and the fast-moving aspects of the web makes a lot of sense. The article however does not in any way I can see make a case for the enterprise being wrong either, they are just leveraging the web in a different web and of course many find it a bit unfortunate that only IE seems to cater to what they need to run critical business on internal webapps.
To be honest though, I fully expect some company to take it upon themselves to fork either Firefox or some Webkit variety to make a stable browser exactly for the sake of predictable release management. IBM seems a very likely candidate, possibly RedHat.
It is “Christmas in July” for Microsoft.
“Final Cut Pro” signaling Apple’s exit from the professional market. Mozilla broadcasting its own contempt for the enterprise.
There were two zingers from Net Applications this month:
“The iPad has 0.92% share of all browsing. In other words, the iPad has 53 times the usage share of its nearest competitor.”
“When Microsoft decided not to support XP for Internet Explorer 9, they narrowed the front for the browser wars to Windows 7. We’ve been tracking this strategy ever since, and in May, Internet Explorer 9 on Windows 7 reached 12.2% worldwide (including custom editions). In the U.S., Internet Explorer 9 on Windows 7 averaged 17% usage share during the last three days of May.”
http://www.netmarketshare.com/
To put this in perspective, Firefox, all versions, all platforms, has a combined 22% market share in the Net Applications stats.
IE 8 32%.
In its most recent financial report, it is quite clear that Mozilla is still bound hand and foot to Google and AdSense.
If Google pulled the plug, that would be the end of it.
The article completely misses all the non-MS enterprises out there, which are now left in the lurch.
For example, we just this month completed the upgrade to Debian 5.0 with Firefox 3.6. Since Google has ended support for FF 3.x we have just finished testing a manual install of FF 4.0. And now that’s been EoL’d.
There’s no way we can test and implement a new FF every 6 weeks.
I read this “6 weeks” number a lot, but I think it’s in fact 16 weeks (a bit less than 4 months).
However, we totally agree that Mozilla has EOLd Firefox 4 unacceptably too soon.
From the article, where he talks about what the corporations did and the IE6-problem:
“Progress is severely retarded by corporate foot-dragging.”
What is annoying about this is that Mozilla is in a way now punishing those corporations that did get it, got onboard, worked to deploy Firefox and are using something else than IE6.
And I think this is why people got so fired up.
The problem here is that there is no difference between Firefox and Chrome (in fact this policy is attempting to get some parity between them). The shorter release cycles mean that there is far less code and feature churn this reduces the likely hood of major bugs which in turn means that corporate customers need to do far less rigorous testing.
Many have accepted chrome as an enterprise browser; How is the firefox update policy any different?
Also as the article points out, Mozilla isn’t going out of it’s way to punish enterprises, It is just that users were always the primary focus.
Edited 2011-06-28 09:03 UTC
I’m fairly sure different enterprises have different policies/needs. Some chooce Chrome.
With Chrome you ‘always have the latest version’ on a daily, if not more frequently basis.
So if you are an enterprise that want to do testing, you just can’t with Chrome. At all.
If you’re testing web apps then this should not be an issue at all because you shouldn’t be targeting a specific browser release.
If your application needs an ultra stable api then maybe it shouldn’t be a web app.
If you’re testing the browser itself then, yes, you will have difficulties. Those difficulties assume that the browser changes massively between releases which isn’t the case with Chrome (and now Firefox). This means that you can start to treat each release as a minor release.
The article does point out that historically,even in minor releases Firefox changed/added/removed features. People have accepted that, how is this different?
In a word, scale. In the past, (say) 10% of new features would be backported to a stable branch. Now that number goes to 100%. That means a ten-fold increase in potential compatibility/functionality regressions.
Actually, unlike Firefox, Chrome supports both .MSIs and GPO so I’d say this latest move by Moz was just dumb. If I was still working corporate there is NO way I’d allow FF anywhere near my machines. Whether Moz likes it or not there are often REAL apps that REAL people require to do their job, and with a schedule as fast as Moz is advocating (in the blog they said they want to get SIX WEEK turnover) it simply can’t be tested AT ALL.
I personally have removed Firefox from my standard install images here at the shop and am currently switching my customers over to Comodo Dragon, which is Chromium based. Firefox killed about a third of my extensions with their little stunt, including some that my customers counted on for security. that is simply unacceptable so Firefox has to go.
So sorry Mozilla, it was nice while it lasted, but with the Chromium based I don’t have to deal with broken extensions. So far all my extensions have continued working just fine after updating, soemthing I can’t say about your browser. so long, and thanks for all the fish.
Does Chrome auto update?
I’m pretty sure minor revisions update automatically while full point releases require interaction, though I’m sure there is some way to automate it.
The real issue isn’t the updates/testing, but predictability and security-options.
Does Google Chrome 12 still get any security updates if you don’t upgrade to Chrome 13 ?
That was one of the biggest reasons people complained about with Firefox 5, Firefox 4 does not get any security updates anymore.
The Dragon which I’m using (based on Chromium not Chrome) will tell you there is an update but you have to download it manually. More importantly I’ve been from Dragon 8 through Dragon 12 and NO dead extensions, none. And I’m using the same extensions I was on Firefox.
Frankly to me this is Firefox taking a big old dump on the people that made them great. the ONLY real selling feature Moz had over Chromium was the extensions, but thanks to their crazy release schedule more and more extension devs are switching over to Chromium which means both Chrome and Dragon benefit. I only hope the Chromium guys manage to give the NoScript guy the hooks he has been asking for so noScript can come to the Dragon. that is the ONLY extension I need now, everything has been taken care of thanks to FF devs jumping ship.
So as I said before it was fun while it lasted, so long Firefox and thanks for all the fish
Actually, unlike Firefox, Chrome supports both .MSIs and GPO so I’d say this latest move by Moz was just dumb.
Is this what you are looking for?
http://www.frontmotion.com/Firefox/
Firefox killed about a third of my extensions with their little stunt, including some that my customers counted on for security.
Like what? Really, if the extensions don’t follow firefox update speed, you should not blame Firefox. Maybe you made some poor extension choices?
No…. We should blame Firefox. For not giving us any time at all to test whether their new browser is compatible with the Web apps / plugins we use before forcing it on us. This really is going to hurt Firefox adoption in the enterprise. It was a really dumb move on Mozilla’s part.
As they have stated, they are not doing what they’re doing for the sake of enterprise. So it is rather pointless of you to call it a stupid move. They knew what they were doing and they made the choice.
eldarion,
“Like what? Really, if the extensions don’t follow firefox update speed, you should not blame Firefox. Maybe you made some poor extension choices?”
Wow, I can’t believe what I’m hearing.
I’ve personally seen FF updates break extensions, it’s very frustrating. According to you we should blame the extension author, but not every 3rd party developer has the resources to keep up with mozilla’s betas, I know I wouldn’t. Doesn’t mozilla deserve more of them blame since it was they who caused things to break?
Edited 2011-06-28 17:04 UTC
Are you sure about that? I haven’t seen very much adoption of Chrome in any enterprise environment I have been in.
Corporate customers should be complaining. And no, it is not the enterprise that is wrong. it is Mozilla that is wrong. Enterprises that have thousands of users running mission critical internal Web applications simply can’t work with this kind release schedule. They need time to test their application and fix any bugs that are found. This is going to hurt adoption of Firefox in the enterprise. There’s no way around that. One of the major gripes about Apple, and one of the reasons they have seen little Enterprise adoption is because of complaints about OS X release schedules. Mozilla has made the problem 10 times worse than Apple ever did.
It’s not going out of its way to punish enterprises. But it is giving enterprises a great big F-U that basically says says “We don’t care about your concerns, wants or needs”. Kind of like what Apple does. And just like Apple, it’s seriously going to hurt adoption of Firefox in the enterprise.
Edited 2011-06-28 13:08 UTC
Firefox’s success is based upon the add-ons. As long as Mozilla did major updates every year or so, it was acceptable to require developers to update their extensions every version.
But now, it’s just plain ridiculous … They should choose an API, label it as 1.0 and keep it stable for, let’s say, one year, no matter the changes in the browser. Then, if it’s deemed necessary, they can make a different API, call it 1.1 or 2.0 and also keep 1.0 for compatibility reasons and map its calls to the new API. New add-ons can use the new version, older ones will still work in compatibility mode.
+1000, except that I think there are other appealing sides to FF than add-ons.
Yes, there are. Besides add-ons, I use Firefox because it’s one of the most forgiving browsers (if not the most) when it comes to malformed and not standard compliant markup. Also the Panorama feature is simply brilliant … you can group your tabs based on activities and never worry about bookmarks or clutter.
They actually have. It’s called Add-on SDK and has been released with Firefox 5. The downside is that it is far less powerful than Firefox’ “old-school” extension mechanism. Firefox developer Dave Townsend explains in a recent blog post why extensions tend to break with each new release:
http://www.oxymoronical.com/blog/2011/06/Why-do-Firefox-updates-bre…
Basically “old-school” extensions can hook into every subsystem of Firefox, there is no “API” per se which could be stabilized. That makes it ridiculously powerful but also causes the known compatibility headaches.
You know that they do NOT forget about this and Addo-Ons will be handled different in future?
The trouble here is that no-one seems to get what the actual problem is here. The linked article touches on it, but doesn’t really deal with it.
The problem is not just that the version number has jumped, but that it is an abrupt change in how Firefox’s versioning has worked up until now.
Enterprises are okay with version numbers that go up by major numbers each time. They’ve already got apps that do that. They’re also okay with apps that have rolling releases. Again, there are precedents which sit quite happily in the enterprise.
What is a problem is that Firefox moved to this new model without adequately explaining the change, and without any kind of transition period to allow their users to adjust.
Enterprises started testing with FF4 in the way they were used to for Firefox, and then found themselves still testing it when FF5 came along.
The article claims that now that FF is into a short release cycle, people should realise that new releases don’t need so much testing, but – and this is important – FF4 *was* a big change. It *does* need to be tested thoroughly when upgrading from 3.6. That’s what enterprises were doing when FF5 was thrust upon them.
If Mozilla had said “we don’t want to support old releases any more, but okay we’ll do it for a bit for FF4 to give everyone a chance to adjust”, then this storm wouldn’t have erupted at all. The fact that they didn’t do that, and then started lecturing the world about it not being a big deal just comes across as arrogant, and that’s the one thing that an open source organization can’t afford to be.
The problem is the abrupt EOL schedule of suddenly deciding 3 months after they released a major revision of their browser, that it’s no longer supported and won’t even receive security patches anymore. That’s really the basic issue. It forces business customer’s into an update upgrade schedule that is simply too aggressive for them to deal with.
If Mozilla wants to survive in the enterprise, they need to do something like Ubuntu does, and have a long term support version. Firefox LTS or something. Otherwise business users are going to drop Firefox like a hot potato.
Edited 2011-06-28 13:25 UTC
Not quite “no-one” – there are quite a few Mozilla folks who do get it, and are saying much the same thing as you. Unfortunately, there are also a lot who either don’t get it, or who just don’t care.
Exactly. I mentioned this on one of the developer’s blogs, and was told that the changes weren’t a secret, and that I should have been paying closer attention. Which is funny, because I *do* pay close attention to Firefox development, and while it was clear that there wasn’t going to be such a gap between 4 and 5, it certainly wasn’t clear that we were talking months.
Also, what about the Linux distributions? Does Mozilla actually expect Linux package maintainers to keep up with this aggressive new release method? Hell, Ubuntu was still on Firefox 3 for a very long time after Firefox 4 was released. So this new method also also gives a great big middle finger to the Linux distributions that package Firefox and update it through their update managers, such as apt, or yum.
I can see tomorrow’s headline already. It reads something like this “Bad Management Decisions Kill Netscape / Mozillia’s Browser…. Again…”
Edited 2011-06-28 13:26 UTC
I’m the only IT guy for a small business (100+ people). I detest IE, but it’s the only approved browser because we still deal with sites that only support IE. We even have a few that refuse to work on anything over version 7.
Rolling out other browsers just generally doesn’t work corporate wide for lots of reasons. Testing is one reason, but since I already know we have sites that don’t work it becomes pointless.
We are doing a test run of chrome frame through group policy. For sites that require IE we force them to render under ie and everything else under chrome. So far it’s going quite well.
I get what Mozilla is doing. Their target audience is end users and not corporate users. Given how Firefox works and such I can’t believe corporation are rolling it out anyway.
By how it works I mean that it Mozilla doesn’t actually make and network administration tools for it. You have to go through a third party for group policy administration and download a special version of Firefox. (I assume it still works that way. I haven’t looked in a long time.)
This may sound rude, but Enterprise sound like just another excuse for being slow. I understand why you have to be careful for server side stuff, but what’s really the problem with more web standards? Mozilla currently puts a lot of effort into making updates completely painless. They didn’t just think say “Hey, let’s change our release schedule” without having a look on technical implications.
For Enterprises: There IS support Firefox3!
The change HAS been announced, so every serious enterprise having problems with the new schedule should still be using Firefox3.
So what’s the problem?
Edited 2011-06-28 14:01 UTC
The problem is that Firefox stated plan going forward is that every release will be end-of-lifed in six weeks, meaning that there is no such thing as staying with an old version. Unless you are fine with security issues and bugs going unfixed.
As for “an excuse for being slow”, how do you mean slow? Upgrading browsers just for the sake of upgrading browsers is not a good value proposition for a company. Every upgrade involves a huge testing-, and usually fixing-, effort. This takes time and costs money, and doing it every six weeks would be ridiculously expensive. In any large organization you will have amassed a lot of legacy applications that no one is actively working on, but which are running fine as long as you don’t mess with them. Bringing someone in to test it is bad enough, but if the upgrade turns out to cause problems then you will have to have them familiarize themselves with the code base and set up a development environment to fix the problem (and then do an even deeper regression test).
I could go on, but I hope I made the point, this really is impractically expensive and cumbersome. Even if you had infinite resources it is not that easy to keep a six week test/fix release cycle for complex business applications, and if you fall behind you are in trouble. Not only the currently running browser is unsupported already, but the one you are working to validate for an upgrade to has already fallen out of support!
Still, of course Mozilla gets to use any release procedure they want, but it is clear that the current plan doesn’t really work for most enterprises. In that sense it is nice that Asa doesn’t pay lip service to enterprise support, both sides of the fence knows what the deal is this way.
if mozilla changed nothing but called the version 4.1 4.2 etc or even 4.01 no one would complain.
its what they have been doing for the past couple of years.
FF 3.6 and friends did bring a lot of features. even sub-sub releases such as 3.6.4 did !
True enough, and Firefox has never been very big in enterprise settings, but what this announcement did was quite thoroughly crush any hopes that Mozilla would take an enterprise-friendly stance. It is pretty notable that John Walicki at IBM, who has 500,000 users on Firefox, calls the announcement “a kick in the stomach”, since he was halfway through validating Firefox 4 for deployment in Q3.
Still, I maintain that I much prefer the way Mozilla is being clear about their plans and priorities to some lip service. Being a consumer browser is probably a more interesting and ultimately useful place to be than to cater to enterprise at the cost of mobility.
Web standards… Oh right. Those are those things that are supposed to be “standards” but in reality don’t work 100% correctly in any browser. Which means that yeah, careful testing of Web applications is required when new versions of browsers come out to ensure compatibility.
If you standards how much would be broken on each firefox release? IE yes but firefox?
I work on application using firefox on Ubuntu with basically 95% end users on windows using some version of IE. Most of the issues we deal are wrt to IE. Safari and Chrome are not tested at all but it does work on them without any change.
Extensions: There are extensions that do not need to be updated and do not even need browser restart on update. Its called Jetpack / Addon SDK / etc.
If you want that, use that (its what chrome does too btw). If you want more powerful extensions, no dice (chrome does not support that, chrome’s extensions are limited).
If you want to use FF with MSI, go right ahead they’re available.
If you want to use a LTS FF go right ahead with 3.x
If you think that minor versions make things stable, go get a clue. Minor versions of FF always included non-trivial updates, not just bug fixes.
If you want to blame FF because its old and it looks cool to blame it since Chrome is the new kid, you’re doing well. But that doesn’t make it accurate, or right.
Why isn’t “the enterprise” using site-specific browsers to deal with applications which require testing if regular browsers need updating?
It’s getting ridiculous that the progress of web browsers is being held back. I can understand that internal applications need to be well tested with particular version of a browser to ensure that there are absolutely no bugs, otherwise hours of productivity could be lost across an entire organisation if an update causes issues. It’s increasingly difficult to deal an increasing number of sites online which are dropping support for particular browsers (and to be fair, they have every right to do so).
More or less how it works, but the site-specific internal browsers still kind of need to be secure and possible to update if they break in some unexpected way. Which IE promises (IE9 will be updated until at least late 2020) but Firefox does not (Firefox wont be updated past August 2nd).
To some part this issue is overblown in that sense, IE will stay around since it is an enterprise-friendly browser, but for general-purpose browsing one could easily keep Firefox/Chrome around. It would be nice to have another good enterprise browser, and Firefox makes sense since it is so extensible, but if Mozilla doesn’t want to build things that way there is no reason to force them.
Well, site specific browsers kind of defeats the purpose of Web applications. If we are going to do that, we might as well go back to traditional thick client desktop applications. They are easier to program than Web applications anyway.
The browser vendors are the ones holding back progress though. Not the users. The browser vendors hold back progress because of incomplete / buggy implementations of standards. CSS is still a pain in the ass because some parts of it don’t work the same in all browsers. Same with Javascript. So we have to put in workarounds and hacks to make things work. And then when the browser vendor updates their browser, the hack breaks, etc.
And that’s exactly why IE6 is taking so long to go away. Except that because people don’t like having to use one browser for one thing and a different one for something else, they end up using the lowest common denominator for everything.
Rule #1: The use of logic and/or reason is NOT supported in the Enterprise. For further details, please see Rule #1.
Working in the enterprise, I can tell you I see old software everywhere. Operating systems, DBMS’s, browsers, JREs, office suites, compilers, libraries, you name it. I still see DOS stuff.
The same applies to hardware – I’ve seen Alpha servers dismissed recently, and I know of a case of a 1970 computer dismissed in 2003 (note that the first microprocessor is from 1974).
That’s because after 40 years of using computers the enterprise systems have become huge, instable, mysterious behemoths, where every little change may cause everything to fall apart. Who would accept such a responsability? That’s why every little change has to be designed and tested carefully, and this is costly.
That’s so simple.
Yes, I know it shouldn’t be like that, but – as you can guess – it’ll take long time for enterprises to change.
Edited 2011-06-28 15:17 UTC