DiCarlo: Everybody hates Firefox updates

That's already know, it was formalised by Jakob Nielsen who said something like that "unless something is a 100% improvement over the previous version, it should not be changed" ( not sure of the exact wording ).

However, that's also contradicting the nature of free software, ie that's open innovation. So people do see version that are not perfect, because we do not restrict them from using ( or they would not give feedback ).

So unless you advocate to never change anything ( cause every UI change requires to relearn it ), this cannot be realistically applied to free softwar. IE, people complain when it change too much, when it change too often, and when thing are broken and do not change.

Basically, what people want is something right from the first version. And that's not how it work, not only for free software, but for most software, no one can be right at the start, or at least, not without being wrong a couple of time before. But you can have a pony if you want.

I vote for free tattoos! Seriously, I'd put money into sponsoring that.

Could we please put him in charge of GNOME?

The basic rule is that you can change your internals at a fast pace, but breaking interfaces with others is the ultimate sin. That's just as true for the kernel as for the interface with users.

Sadly that's not something the average developper wants to hear

Yep, I'd say this is by far the biggest issue. And it's not like this caught anyone by surprise, it's been obvious that the reason Chrome can pull off rapid updates is due to its /unobtrusive/ updater. Firefox, even before the rapid release cycles, already had issues with its updates. Firefox for some reason had completely ignored UAC and non-admin accounts on Windows (I guess presuming everyone would disable UAC and run as an admin?), because for the longest Firefox's updater would crap out in that situation. Even now, it still doesn't work if you try to use it from an instance running as an unelevated Standard User (it fails silently as well, even on 13.0!).

The only way Chrome's updater could be topped is if a browser figured out a way to perform them without restarting at all (with a stable inter-process interface and some cleverness I'm sure it could be pulled off, although it may or may not be worth the effort).

"However the difference is Chrome updates are transparent."

which I find quite offensive and is one of the reasons I will never run chrome.

As far as I know, the problem is that there really is no API for extensions. They have access to all of Firefox' internals, thus keeping the API stable would mean to stop Firefox development altogether.

This doesn't apply to the extensions built with the Add-On SDK
https://addons.mozilla.org/en-US/developers/builder
So let's hope that most Add-Ons are migrated to it in the foreseeable future.

Perhaps you are not using Garmin plugin on Windows.

Firefox 13 and NewTabUrl, for example.

> Your users do not "love" your software. Your users are temporarily
> tolerating your software because it's the least horrible option they
> have -- for now -- to meet some need. Developers have an emotional
> connection to the project; users don't

I think developers should aspire to making software that is lovable by users. The broader the audience, the harder it is to make a large percentage of users love your software, but still, should we really give up? I love vim and gimp, without having donated a speck of code to either one.

I love a lot of the software I use.

... although, I'm not sure how much of it I'd still love if it changed non-trivially!

Occasional breakage of plug-ins or sudden displacement of Tabs to top, or the awesomebar behaviour changes were not enough for me to start using Chrome, thanks to all the workarounds. People get emotional about what they love, not what they hate. Rapid UI changes do drive people away from the browser we love.

I'm also happy with the new Firefox release model. There's nothing intrusive about 'yum install firefox -y' for me.

> Am I the only one who remembers what the *old* Firefox release process was like? You'd wait 2 years for the next version, which was then guaranteed to break *all* of your extensions. And it took at least a month for the extension writers to catch up. The new release would always come with show-stopper bugs, crashes, etc. and would stabilize in a few months.

That is nothing. I remember an even older Firefox, back from before it was called Firefox. You had to run one single window at a time, else it would crash too often. Even with a single window, it crashed often. Tabs? There were no tabs back then. I do not think we even had extensions. The versions were also a single number, only it had a "M" in front, and were also released with a rapid schedule.

*insert Four Yorkshiremen sketch here*

Indeed, I have been thoroughly enjoying the rapid releases. I can finally use the stable versions.

As for GUI changes, I use the vimperator extension, so I have no idea what any of you are talking about. :)

The Firefox ESR page makes it quite clear that the ESR is not intended for 'personal' usage (only for organizations). I also couldn't find any links to the download from the main ESR page (there ARE two forms to sign up for the mailing list, though!).

I'd imagine most people also haven't heard of the ESR, anyways, since it's fairly new, and is a reversal of Firefox's long history of quickly abandoning releases.

As for distribution packages? Not an option for Windows/Mac users. Distros often lag behind upstream's major versions as well.

If I _used_ Windows I'd stop using Firefox.

On Fedora, I get new Firefox updates along with everything else, and the (easily dismissed) dialog telling me I need to restart everything. That's the only sane way forward.

Most of the software running on Fedora in fact tolerates not being restarted until its convenient. Firefox mostly works but a handful of things break, the built-in search used to be one of them but maybe they fixed that a while back. But you don't lose your open pages, half-filled forms, etc. and it doesn't interrupt what you were doing to demand an update.

On Windows every tuppenny-ha'penny program has a separate "updater" that runs at the worst possible moment, demands focus, then insists you fill out forms acknowledging that you now agree to be sued in Las Vegas if you think bad things, and finally restarts the entire computer. And that's if it works at all because (as somebody pointed out above) a lot of stuff can't handle the idea that a Windows user might not run constantly as Administrator, so it breaks.

"In other words, the problem now seems to be that people's perception is that firefox has remained behind and is playing a catch up game."

It was probably due at one point of time when firefox 4 was being pushed out the door. Afterwards that, they started the "be more like Chrome"-trend, while simultaneously avoiding any mention of word Chrome in public communications, and I read between the lines that they themselves thought they had fallen behind.

The rapid update process was imho a smart thing to do, but Firefox just could not work out how to shut up about it, and/or do it without breaking those extensions everybody seems to rave about. (I use none, personally.) In this, they failed to be more like Chrome, sadly. If Firefox knew how to update itself without showing the process to end users in any way whatsoever, they would reach the user experience parity with Chrome in that. Chrome has the pleasant apt-get/yum -like usage experience in that it appears capable of updating itself while still running, and shows the new version a quick restart later. There's literally nothing that interrupts users's workflow.

The reason people do not install the "enterprise" version of the firefox is that they don't even know it exists. If you go to mozilla.com/firefox/, you can only see a "Get firefox" button. (And yes, this is how it should be, also.) The normal response people have when they are unhappy with a product is that they look for competitor's product. So is it any surprise that they pick another browser? It's not like there's any shortage of alternatives. I don't use Chrome personally because it renders text incorrectly on OS X -- a bug I recently noticed when doing some text rendering tests -- while Safari renders correctly and is probably 99 % the same engine.

The reason I think some people to start getting fed up with open source development because the incessant churn in the stack that requires programs to be rewritten for no visible user- or developer-related benefit. It soured me when I was happy with KDE 3.5 and had to face the disaster that was 4.0. It was when I realized that in the past 10+ years that I have used Linux, it's been slow steps forward followed by an instantaneous giant leap backwards. And I gave up, thinking that this is how it always would be. For a brief time, I tried using Windows, but realized that the environment was simply too far away from the sort of things I needed to do---it required too much customizing to allow decent command shell and tool suite. OS X, on the other hand, is perfect.

>I find it hard to believe that people moves to chrome due to the fact that
>the rapid release process breaks their extensions, since they will probably
>not find those same extensions in chrome.

That's one of (if not the main reason) why I moved. By the time I moved, a lot of the extensions I used to use no longer worked. Of the ones that remained, some weren't necessary in Chrome (tab handling is nicer out of the box), most had equivalents (ad-block), and the one or two that didn't I could live without.

In the case of my wife, it seemed (to her) that every time she turned on the laptop Firefox wanted to update itself. So she's now on Chrome too - now, no more problems.

Same for my parents.

Regarding the long-support release, I didn't know it even existed, let alone where to find it. If I didn't, there's no reason less technical people should. You're also assuming that people have an attachment to Firefox. By this time, they don't - they've got fed up with the restarts, and they're already looking for something that's not Firefox. As the blog entry says - developers have an attachment to the software, users don't.

That should read:

This issue with updates will largely disappear once they are installed without user prompting by default (windows).... so by the end of the year :-)

I recommend to update somewhat newer "stable" browser from backports.
Please try Iceweasel ESR (now at version 10).
http://glandium.org/blog/?p=2573
http://backports-master.debian.org/
http://mozilla.debian.net/
It works fine for me and has many useful new features.
3.5 is now quite ancient and even security updates are hard to make when upstream don't support 3.6

Now that I've caught up on all the comments, I *am* sure that lots of extensions fell over dead. My favorite tab wrangler was one of them; on a 12" laptop, that's pretty critical...