“One of the most popular things about firefox 3 is that the plugins provide so much useful extended functionality.”

Except that there are about 200 million Firefox users and only a tiny percentage of them use any extensions.

To claim that no comparison is fair unless Firefox is loaded up with add-ons is to suggest that you must compare the worst, and tiny minority case scenario and that’s simply wrong.

Should performance tests be for the most common case or the least common case. That’s my question to you. If it’s the former, then the no add-ons case is the best comparison.

And it’s not even close. The most popular add-on gets used by about 5% of Firefox users. It’s likely that only around 10% of Firefox users use any add-ons—with most of them using only a couple of well tested and popular ones that likely don’t cause significant performance problems.

You’re setting up a poor comparison based on your power-user usage of Firefox that doesn’t hold for the overwhelming majority of Firefox users out there.

Mozilla is up to 65 million Active Daily Users which translates to about 200 Million total users (at least, the multiplier might be a bit higher.) See Mozilla CEO’s blog post for how that works out http://john.jubjubs.net/2007/11/27/mozilla-firefox-market-share/

Or, if you don’t want to take Mozilla’s word for it, how about the this study from Google’s logs, that puts Firefox users at 16% of 1.4 billion total Internet users (I’ll let you do that math.)

What Download Day users has to do with this I haven’t a clue. You said “I’d love to see the test results in a common configuration rather than minimal configuration with a handful of the most common plugins installed.”

So, “common configuration” to you means the setup from the most enthusiastic early adopters and power-users. How about not changing your argument mid-stream and sticking with your original suggestion.

There are about (maybe a bit more, maybe a bit less, depending on the source you trust) 200 million Firefox users. The most “common configuration” out there is without a doubt zero extensions. It’s not just the most common, it’s the overwhelming majority.

Apparently learning that doesn’t fit with your argument very well and so now you’ve changed it and you want to talk about the 4% or so of users who downloaded Firefox 3 immediately upon release.

I’m happy to concede that a good chunk of those users are ones most likely to be running with one or more add-ons. That’s also the group that’s most likely to have done other tweaks to their Firefox like disabling add-on compatibility checking, installing alpha and beta add-ons and plug-ins, and making any number of un-recommended about:config settings adjustments.

That’s the group that’s going to have the most problems but they are absolutely not “common configurations” and they are not representative of the overall universe of Firefox users.

>Given the browser integration I fear Google’s data to
>be somewhat flawed

I’m not sure what you mean? They’re measuring users as well as usage. Usage might swing a bit heavy for Firefox/Opera/Safari/I.E.7 users since those browsers (making up about 70% of the total usage) have built-in search and most people prefer Google’s search. But independent surveys put Firefox usage even higher. The most often quoted stats come from Net Applications which puts Firefox usage above 19% (and they predict 20% this month) of the total web . XiTi Monitor which polls Europe primarily but also globally puts Firefox near 35% usage in Europe and over 20% globally . Combine that with what Mozilla sees directly (about 65M active daily users at a 3x-3.5x multiplier for total users) and average them all up and it’s no stretch at all to see Firefox usage at or around 20% and users at or above 200M.

So, as you can see, it’s not just Google data, though they are probably the most representative since they have the broadest reach. It’s also Net Applications, XiTi Monitor, Mozilla, and others if you care to look them up. They’re all in rough agreement with the 200M number I gave you in my first comment.

I’d wager that numbers available at the end of this month show a percent or two more than what what you and I are seeing today and I think it’s probably a pretty darned conservative estimate 200M users. The Web is about 1.4B users and Firefox is accounting for 20% of the usage. I’m not gonna claim that Firefox users have “average” usage patterns in terms of visits and time online—that would put Firefox users at nearly 300M, but given how far beyond the “early adopters” it is with 20+% of the market, I’d bet that they’re not more than 50% heavier in terms of browser usage than IE+Safari+Others.

All that being said, the data’s of course imperfect and to get more accurate, Mozilla would have to engage in more deliberate tracking (as opposed to just counting daily update pings) but there’s widespread agreement that there are “a lot” of Firefox users out there, probably around 200M or so.

>However I’d bet beer on the fact that the plugin
>users group is far bigger than the about:config
>fiddlers but we digress.

I don’t think we disagree at all on that. There are probably 10-20 million extension users out there, and there are probably only tens of thousands, maybe hundreds of thousands of about:config fiddlers. It’s a total guess, but there are probably about half a million to a million alpha-beta extension users too (at least two of the most popular add-ons available for Firefox 3 today are only available in pre-release form right now.)

If you overlay all those on top of each other, there are a non-trivial number of power users out there who have created less stable or performant Firefox installs through not-common usage. Those people are over-represented among the first downloaders and the “blogging” or “twittering” types so the problem cases are going to be much more visible than the actual reality across the Firefox landscape.

>4% of the userbase is obviously worth worrying about
>as are the power users who are clue-full enough to
>log bugs once in a while.

Absolutely. And enough of the clue-full are reporting genuine bugs to Mozilla and those bugs are being actively investigate, triaged, and fixed. As for the couple thousand Twitter users who aren’t clue-full bug filers, the best thing for them to do is to get their Firefox install into better shape ASAP so they can go about their lives and for a very many of them the answer is to find out which of their add-ons or other tweaks is causing them problems and either locate an update, or an alternative, or do without it until one of those is available—or endure the bugs introduced by that add-on or tweak.

>Does that figure include flash and acrobat too?

No. Approximately 100% of Firefox users have Flash installed Not sure about Acrobat but it’s probably pretty high too. Right now, a somewhat disproportionate (to the general population) number of Firefox users are running a Flash 10 beta from Adobe which appears to be much better behaving for some and much worse behaving for others—such is life with betas.

>Since many of the resource complaints around resource
>usage are apparently plugins, how come there isn’t
>more in your face visibility or protection around that
>feature too?

Plug-ins? Or do you mean extensions? Plug-ins are mostly commercial products added to the browser to facilitate non-open web content types like Flash, PDF, QuickTime, etc. Extensions are Firefox-specific browser enhancements, mostly non-commercial, that can change anything about Firefox from the chrome to the content. Themes are Firefox-specific browser enhancements that can change the look and feel of the chrome of Firefox. Together, plug-ins, extensions, and themes are called “add-ons”.

The terminology is a bit confusing and often mis-used (even by me, occasionally,) but it’s important. Extensions can change anything. There’s literally nothing Mozilla can do to control what they can or cannot change and there’s little Mozilla can do except encouraging or otherwise making it easier for extension authors to follow a set of best practices when developing them. But that’s why they’re so powerful (when done right) because they’re not limited to a small surface of change. The downside is that if an extension wants to break Firefox, it can damn well do it.

Some extension authors are OK with the performance or resource trade-off they make for their added features. They may have a different perspective than you do or even than their users do. Mozilla can try to advice them on better trade-offs but if there was a perfect solution with no negative impact on any user, it would probably be a Firefox feature and not an extension

Also, Mozilla doesn’t have the resources to individually assist the thousands of extension authors out there. Hopefully the extension-using community will complain more when they find flaws in their extensions and authors will get better and the best extensions will rise to the top and the poor ones will fade away. Think of the Windows and Mac software world. Microsoft and Apple can’t individually manage everyone developing for their platform, but over time, users learn what’s good and what’s not. Those vendors cannot prevent people from writing poor software that can even slow down or crash their platforms. Firefox is in roughly the same boat except that it’s much easier to write for Firefox than for Windows or Mac since Firefox extensions are mostly constructed using the languages of the Web (though some do have compiled code too.)

Plug-ins are about the same. There’s a very specfic API for plug-ins so they’re supposed to be a bit more limited in their scope if they follow that API (though nothing stops them from using extension hooks along with their NPAPI plug-in.) But ultimately the plug-in is responsible for how it deals with its content. Mozilla cannot intermediate much between Adobe’s Flash plug-in and Flash content that Firefox runs across just as Windows cannot intermediate much between Firefox and the HTML, CSS, and JavaScript it runs across on the Web

Mozilla can work to make it easier for plug-in authors to build more performant and less resource hungry features, but again, that burden should really be on the plug-in authors since they have big companies and big budgets. When Flash is sucking, people should blame Adobe. When QuickTime is sucking, people should blame Apple. When Silverlight is sucking, people should blame Microsoft. All the while, Mozilla should be working with those vendors to help unsuck their products as much as it has the resources to do.

There are a few things I’d like to see Mozilla do in Firefox to make it more obvious what’s going on to the user.

It’d be nice if Plug-ins could fail without crashing the browser, for example. Just let the plug-in die or restart. That’d be nice. With extensions, that’s probably impossible since the extension can hook into anywhere in Firefox, including the code that would supposedly keep the extension from crashing Firefox

Mozilla could also do something in Firefox to identify when a plug-in was spiking the CPU or sucking abnormal amounts of RAM and alert the user, kind of like it does with a script that’s been running too long—letting the user decide to keep it running or kill it.

Finally, Mozilla can keep working on Firefox to get even more performance and lower resource usage (often though, the trade-off is exactly between those two—increase one, increase the other.) With super-fast JS performance and solid SVG and Canvas support, for example, plug-ins like Flash become less necessary. Think of Google Maps or Zimbra email. Those are done with JS and HTML rather than Flash, so everything Firefox does control about how it renders real open web content makes those better—it’s not about Adobe or Apple or Micrsoft. Blame and reward both belong to Firefox there since that’s its native content.