Sort of a shame that the move over to Firefox 3.5 means that a version of Firefox won't really have a chance of becoming the most popular browser version on the web for say a week or so.

Fx 3.5 could be on the market for a long time, maybe it'll become the most popular (misleading term btw) browser version sometime between this and next year :)

I don't want to get into the relative merits of the different stat trackers (although I agree with Asa) but if you look at StatCounter, Firefox 3.0 was already, briefly, the most used (less misleading term) browser on the web and would continue to be if not for the release of 3.5. Perhaps if the transition is quick enough, we could see it happen again with 3.5 before IE8 takes over the top spot.

Asa, how much of IE8s uptake is from Win7 betas? I sadly know a few techies that have been running it for a while…
Looking at the totals cumulatively, am I right to think IE is now holding ground? Would the new 7 installs be masking the trend to leave IE?

John, I don't think there's any measurable IE 8 usage on Windows 7 betas. They just don't have the reach. Remember, to get even a couple of points of global share, you have to have several tens of millions of users.

And I don't think IE is holding ground. I think it's still falling and has lost about a point of total share since IE 8 shipped if you look at Net Applications data and about 3 points if you look at StatCounter.

I think the truth is probably right about in the middle of those two accountings and IE has lost about 2 points in the 4 months since IE 8 was made available.

This is right in line with IE's average monthly loss over the last 5 years. The long-term trend is half a point loss per month. A couple of years have been better for them, a couple worse, but overall it's a steady downward trend.

- A

any comments to Google's chrome bug reporting?, that's "hidden till everyone upgrades".. considering the recent Firefox security bug was completely open and transparent to all users.

http://googlechromereleases.blogspot.com/2009/07/stable-beta-update-bug-fixes.html

fanboy: Mozilla security bugs are normally hidden until either they've released an update for a few days or everyone knows about it anyway.

fanboy, that's pretty far away from the topic.

I'll just note, though, that we do our best to keep security bugs limited to the group of people who work on security issues, have a need to see the particular bug, or are distributing products which might be impacted by the bug. We don't, as a practice, keep these bugs open to the general audience and we do work very hard to get our updates into the hands of as many users as possible before releasing the details of the flaw to the general public.

I don't think that's very different from what Google is saying.

- A

And continuing down the off topic road, it's not a big deal, I don't think, to keep the specific exploit details restricted to the security group for the 5 or 6 days it takes to migrate ~95% of our users to the new safe version.

Now, when someone decides to release the exploit details to the public, rather than to Mozilla's security group, then there's nothing we can really do except push the pedal to the metal to get that fix to users as quickly as we possibly can.

- A

Asa, why is it that Mozilla still haven't pushed out a major update for Firefox 3.5 to all users? It's a great product, a great upgrade to all. Wouldn't you want it installed as quickly as possible?

I know it's possible to upgrade by clicking "check for updates..." (or just downloading) but it seems like a missed opportunity.

Chris, a couple of reasons. First, we're trying to drive a little bit more stability into Firefox before we unleash it on all 300 million of our users. That means fixing a few more bugs and shipping Firefox 3.5.2 (would have been 3.5.1 if we weren't zero day'd.) Second, we're hoping that more add-ons will become compatible before we tell everyone to get the new versions.

This is normal Mozilla process. We ship a new version, and get feedback from all the "early adopters" who regularly seek out and download new software. We respond to that feedback, fixing any stability or security issues that come up, and then we ship a dot release. Once that dot release is vetted, ideally no more than a couple of weeks of it being in use, we can turn on the notification for all 3.0 users.

There's no missed opportunity here since we don't depend on our "release buzz" to get the attention of most of our users. We have a nice built in update mechanism that we can use for "direct contact" and so we can afford to wait a month or two after a major release before we push everyone to get the new version.

- A

Yeah, it'd be a shame to give some users a bad impression of 3.5 due to pushing it too early, before having fixed bugs / regressions that may be important to them. The first few revisions of a major release are usually the most important ones, so it's better not to rush the major update.

Asa: Any chance you would put a title over that graphic? It was not completely clear to me if the Y-axis was % worldwide or % of IE-only. I started forwarding some of your posts to my current employers' product group. Some of the readers are less familiar with this topic, and that change would help a lot.