I've been working on my Firefox 3 review lately and one thing I'm a little bit concerned about is the reliability of this product. As many of you know, Windows Vista includes an amazing tool called the Reliability Monitor that monitors the reliability of the system and provides an ongoing reliability score, based on such things as software installs/uninstalls (which are noted, but don't affect the score), application failures, hardware failures, Windows failures, and miscellaneous failures. A failure of any kind affects the score dramatically; too dramatically, in my opinion, but give Microsoft credit for taking this seriously. If you have a day without issues, the score inches back up, but slowly: A day without issues is not seen as a big deal, whereas a day with even one issue is quite problematic. Put another way, if you have a single failure on one day, it will take three days of utter perfection to get your score back up to the previous pre-failure score.

OK, so what does this have to do with Firefox 3? Well, I've been testing pre-release versions of Firefox 3 since, well, there have been pre-release versions of Firefox 3. And on my daily-use desktop, which was most recently brought online with Vista Ultimate x64 on April 8, 2008, almost two months ago, I've been using Firefox 3 as my default browser all along. Though I don't feel that this system has been unduly unreliable, the Reliability Monitor tells another story, as seen below: The system's score, today, is 4.80 out of a possible 10. That sounds horrible. And maybe it is, though I don't think of this PC as particularly problematic.

Looking over the data, I see the following results:

Software Un(Installs): n/a
Application Failures: 16 days out of 51 days; 24 actual failures
Hardware Failures: 0
Windows Failures: 0
Miscellaneous Failures: 1 (Disruptive shutdown on 4/23/2008)

Clearly, the majority of issues were caused by applications. Here are the applications that failed over this time period and the number of days in which they failed:

vlc.exe 1
firefox.exe 12
ieuser.exe 1
Amazon Unbox Config.exe 2
safari.exe 1
pi.exe 1
photoshopelementseditor.exe 2
quicktimeplayer.exe 1
skype.exe 1
iexplore.exe 1
realplay.exe 1

Looking at this list, a few trends emerge:

  1. Fully 50 percent of all applications failures were caused by Firefox 3.
  2. Microsoft was only directly responsible for 2 of the 24 applications failures, and both were IE related.
  3. Some applications I very rarely use (Safari, RealPlayer) somehow managed to crash anyway. Good stuff.

Looking more closely at Firefox, I have to wonder if there are mitigating factors. For example:

  1. It is pre-release software after all, and I've upgraded beta-over-beta a few times in this time period.
  2. I have installed a few add-ons over time as well, and Firefox add-ons can be buggy especially when running on the next version of the browser. I currently have the following add-ons installed in Firefox 3:
    Glasser
    Hide Menubar
    Prism for Firefox
    Stylish
  3. Looking at my Wakoopa profile, I use Firefox more than any other application, by far. Using the "last month" stats, Firefox accounts for 39 percent of my PC usage, compared to 30 percent for Microsoft Word (which hasn't crashed once, naturally), 5 percent for Windows Live Writer (no issues), and 4 percent for Windows Explorer (also error free).

So. What am I to make of all this? Nothing? Or is Firefox 3 really buggy?

I'm going to try and test the browser without any add-ons for a while and see how that affects things. But this really has me wondering. Has anyone else seen huge problems with Firefox? Or I am I just seeing a side-effect from a particular add-on?

If it is add-on related, however, where do we point the finger? Is Firefox's single greatest feature--it's extensibility--really an Achilles Heel?