In the Editor's Note for this week's Linux Pipeline newsletter, I praised Mozilla for fixing the memory-management bugs that had plagued Firefox for such a long time, but which no longer seemed to be a problem in Firefox 1.5. Today, if I could add just two words to that Editor's Note, I would call upon the immortal wisdom of Rosanne Rosanna Danna: Never mind.Firefox 1.5, as it turns out, still eats up system memory the way my dog eats his dinner: No matter how much he gets, it goes fast and it's never enough. I got a vivid reminder of this two nights ago, when I gawked in amazement as Firefox tore through more than 400MB of physical RAM and virtual memory on a Windows XP system.

That's 400MB apiece, and that number continued to climb even after I took the Windows Task Manager screen shot you'll find in Scot Finnie's feature detailing this and other serious problems with Firefox 1.5. I finally killed this voracious little pig after it had gobbled nearly 500 MB of RAM -- half of the total memory on this system -- and another 500MB of virtual memory, brought a cutting-edge system I had built just a few months ago to its knees, and then finally quit responding.

Scot discusses this and other problems in his article on Firefox 1.5, including some Javascript problems that are also capable of taking out both the browser and, in some cases, sucker-punching your operating system into a hard restart. Curiously enough, even as Firefox devoured memory on my system, it scarcely touched my CPU, using 10 percent or less of its capacity right up to the point it quit responding.

This is the nature of these problems with Firefox: They can be spotty and difficult to diagnose. As Scot and I both know, based upon our own research and conversations with other Firefox users, many of you experience these problems with Firefox and consider them a serious nuisance -- yet others, a small but significant minority, have never seen any of these problems.

In fact, I have heard before that because these bugs can be slippery and hard to pin down, Mozilla developers can't be expected to do anything about them. Some people even go so far as to deny that these are even real problems, since you can expect a certain amount of undesirable behavior in any widely-used software created today. That's true, but it's also absolutely irrelevant to these problems, which date back to the pre-Firefox 1.0 days and have spawned a number of Bugzilla entries along with countless complaints and debates in various online forums. These bugs are either the biggest collective hallucination in history -- along with all of the Task Manager, Activity Monitor (for Mac OS X) and "tops" output (on Linux systems) screen shots that show the same things I saw two nights ago -- or they're among the most persistent and well-documented software bugs to appear in any mainstream software product, open-source or otherwise.

I like Firefox, and I support Mozilla, just as I support other open-source software for its many political, intellectual, and economic advantages. Yet Firefox has reached a point, with its climbing market share and Mozilla Corp.'s nascent efforts to give the browser a real corporate IT presence, where these types of problems can't keep hanging around like some sort of extended family secret that we all hope to see resolved someday. I can also guarantee that Microsoft is unlikely to let these types of problems slide by when it releases Internet Explorer 7, which by all accounts will give Firefox a real challenge in its Web standards support and other features.

Software bugs shouldn't be allowed to celebrate birthdays. It's time to deal with these problems, so that Firefox can move ahead with quality code and a clean slate.

Update: What About Extensions?

A reader posted a comment below, asking whether a Firefox extension could cause the memory-management problems I cite here. It's a good question, since it's my understanding that at least some extensions can cause memory leaks.

In this case, however, I didn't have ForecastFox or any other extension installed that is capable of sucking up 800 MB of combined memory without so much as a burp.

This does, however, raise another issue: Whether Mozilla should catalog extensions that exhibit such behavior or, indeed, allow extensions to run without any safeguards against this sort of problem.

- Updated 12/09/2005 15:51pt

As Scot notes in his article, we'd like very much to keep hearing from all of you -- whether or not you've encountered these or other problems with Firefox 1.5. Drop me an email detailing your own experiences, problems, and any other thoughts you'd like to share.