I got a terse response not long after my review of the second Netscape 6 preview release demanding to know why I'd "kept the kid gloves on" where the package was concerned. As far as the commentator was concerned, it was apparent that a "boost Open Source Software at all costs" agenda was involved in the article.

The review read, in part:

"Netscape 6 will not present itself as usable to people in need of a stable browser. It has bugs, it misbehaves, and it disappears from the screen over mysterious vexations known only to itself. If casual users can spend an hour of worry- (and crash-) free browsing, we're happy for them."

Clearly a gloss job aimed at boosting a high-profile project at the expense of accuracy or fairness.

I will confess, though, that I've wanted Mozilla to succeed like few other projects I've tracked over the years. That need isn't derived from Open Source boosterism as much as it is the simple pragmatics of getting work done.

Most of my work day is consumed by time in front of a browser. Even if I had Windows and IE on one of my work machines, I'd have to use Linux to do my job. Using a Windows-based browser risks missing moronized HTML or directing readers to a site that isn't going to render with Netscape running under Linux. It's important to me that Mozilla work and work well. The alternative is spending the rest of my days working with the "finished" Netscape 4, which also crashes a couple of times a day and shows its age daily.

So for the past several months, since I acquired the bandwidth to do so, I've been regularly grabbing the builds from mozilla.org's download area.

My workday has involved a "Mozilla session" for that long. I unpack the tarball, fire the browser up, and work until it dies. For a while, it crashed when I tried to use sites that required authentication, which kept the sessions short. For a while longer, until I could locate the bug that explained how to fix the problem, use of the browser with Junkbuster eventually caused the navigation buttons to become confused. Up until a week ago, it mangled CGI form submissions.

In addition to the "show-stoppers" that kept it from being usable at all, there are the random crashes and the sense of pokiness from the interface itself that made it frustrating. My rule, though, has been to use it each day until it breaks.

Since the M17 release, though, something pretty cool has been going on with Mozilla: it gets noticeably better on a daily basis. Not the sort of "better" that had people calling it "rock stable" six months ago, when it clearly was not. Rather, it's the sort of "better" than finds me using it continually for most of the day. The sort of better that involves enjoying it enough that I break my rule and fire it back up again if it crashes two or three times in a day. This improvement isn't be a big surprise. The game plan called for a push on optimization and bug-fixes after the M17 release, and the Mozilla developers are making good on their plan.

Stability isn't the only area where it's improved. The interface itself is more responsive and faster. Some of the simpler features like editing bookmarks work more reliably, and setting preferences isn't an exercise in futility.

It still acts like beta software here and there, but I can use it for work, and I find myself much more distracted by the slightly different keybindings and occasional differences in behavior than any shortcomings of the package itself. It's clearly down to hunting bugs now, and I hope those who have held off from dealing with Mozilla on a daily basis will reconsider their abstention and start pushing the software hard. The end user bears some responsibility to help the developers figure out what's left to deal with in the home stretch.