Mozilla provides more details on Firefox 3.0 RC2

The official release notes are now live on Mozilla's servers, along with Release Candidate 2 of the organization's next Web browser, which was published yesterday.


Download Mozilla Firefox 3.0 RC2 for Windows from FileForum now.


The festivities began yesterday -- two days ahead of schedule -- with the publication by Mozilla of a package that ended up being confirmed as Release Candidate 2 of its Firefox 3.0 Web browser. The final release certainly appears on schedule for "mid-June," which is true to what the organization has promised users for several weeks.

Thus far, BetaNews tests have shown RC2 to be mostly solid. We noticed the absence in the new custom installation option of DOM Inspector, a tool developers use to look into the code for a page's Document Object Model (for pages that use DOM). There's a good reason for that, as confirmed this morning in Mozilla's release notes: DOM Inspector was separated from the 3.0 code base, and is now available as a free add-on.

Upon installing the add-on this morning, for reasons we haven't been able to trace so far, RC2 crashed. This was with a full complement of Web pages spread over two open browser windows. Upon recovery, however, RC2 pulled up all our Web pages and so far is running the new DOM Inspector add-on smoothly.

What we have consistently noticed in our tests of Firefox 3.0 versions for Windows leading up to RC2 is the absence of easily our least favorite feature of Firefox 2.0: the memory leak. In almost 100% of our trials where two or more FF2 windows are loaded with a dozen or more pages each, and at least one of those pages in each window has active JavaScript, we can watch the leak happen in front of our eyes in Task Manager.

Ever since Beta 3 of FF3, we have not seen any evidence of the same leak, and this fact alone lends substance to Mozilla's claim that any of its FF3 pre-releases could easily replace FF2 today for most users. (Those who prefer a multitude of customizations and add-ons will still need to wait a few weeks for them to be officially approved for FF3.)

But this does not mean FF3 doesn't produce a big memory footprint. Granted, I'm the heaviest Web browser user there could ever possibly be, but on an average day, my active FF3 pre-release consumes up to 400 Mb of memory in my 2 Gb production system. However, apparently due to FF3's improved memory allocation, that footprint is stable. I can overload FF3 with pages with active scripts and even running videos, while noticing far less of a performance hit on other running applications than with FF2.

A few years ago, my friend and colleague Angela Gunn wrote for her USA Today blog (the evidence of which has apparently been wiped clean from the face of the planet) that Firefox 2 was the "worst upgrade ever." To paraphrase Pres. Johnson, if you lose Angela Gunn, you've lost the everyday American user.

But in the same way that Windows 98 SE was easily the greatest Windows 95 upgrade ever produced, there's a very genuine possibility now that Firefox 3 will be the best Firefox 2 upgrade in human history. We'll keep testing, of course, and we'll let you know more.

Viewpoint ribbon (small)

Yesterday, a few of our regular users asked whether BetaNews jumped the gun by posting a link to Firefox 3.0 RC2, when an openly posted schedule showed the QA for the product was supposed to be released today. In other words, why would we enhance the availability of an already available product before the company's posted schedule says it's no longer suspect?

My favorite comment was this one: "Does anyone have an address at hand to get my dog's excrement routed to SF3?"

Well, here's the deal: I like to leave it to each individual dog owner to handle his own excrement problem in his own way. In other words, when you make something public, that's publication. At the risk of sounding like Judge Neil Wake's decision in the KaZaa case in Arizona before he himself overturned it, if you place something in a public folder, it's generally with the intent that the public uses it. And when that public folder is clearly marked "RC2," well, it's usually for a good reason.

Usually. That said, we've been wrong on this one before. Last November, when it looked like Firefox 3.0 Beta 1 was released, it actually wasn't, and Mozilla had to take it back. Then last March, our FileForum posted a link to what appeared to be Beta 4. But what ended up being posted in the proper Beta 4 directory on Mozilla's servers wasn't Beta 4, but instead a nightly build.

The whole matter has prompted many to split hairs even more than they've already been split. It was a publication, for instance, but was it a release? If a publisher makes something public but doesn't publicize that it's made it public, is that a publication?

That type of hair splitting makes me want to go into the dog excrement-routing business. Here's how I look at it: There's all kinds of information that's shared with us at BetaNews at a time prior to when the producers of that information want to make it public. And we honor those requests, at all times. It's part of doing business. Information, I sincerely believe, can only be free if you're willing to pay the price.

But a Release Candidate is, by definition, a product meant to be tested -- which already means that using it comes with risks. We believe Mozilla is a smart enough organization to be able to handle the mitigation of those risks, which means we trust the organization to be able to make the decision for itself about when something's ready for the public, without us having to second-guess it. And if they're wrong, well, they're wrong, but that's their business.

In this case, as it turned out, they weren't wrong. It was RC2, it was released intentionally, and no one -- dogs included -- had an accident.

We're in a risky business. We are risk-takers by nature. That is why, to paraphrase the philosopher Chevy Chase, we're BetaNews and you're not.

-Scott M. Fulton, III, Executive Producer, BetaNews

© 1998-2019 BetaNews, Inc. All Rights Reserved. Privacy Policy - Cookie Policy.