There's a lot of work underway now to improve cycle collection times, which is where many of the pauses come from. Also, work is underway for both Generational and Incremental GC, which should improve things on the GC side. At least with a rapid release schedule, those improvements will ship when they're ready rather than waiting for other things to finish up first like they would have in the past.
bz pointed out elsewhere in this thread that while ESR releases will get critical security updates backported to them, they won't be receiving any major architectural updates/refactorings that inherently improve security. So in that respect, ESR releases will be more limited security-wise than the mainline release.
You'll be happy to hear that Asa isn't the only decision maker at Mozilla, nor is his opinion held at any higher esteem than anyone else's there. Nice troll, though.
Actually, google said, "Mozilla, can we work with you to make firefox radically better?" and
Mozilla said "no, we have our own ideas and we don't want you telling us what to do!" and
so google created Chrome with the goal of forcing all the vendors to make their browsers better.
It's still supported on an opt-in basis. But given that it was prone to numerous security holes and was unused by the vast majority of the web-facing population, the decision was made that it made more sense for those who need it to explicitly turn it on (whitelisted, no less) rather than exposing all users to the risks that come with it.
Mozilla has a pretty extensive automated test infrastructure that runs on every checkin. Also, the major work is done on different project branches and made stable before landing on mozilla-central (the mainline Nightly branch), hence the better stability of the Nightly builds.
I can't speak for the entire college of engineering, but at least in my department, it was typical for the BS (or BS/MS) to take 5 years, with 2 or 3 co-ops depending on whether you're getting the BS or BS+MS.
That sounds like an incredibly fragile idea. That said, a lot of work is being done to add Type Inference to the JaegerMonkey engine, which performs static analysis to determine type information.
Also note that Win2K support will be going away once the switch is made to start compiling with MSVC 2010 instead of 2005. Win2K support in Fx4 is more by accident than by design at this point.
Slashdot Mirror
See bz's post and subsequent reply: http://news.slashdot.org/comments.pl?sid=2615802&cid=38664528
There's a lot of work underway now to improve cycle collection times, which is where many of the pauses come from. Also, work is underway for both Generational and Incremental GC, which should improve things on the GC side. At least with a rapid release schedule, those improvements will ship when they're ready rather than waiting for other things to finish up first like they would have in the past.
It doesn't make a lot of sense to backport big architectural overhauls to long-term stability releases, no.
bz pointed out elsewhere in this thread that while ESR releases will get critical security updates backported to them, they won't be receiving any major architectural updates/refactorings that inherently improve security. So in that respect, ESR releases will be more limited security-wise than the mainline release.
You'll be happy to hear that Asa isn't the only decision maker at Mozilla, nor is his opinion held at any higher esteem than anyone else's there. Nice troll, though.
So you're going to switch from one browser that rapid releases to another browser that rapid releases, over......rapid releasing?
For the umpteenth time stated in this thread, Firefox 10 defaults to addons being compatible.
Firefox 10 defaults to addons being compatible by default, FYI.
You do realize that you're replying to a core Firefox developer, right?
So a license update to make the MPL simpler and *more* compatible with other open source licenses is somehow discouraging re-use? Do explain.
Actually, google said, "Mozilla, can we work with you to make firefox radically better?" and Mozilla said "no, we have our own ideas and we don't want you telling us what to do!" and so google created Chrome with the goal of forcing all the vendors to make their browsers better.
[citation please]
How exactly would Microsoft "gobble Firefox up"? It's an open-source browser put out by a non-profit organization.
It's still supported on an opt-in basis. But given that it was prone to numerous security holes and was unused by the vast majority of the web-facing population, the decision was made that it made more sense for those who need it to explicitly turn it on (whitelisted, no less) rather than exposing all users to the risks that come with it.
(Hint, it is not Chrome, Chrome gets people out of Firefox mainly.)
I would disagree with that statement. While Firefox has lost a bit of market share to Chrome, most of Chrome's gains have come at the expense of IE. Look at the trends.
http://arstechnica.com/business/news/2011/12/internet-explorer-stops-its-slide-as-chrome-nears-firefox.ars
Mozilla does a lot more than just make a web browser.
Mozilla has a pretty extensive automated test infrastructure that runs on every checkin. Also, the major work is done on different project branches and made stable before landing on mozilla-central (the mainline Nightly branch), hence the better stability of the Nightly builds.
Except that there's an Adblock Plus extension for Chrome too...
Those hangs are probably GC pauses, which ironically would be tied to memory usage :-)
Actually, Firefox 3.5 hit EOL (at least from Mozilla) as of the last security update. They will not be updating it further.
I can't speak for the entire college of engineering, but at least in my department, it was typical for the BS (or BS/MS) to take 5 years, with 2 or 3 co-ops depending on whether you're getting the BS or BS+MS.
I wouldn't expect two people to skew a median value significantly when dealing with a population size many orders of magnitude larger than them.
Yay Drexel! (MatE alum here). Surviving the first two years sucks, but it got better after that IMO.
That sounds like an incredibly fragile idea. That said, a lot of work is being done to add Type Inference to the JaegerMonkey engine, which performs static analysis to determine type information.
Same place as always: http://nightly.mozilla.org/
Also note that Win2K support will be going away once the switch is made to start compiling with MSVC 2010 instead of 2005. Win2K support in Fx4 is more by accident than by design at this point.