Honestly, the absolute best thing you could do for any dev, is lighten the footprint FF currently has on my ram, and continue to improve the api for add-ons/plugins. The more stable, extensible and fast FireFox is, the more useful I will find it for development. But if more kludge is added to it, and it continues to get slower/bigger, the less likely I am to continue to use it, regardless of what amazing plugins are available (FireBug, Web Developer Tool Bar, etc)
If you are just looking at what's best to add to the browser itself, just look at the plugin market to see what's in wide use.
Indeed, everything I've seen recently points to FF as having a slim lead over Chrome. It's just that for many of us, the last time we used FF was right before we switched to Chrome, so our perspective is frozen at that point.
Browser
Google Chrome 25.0.1364.97
Memory
Private Shared Total
3,950,040k 12,805k 3,962,845k
Virtual memory
Private Mapped
4,728,936 k 937,320 k
This is with 18 tabs and ~20 extensions. Chrome isn't really light on memory usage anymore. 1.5GB of that is split between Google Reader and the Spotify web player though, anything Javascript heavy seems to use an absurd amount of RAM after it's been left open for a few days.
We already have a team working on memory consumption (and we already made some gigantic progress). We also have a team working on improving addons APIs (see the new builin SDK).
What about an easy way to redistribute customized versions of Firefox? What im thinking about is something like a software pack or plugin bundles, similar to ninite. Very much like the spf-13 vim distribution. Or like some of the bundled eclipse distributions, like PyDev or Aptana. It's a batteties included idea.
There are a lot of cool plugins, but new users don't always know what to get. But if if it was easy to fork Firefox and add some plugins and redistribute it in whatever forums or communities they prefer, that would be cool.
So there may end up being a Firefox for reddit, for image processing, for web design, for web dev, etc.
This has been considered and rejected as inferior to the current approach. The Android browser used this and it was rewritten to get rid of it.
Let's focus on doing things that improve the user experience rather than blindly copying features from other browsers without considering if they make sense.
Moreover, it has absolutely nothing to do with the question asked.
Plugins and add-ons AFAIK. There are also memory/performance implications, which apparently were not insignificant on mobile platforms. Given that Chrome uses more memory than Firefox, it's perhaps not insignificant on desktop either.
Meh, if you're referring to making every tab its own process, I personally don't like that. I'm sure there's performance or stability reasons for Chrome's decision, but I also don't really like having 30 different processes open for the one application.
Why not? I'd say 30 processes is a lot better than 1 process. You also need to take into consideration that the number of cores is increasing pretty quickly. Pretty soon consumers will have 32 cores and whatnot. I'm a Firefox user but really it's the largest weakness Firefox is having. If I open two new tabs simultaneously, the whole browser can freeze for a bit.
Why not? I'd say 30 processes is a lot better than 1 process.
1 process is better than 300 processes, and I sometimes have more than 300 tabs open. (I always have more than 100 tabs open.)
If I open two new tabs simultaneously, the whole browser can freeze for a bit.
People who prefer one process per tab tend to mention that tabs freeze up their browser, and I believe them, but this never happens to me. Maybe it's because I use NoScript and have all plugins disabled, except for Flash, which I block with Flashblock.
Honestly (and I'm fairly certain this is a bad reason), it's a cleanliness thing. If I use top or the process manager, it gets annoying to scroll through countless variations of 'chrome'. Maybe if the process manager could group those into one heading where I just see "Chrome (30 processes)", I'd feel better about it?
My favorite thing about Chrome is that I can open up "View background processes", get a list of which tabs are using up how much memory, and then click-and-kill selectively.
But this is an end user experience, rather than a web dev experience.
Your comment is somewhat self-contradicting. Baking plugins into firefox will increase Firefox's memory footprint. Plugins are plugins for a reason -- some people will want the functionality and others will not.
How would you like to see the API for plugins/add-ons improved?
This and tabs on the left. It feels like using Tree Style Tab is not really fully optimized.
I don't think you need to add a more things for developers. The tools already present are good enough, I saw someone suggesting a built-in sFTP. That's just too much.
Yeah it's not a developer feature but I'd love this too. It's just a good way of managing lots of tabs and developers do have lots of tabs open.
Tree tabs are the more conventional way of understanding lots of items (and tabs are hierarchical -- parent tabs are where you opened the current tab from).
How much RAM do you find FF using? And how much do you have in your machine?
Just curious because I rarely find FF creeping up towards a gig and RAM is so ridiculously cheap most of what I have just sits around doing nothing. (Edit: shouldn't have said nothing, but I'm not maxing out even with 4GB unless I am working in Photoshop/Illustrator)
Mozilla is working hard to manage memory/RAM in the most efficient manner. In many 3rd party tests, Chrome uses more RAM than Firefox for the same set of tabs. Mozilla is tracking this here: https://areweslimyet.com/faq.htm
What about memory analysis tools for javascript?
Im my experience ATM most developers don't have a clue about whether they have memory leaks and where the could optimize (javascript) memory usage.
Chrome does have heap snapshots https://developers.google.com/chrome-developer-tools/docs/he.... Also chrome has some tools to analyze thos snapshots they are still far from what the Eclipse Memory Analyzer http://eclipse.org/mat/ can do for Java apps. Fortunately chrome heap snapshot support will be coming soon for the Eclipse Memory Analyzer.
IE BTW now has similiar heap analysis tools.
I have this thread, HN home, espn, and techcrunch open in both FF and Chrome right now
Chrome 454,920K
Firefox 487,156K
a 30 meg difference is really _not_ a big deal.
Anywho I find myself using chrome for dev while FF is for browsing. The AngularJS and Speed Tracer addons are nice for Chrome. I browse with FF because they support tags with bookmarks, which I'm addicted to
If you are just looking at what's best to add to the browser itself, just look at the plugin market to see what's in wide use.