Help making Firefox better – share your thoughts!
Last weekend, I had the pleasure of being invited by Mozilla to Prague to, amongst other things, discuss the future of Firefox.
Background
As we all know, Mozilla, with Firefox, took back the web from a Microsoft-dominated era with Internet Explorer, made it open again and put emphasis on web standards. The way I see it, we owe gratitude for the openness and much healthier web browser landscape we see today, and it’s a very important step towards an open web.
Following that, Apple started developing Safari, based on WebKit, and last year, Google got into the game with Google Chrome (also based on WebKit). Somewhere at the sidelines, marketshare-wise, Opera has been around for quite some time, and has lately put a lot of effort into having people evangelize about HTML5 and future technologies.
With all this innovation, faster release cycles, new features and better overall support is brought to us, and I believe competition thrives from this. For many years, Firefox was known for setting the bar, both with web standards support and new exciting features, but as of lately, other web browsers have started seriously competing for this position, in terms of feature-support and performance.
Meeting in Prague
In Prague, I got the opportunity to discuss with Firefox Director, Mike Beltzner, and Firefox developer Vladimir Vukicevic about the future of Firefox, what to focus on and what web developers find most important. I also featured in a panel about HTML5 moderated by Vladimir where we discussed the balance of fixing old issues compared to implementing new things.
Before the panel, Vladimir had also expressed what they are looking for:
We are interested to hear what you think we could be doing better at, in terms of support for current or emerging web standards. Are there existing features in other browsers that you want to take advantage of that we don’t support? What about those features is compelling?
What’s missing from the web platform? Where do you want to see us take it? If you could pick one capability to add to the web, what would have the biggest impact on your web app development?
Of the currently supported standards, what’s painful? What would you like to see us focus on improving, whether through enhancement or through change?
What I think
The way I see it, Firefox took, and held, the lead for some time, but it is facing a huge challenge right now. The parts I find most important are:
- Start-up time
- In my daily work, with colleagues as well as people I met in the various companies I visit or do work for, a lot of them have chosen Google Chrome as their number one web browser. Anyone I ask about why, the answer always comes out the same: “Startup time – Firefox is so slow”. I know lots of work is being done by Mozilla in this area, but, if possible, I believe it has to be given even more priority. One suggestion is to postpone checking for add-on updates till after the web browser has actually started. Complement performance focus with looking at perceived performance vs. actual performance.
- Performance, performance, performance
- I think this can not be stressed enough. Without a doubt, the reason behind a fast adoption of other web browsers, especially Google Chrome, is spelled speed. Page rendering and JavaScript performance have to increase even more. In all fairness, it has to be said, Firefox 3.5 without any add-ons is very fast and almost up to par with the others, but at the same time add-ons is the main competitive advantage and in many cases, the sole reason for people choosing Firefox. Therefore, add-ons implementation has to get faster as well, and I believe Jetpack is an attempt to achieve that – problem is, native Jetpack in Firefox and people porting their add-ons to it is far away in the future, and other measures need to be taken before that.
- Separating processes
- This goes a little hand-in-hand with the performance points made above, but with separate processes for each tab, and especially one for the Firefox UI, it gives a great playing field for accomplishing that. I know this is in the long plan for Firefox as well, but I just want to emphasize how crucial it really is.
- Release cycles
- I think it comes down to how often a new version is released, but also, more importantly, how many features that are being tried to be packed into each release. The important choice to make is incremental additions and enhancements, and not believing each version will contain everything that is desired.
- Acid3
- Why I think the Acid3 tests matter is not necessarily what support that comes with a 100/00 score, it’s about the message it sends out to developers. If Mozilla can’t deliver a score of a 100, while WebKit and Opera can, it conveys the feeling that Mozilla have a harder time and is a little bit behind implementing things.
- Implementing new features
- What was good with the Firefox 3.5 release was support for
video
andaudio
elements, Location Aware Browsing, general HTML5-related support in the form of CSS enhancements, query selectors etc. Keep this up, but most of the above are just following WebKit examples – I would like you to be first with the most mind-blowing features! (I think the AwesomeBar is one example of such a great feature)
I want Firefox to be the best web browser out there again, all categories, and I believe these are the vital steps to achieve that. Make me proud! π
What do you think?
Those are my thoughts. Is Firefox your main web browser, or do you just “use it for developing”, as I mostly hear? What would it take to make it your number one web browser again?
Help Mozilla out by contributing your thoughts!
I would like to get rid of the search window, and have only one input window for both URL’s and search strings ( like Chrome )
Is there need for two input fields ?
As a developer I would like to see more HTML5 features implemented.
I would like for Firefox to explore more ways of lowering the bar of participation and further emphasize the open source dimension of the browser.
Test Pilot is a very good initiative and it would be great if it would become funnier, easier and more rewarding to contribute to the future of the browser.
Focus more on development of things like Bugzilla and Test Pilot and be inspired by GitHub, Uservoice, Twitter etc.
Great write up! I think it put all annoyances I’ve had with firefox on “paper”.
It would surprisse me a lot if mozilla is not aware of these issues already. On the other hand I can really understand why firefox has this performence issues since chrome is pretty new and built from the ground up with performence in mind, while firefox has been around for a while and theres probably code in there from the mozilla browser days aswell.
The comparatively poor performance of Firefox versus Chrome is underplayed here. Chrome is faster by many orders of magnitude for every single operation. Even switching to a Firefox window is slower than switching to a Chrome window.
The chrome of Firefox has to be reduced and the handling of tabs needs to be a little slicker for me to even consider using it as a browsing tool instead of a development tool.
As a developer, what I want most is for Firebug to be an integrated tool and not an add-on. That might improve perf a little bit too.
Acid3 is definately important.
Allowing for several usage scenarios would also be helpful, mostly for developers. Say for instance that I load up Firefox with all sorts of wonderful developer extensions (Firebug, Firefinder, et al), then I have to pretty much gimp my entire setup to get the same experience as the average user sees. Allowing developers to use several “scenarios” where you can cherry pick which addons to load, would go a long way.
I suppose that this might also win back some users for using Firefox as the default browser, as it will not be slow for everyday use, when you don’t load all your developer extensions just for accessing YouTube π
The UI could certainly be more polished… taking a good hard look at providing a clear progress indicator would go a long way. Maybe the little progress bar at the bottom of the window just needs to find a new place to live? I for one spend a lot of time in the address bar, and expect the progress indicator to be very close to this and obvious, but it isn’t.
XUL needs to be faster, or just compiled to native UI … when switching focus to and from Firefox on OSX you can sometimes see the background of the topbar being redrawn (Camino doesn’t have this problem).
Smaller and faster release cycles, will also help new platforms like Maemo to keep their browser fresh, instead of leaving users with a very old browser, and only updating once a year.
Robert
A short while back (prior to FF2.0) I entered into discussion with the development lead that there was a feature they could add which would change numerous peoples lives. I mentioned a potential text resizing feature and was assured that they were soon coming up with something better. It was zooming. The guy missed the point totally. I was actually asking for two new buttons to be placed by default in the navbar. Next to the Refresh / Stop / Home. My suggestion was + and – buttons to increase and decrease size of text. Whether they represented text resizing or zooming was immaterial.
It is fine to say View -> Zoom or CTRL + are there, however the regular web user doesn’t do browser menus or CTRL keys. I’m thinking of the types who think Google is the Internet. I Introduced the concept of text resizing to my short sighted father In-Law after he had been surfing the web for several years. He was amazed and now doesn’t suffer when going to websites where the default text size is too small. I’ve also formally taught many net newbies / older people and text resize/zoom is a great feature for lots of people. OK it’s there right now – how helping about making it easier to discover and subsequently use?
It’s a difficult question, since most things I can think of can just as well be implemented as extensions. But anyway, here are a couple of ideas:
Automatic login: I’d like a way for a web site to request credentials and log me in automatically (if I have authorized the site). This would be safer and easier to use than the (cookie based) “remember me” and form autocompletion. It would need a protocol that web sites should to implement, perhaps something based on OpenID.
Edit image before upload: When uploading a photo, the possibility to scale, crop and recompress the image before it is sent to the server. Useful for uploading avatar images in the correct size, and for photos to be published on a blog.
@Morgan Roderick: Firefox supports profiles which should do what you want.
@Martin Vilcans:
Well, profiles doesn’t exactly do what I want, as it’s mostly about having several USER profiles, and not several CONFIGURATION profiles… or at least that is my understanding of them.
Maybe I should look at them again, to see if I can make them do what I want.
I agree with many of your points Robert but one thing I really have waited for a long time is that Firefox should save background images used in CSS when saving a page. It feels real bad when a web browser fail in that area anno 2009. I really don’t know how good the other browsers are in this regard (I believe that Opera can do it) but well, it feels bad when Firefox fail that. People often save web pages (or I so assume at least XD) and I for one expect the saved copy to look like the page on the web. And with the increased adoption of CSS for design, it just looks real bad. Heh, especially since this happens with imported CSS too, so some sites just get saved naked. So… well… that’s what I have to say right now.
(Not the best written entry I’ve ever made, blame tiredness XD)
In a prioritized list:
1. Mobile. They must get Fennec out of the door and on more platforms, not only Maemo and Windows Mobile. They need deals with carriers to set Fennec as the default browser. Weave on Fennec would be a killer feature.
2. Acid3 – for pure marketing reasons. The missing pieces of the puzzle is SVG fonts and SMIL. I know no dev who would like to use that today, but that 100% score is seen by the masses as important, even though they have no idea what the test actually does.
3. FFox 3.7 must have a complete ECMAScript 5th edition implementation. I have high hopes it will.
4. TraceMonkey must start to trace recursion as well as iteration. DOM manipulation must be faster as well. Dromaeo must be marketed as a more relevant speed test than Sunspider (it is!)
5. Ubiquity is a power user killer feature, just like the awesomebar was.Integration into the toolbar and less resource using would be great. (Ubiquity occasionally sucks up all my CPU for 5-10 seconds.)
A few additional thoughts: Mozilla often comes 2nd in implementing some stuff, but when they do, they seem to do a much more thorough job. Case in point is quite a few of the bugs in Acid3, that Opera and Safari quickly “passed”, but Mozilla people found spec holes, edge cases, etc Another example is when Webkit implemented Canvas as background images, but Mozilla implements the use of any element as a paint server! Even Dave Hyatt thought that was neat.
Of course Opera and the Webkit teams do tons of cool and good stuff as well. I just wanted to balance the equation somewhat on a few issues.
And finally:
Personally I like having a status bar, dislike tabs on top, but think tabs on the side is a neat idea (Opera 10 option), and dislike unified location and search. A neat Ubiquity integration into the location bar might change that, though.
– PERFORMANCE!
– Better Memory Usage!
My 2 cents π
Printing! I hate paper, but some people don’t, and Firefox still sucks at printing! DIVs and images disappearing when split between pages, only prints one page when there should be at least two… bugs that have been in Mozilla since 2004.
I wish they were a little faster with some of the new specs. Right now WebKit seems to be implementing a lot of the interesting ideas first.
My pet peeve is multiple background images. I spent hours this weeks in meetings where our team standardised our HTML practises. Almost all the points of contention were workarounds to the fact we still can’t have multiple backgrounds on a single element.
It’s weird how Firefox started out as a lean and fast alternative to Mozilla/Netscape and now starts to feel bloated and sluggish. I’ve been a fulltime Firefox user since.. well, since it was called Firebird β but the only reason I stick with it nowadays is Firebug. I just can’t live without it.
I don’t care about acid tests, I just want a quick and responsive browser able to start up as fast as possible and not feel “in the way”, kinda like Firefox used to compared to its old competitors.
Thank you very much for great comments and input! I have contacted Mike Beltzner and Vladimir Vukicevic with this, and I hope it will help them in figuring out what to focus on.
Please keep more feedback coming!
Firefox is not always working well under Ubuntu. Sometimes it's kind of frozen.
[…] 08: Help making Firefox better – share your thoughts! […]