August, 2009

Aug 09

Eradicating Start Up Dialogs

I’m starting up a new project, which from the title is rather self describing. Firefox’s purpose is to let you access the Web, but it unfortunately often blocks users from accessing the Web until they have responded to some form of question. This is particularly ironic given the massive amounts of time and developer resources that we’ve been pouring into improving our start up time. What’s the point of loading 300ms faster if the user is about to spend 7000ms on reading, 2000ms on thought, and 4000ms on interaction?

In many cases start up dialogs represent some form of failing, ranging from a failing of engineering (unable to find or implement a cleaner solution), a failing of design (unable to determine the best approach) or a failing of security (delegating a decision so that we can blame the victim). In all of these cases the dialog itself is simply the manifestation of a larger problem that we need to attack. So perhaps “eradicating start up dialogs” is an overly negative title, this is really more about “solving a range of complex problems, but entirely behind the scenes.”

To help up us prioritize and also make sure we have everything covered, I have two questions:

1) What do you think is the most annoying start up dialog
2) What do you think is the most obscure start up dialog that hasn’t been mentioned yet in the comments above the comment you are about to write :)

Aug 09

Feedback on Scrolling Changes for Windows

The acceleration based scrolling model on Windows has been landed in the nightly builds for a little over a week now, so I wanted to get a post up for people to provide feedback on how well they think the current implementation is working, now that everyone has had some time to acclimate to the change.

Caveat: One really significant issue we are working on is correctly handling the case where the user already has an acceleration model being provided by their mouse or mouse driver, in which case adding another one is obviously really problematic. If you are finding scrolling to be a bit out of control at the moment, you may be encountering double acceleration.

Aug 09

What’s Next for the Perception of Performance?

The title isn’t a rhetorical question. Now that Margaret’s acceleration based scrolling model has landed for feedback, we are trying to figure out what perception of performance improvement we should target next.

You can check out the full list of ideas on the mozilla wiki. If you have ideas on other things we should be considering please add them to the wiki or comment below.

But more importantly: what do you think is the single perceptual trick you believe would have the biggest impact on how fast Firefox feels? (simply saying “actually make ___ faster” doesn’t really count, and yes we are working on that as well).

I’m thinking we might next work on changing the rate of progress for all of our progress bars. Jeff Atwood over at Coding Horror has a fantastic blog post on the topic that really gets to the root of perceived performance vs. actual performance:

The idea that performance is determined largely by the user’s perception rather than actual wall-clock time can be liberating. Like a magician using skillful sleight of hand to perform magic tricks, you can seemingly alter reality.

Aug 09

Sure it’s Fast, But How Does it Handle?

From the Firefox 3.5 video by community member Vineel Reddy

Currently all the browser vendors are engaged in a pretty exciting race to see who can engineer the fastest Web browser. A lot of different aspects come into play when you are focusing on raw performance, including top speed (captured by a range of different benchmarks), zero to 60 (cold start up time, page load), and the slightly more amorphous topic of how well the thing handles (to quote Jobs, does it “scroll like butter”). We have loads of engineering resources focused on areas like top speed and start up, but we are also taking handling seriously.

Margaret Leibovic, an intern on the Firefox team has been working on what I believe will be one of the most important handling (or “perceptual performance”) improvements: adding an acceleration based scrolling model to Firefox on Windows and Linux.


Check out her post for details, and please comment there with feedback so we can start to refine the scrolling model and work out any bugs.

If you have other ideas of how we can improve handling, please add them to this wiki page about perceived performance.