speed

I’ve been running the Firefox 4 betas, and I’ve been trying to keep an eye on speed, as that seems to be one of the main reasons cited for switching to Google Chrome by Firefox users I talked to.

Firefox

Now, don’t get me wrong, Chrome is fast, that much is true, but I think what a lot of people forget about is the cruft. For many users, it could be a few years since they’ve done a fresh install of Firefox. I mean fresh as in “starting with a new profile” so that all your old preferences, add-ons, plug-ins, bookmarks, and other bits weren’t there.

Right now if you download Google Chrome for the first time, and launch it, there is no cruft. It launches fast. If you’ve been upgrading Firefox over the years and not starting fresh with a new profile… well, there’s cruft.

Here’s what I did, and you can try it at home. Create a new user account (if you’re running Mac OS X, just log into the “Guest Account” that wipes itself after every logout) and launch the Firefox 4 beta. For fun launch Google Chome as well. Both will be cruft-free as they won’t have any previous preferences/profile to worry about, and will be starting fresh.

In my tests, Firefox 4 launched pretty damn fast. If Chrome launches faster, it’s probably by such a small amount that most people could never tell.

Now jump back and forth between Firefox and Chrome and load various pages. Again, don’t use any magical timers, just experience it, and see if you think they are pretty close. They look pretty close to me.

As I’ve mentioned, speed isn’t everything, but it’s nice to see the Mozilla folks stepping up their game and realizing that speed is important to a lot of people. I’ve also read some comments about finding ways to “cleanse” old profile data, to help the folks who have been using the same Firefox profile for years and years, hoping for a bit better performance.

Disclaimer: This is not scientific! I did not run any benchmarks, I didn’t test Javascript engines, or anything too crazy or stressful, I just compared the experience of launching/browsing using completely fresh installs of (the latest beta of) Mozilla Firefox and Google Chrome. I’d love to hear some comments from a Chrome convert willing to do the same tests.

I’m told I need to make some arguments to convince people that speed is not everything. I’ll attempt to do that here.

Should I even mention a time when almost all browsers were named after cars? No? OK, good… Let’s talk about cars then. When choosing a car do you typically go to the car dealer, and ask for the fastest car they have? That sounds like a great idea, I mean, you’re in a hurry, you have a need for speed, and you can’t drive 55, so you want the fastest engine they’ve got. It makes sense.

Now that you’ve got your really fast car, you might also wonder how safe it is. Will it protect you from harm? What if the locks don’t work, and it’s not very secure, and people can just open the doors, and rummage through all the stuff you leave in your car… your shopping receipts, the playlist on your iPod, you know, stuff you might think is sort of personal, and you’re not OK with sharing.

Apache / IIS

Many years ago (I know, I start a lot of stories like that) there was a sort of competition between web server software, and this competition often looked at the speed of which a server could serve pages. That made sense to a lot of people, because, you know, you want your pages served fast. The competitors were Microsoft IIS and Apache’s HTTP Server. In many speed tests, IIS was the winner. If you based everything on speed, IIS would have been the clear choice. Now, that would have meant you ran your web serving platform on Windows, because that’s the only place IIS ran. You also would have most likely restricted your code to ASP or maybe ColdFusion, but you probably would not have even considered PHP or Perl, or any of those “weird open source” languages. It sort of made sense that IIS would be really fast on Windows, as both products were developed by Microsoft, and if anyone could make then work together, and run fast, it would be Microsoft. I mean, I’m not suggesting that Microsoft would use undocumented API calls and what not in the development of IIS, right? Right. I’m sure the Apache Software Foundation wanted to put out a fast web server, but they were also very concerned with conforming to the standards and specifications published by the W3C and the IETF and other organizations that were working towards building the web, and they were also putting out web server software that ran on many different platforms, some of which I’d bet you may not have even heard of. There were also a number of great modules you could add to Apache to make it do great things. I worked with Apache and IIS, and I definitely preferred Apache.

Oh, did I mention Apache’s HTTP Server was also open source? While Microsoft could do whatever they wanted to with IIS, including kill it off at any point, or completely re-write it, or force you to pay exorbitant fees for it, Apache was open source. If you didn’t like where it was headed, you were free to take the code and do your own thing. Or hire someone to do it. Open source is like that… it sort of serves as an insurance policy for the future.

The Apache Software Foundation has been a non-profit organization since 1999, and I believe the web has flourished and grown in no small part because of that. I’m not saying that Microsoft IIS or lighttpd don’t have a place in the world, as they surely do, as do other choices, but I’m grateful for the work the Apache folks have done, as they’ve made it possible for myself and others to do so much over the years to help move the web to where it is today.

But I know.. you all want the fastest damn browser your money can buy… regardless of the other features which may matter today, or in the future.