Square Galaxy

In March I ran the Acid 3 web browser stress test against all the most modern browsers at the time to see how they performed. Since then, there are new versions of browsers which justify a re-run of the tests. Here are the results:

• Mozilla Firefox 3.0.1 - 71/100
• Google Chrome 0.2.149.29 - 79.100
• Opera 9.52 - 83/100
• Windows Internet Explorer 7.0.5730.13 - Too bad to tell. Perhaps 5/100?
• Apple’s Safari 3.1.2 (525.21) - 75/100
• Flock 1.2.5 - 53/100

All the web browsers were run on the same Windows XP system.

Then, out of curiosity, I decided to check the memory usage of each browser and found the following:

• Mozilla Firefox 3.0.1 - 55M (After all add-ons were turned off)
• Google Chrome 0.2.149.29 - 41M
• Opera 9.52 - 27M (But then crashed on subsequent loads)
• Windows Internet Explorer 7.0.5730.13 - 41M (But still couldn’t render the test at all)
• Apple’s Safari 3.1.2 (525.21) - 41M
• Flock 1.2.5 - 54M

The Acid tests are various tests for web browsers to see how well they can adhere to standards. If a web browser was a computer science project, these would the tests the teaching assistant would run on your project to see how well you did and to assign you a score.

The Acid3 test was recently released, and I ran it on a few web browsers I had installed here. Here is how they performed:

• Firefox 2.0.0.12 on Windows: 49/100
• Safari 3.0.4 on Windows: 38/100
• Opera 9.26 on Windows: 46/100
• Internet Explorer 7.0.5730.13: 12? (It was so screwed up I couldn’t hardly see the score)
• Flock 1.1 (based on Firefox) on Windows: 52/100
• Firefox 1.5.0.12 on Linux: 50/100

I also ran it against Konquerer on Linux, but it kept crashing. Other people are reporting other various scores with various versions.

Update March 7. I’m a little confused about how the tests work. For example, I’ve run it multiple times on the Flock browser, but I’ve seen three different scores come out. I’m confused how the same test can yield different results at different browsers on the same browser. I want things to be more deterministic.

Update March 25. I ran the tests against the new Safari 3.1 on windows, and it scored an impressive 75/100.

Update March 26. Firefox 2.0.0.13 on Windows scored for me today a 53/100.

I was programming for Internet Explorer today and found a bug. When trying to run code like this:

td = document.createElement("td");
td.setAttribute("colspan", 2);

IE doesn’t respect the column spanning. Instead, it should be

td = document.createElement("td");
td.setAttribute("colSpan”, 2);

Yes, even though attributes are suppose to be case insensitve, they are, in this case