News for hackers. Stuff that matters.
Re:Learning from Twitter by gfunk911 on Thursday January 20, 2011 @06:18PM (#2125734)
I don't think it's fair yet to say that Twitter won't do these things. The first step was fixing the regression as fast as possible.
Re:Learning from Twitter by jfager on Thursday January 20, 2011 @06:25PM (#2125762)
And the second step was writing a highly-visible blog post slagging the free code they use in their many-million-dollar website before they actually understood the problem?
Re:Learning from Twitter by weixiyen on Thursday January 20, 2011 @06:55PM (#2125843)
Exactly.

It's not hard to pinpoint the exact line of code that is causing the slow down and come to a conclusion. This could have easily been fixed without reverting to jQuery 1.4.2 and risk breaking other things in the process (assuming they upgraded to 1.4.4 for a good reason).

On top of that it was their own code which led to the slowdown. Passing the blame to jQuery seems wrong when it was their own mistake to disregard best practices.

Here is where Twitter failed:

1) Need better testing when updating a version of the library.

2) Need to understand what the library you are using is doing and how to avoid performance bottlenecks.

Basic stuff.

Re:Learning from Twitter by bretthopper on Thursday January 20, 2011 @06:22PM (#2125753)
Well there's not much else to do. It seems other people did most of the work for them.