Impact (Javascript Game Engine) Sales Report

When you're writing a game engine, your main concern is performance. Not using the built-in javascript sqrt() and rolling out your own would be woefully inefficient.

That question would have made more sense for an engine in C, building your own fixed-point sqrt() for example.

sqrt and its approximations seems a bit too low-level, both for what he did and what Facebook is doing. This is the kind of stuff you look up, as opposed to other algorithmic difficulties that arise out of total different situations where without a good general background you wouldn't even know what to research.

Questions about trees, graphs etc. would seem more appropriate for game development (and probably for Facebook in general).

But that's just me. Sqrt might even be a good question if he's a recent graduate and they want to see if he still can think back a couple of years...

There is definitely room for simpler. jqGrid and others tend to favor convention over configuration. This certainly makes example code shorter. But usually I'm adding the grid to a site with pre-existing data sources, UI conventions, and other requirements. I have to override all the defaults, and the result is less maintainable than if they'd forced me to configure everything up front. In particular I always have to override the AJAX stuff, and it's always ugly.

I see here that you are pointing at only one particular problem, which is loading heavy amount of rows in one page.

Can you explain how it can be better than SlickGrid? (other than that)