Hacker News Reader

Well the main difference is, 50ms of processing time and memory usage on the server I'm paying AWS for, vs. 7 seconds of processing time and memory usage on the client machine the user brought to the party. The goals of performance engineering in the two spaces are different.

View on HN

Sure, you are right that the goals are different. But good engineers respect the user. And that includes paying close and careful attention to client side performance. While you may not be directly paying for it, crappy performance there will cause users to bail on you - then you won't need to worry about how much the server is costing you.

Indeed, I think the notion that the user’s resources are “free” is deeply flawed. High load time, CPU usage, RAM usage, etc is merely tolerated, not accepted, by users depending on the value of your product, making whatever gains in productivity or reduction in expenses made by shoveling it all off onto the client more like a high interest loan than a freebie.

Reply on news.ycombinator.com