[shove]

The download, parse, and startup time isn’t considered, so while this excellently captures runtime performance metrics, it’s not the whole story. Does your app replace thousands of rows in a table?

And yes, I’m questioning the time I spent getting Preact to work rather than going with the React herd. :)

[krausest]

The startup benchmark tries to capture that, but doesn‘t simulate network performance. Still I think it shows some interesting results.