What makes Kefir faster than Bacon?

[mhelvens]

I've seen the tree demo. Quite impressive. I'd be curious to know what makes Kefir.js fast / Bacon.js slow.

What do Kefir users sacrifice for this increased speed, if anything?

[rpominov]

It's a little hard to answer as I haven't closely compared Bacon's implementations to Kefir's. For what I know:

• Kefir uses Classes a lot, and stores methods in prototypes, while Bacon often creates methods for each instance. But this probably more affects memory footprint than performance.
• Bacon often implements transformation methods based on other methods. For example this is from the implementation of .debounce: @flatMapFirst (value) -> Bacon.once(value).concat(Bacon.later(delay).filter(false))). While in Kefir most methods implemented without intermediate observables between source and result.
• Also there is a bunch of micro optimizations in Kefir, such as using own concat() function instead of native Array::concat and so on.

About what we sacrifice, I think there is only one feature that aren't implemented in Kefir in particular because of performance penalty — atomic updates. But it's not only about performance, it also pretty tricky to do right. Maybe Kefir will get this one day. (RxJS also doesn't have this, btw :)

Closing this, but feel free to continue discussion.

[asaaki]

This would be something for a FAQ section or comparison overview. Pretty good insight. Thank you!