Hacker News new | past | comments | ask | show | jobs | submit login
Show HN: The Flux Challenge in 139 Lines of JS (github.com/kennytilton)
3 points by kennytilton on May 13, 2018 | hide | past | favorite | 2 comments



I was previous linked to your Cells library because I'm trying to do something very similar [1]. Matrix seem to be in a similar vein (at least from the sample usage here).

Could you tell me if any of your libraries allow live editing of formulas? (Adding new ones, editing existing ones and deleting some.) If yes, do the new input (cI) need to be declare manually?

I've also seen these slides [2] in your repo. Can you say a bit about the solution to "The Glitch"? As you've noted there, topological sort isn't enough if there are conditionals.

I also couldn't tell what "Glitches and RoboCup Sim" was about from just the PDF.

Finally, when using a cells-like library, do you get "discontinuous jumps" when cells are no longer suitable (or the network becomes too complicated) and fallback to imperative callbacks? Or do you manage to used cells for everything?

[1] https://github.com/asrp/persistent_doc https://github.com/asrp/guitktk

[2] https://github.com/kennytilton/cells/blob/master/Lisp-NYC-20...


The Flux Challenge is "tailored to touch Flux's weakest spots", especially "coordinating multiple async data sources", so Flux solutions are sought, but other approaches to state management are welcome.

The Challenge author used his own reactive framework and RxJS: https://cycle.js.org/ https://github.com/Reactive-Extensions/RxJS

My solution leverages my Matrix data flow library (versions in Common Lisp, CLJ/S, and JS). Matrix is much like the JS MobX, if you know that.

https://github.com/kennytilton/matrix https://mobx.js.org/index.html

Moral: The data flow paradigm is well-suited to smoothly absorbing external input, so the async feeds of the Flux Challenge are right in the data flow wheelhouse.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: