DEV Community

Discussion on: You learn the most when you fail - Distributed programming project

Collapse
 
miketalbot profile image
Mike Talbot ⭐

Thanks for sharing that! I'm just looking at my "scaling for robustness" grid system and wondering how much of your issues it will deal with. I guess the issue is the size of the problem being distributed and the cost of the serialization and deserialization. In the end my current system is way simpler than yours - not requiring the fail over of leaders etc and using a "blackboard" style approach for nodes to be able to pick up merging jobs etc.

It makes me wonder in these cases - your approach was imperative and mine inverts control. I can see architectural benefits of inversion here (which is why I wrote it that way) - but I wonder what is has cost me now...

Thanks for the food for thought.