3 ms·

I have been working on a JS runtime that takes away the necessity for databases, by automatically persisting everything that can't be re-computed on the fly. I

I have been working on a JS runtime that takes away the necessity for databases, by automatically persisting everything that can't be re-computed on the fly. If all you need is to store some state somewhere, it works beautifully, but it can't outperform any kind of index-based operation, or anything that needs to operate through all the records.

stephenr 3y ago
... persisting how? Where? Why would this be better than storing what has to be stored, and using some combination of views and generated columns in a regular RDBMS, which has neither of the shortcomings you mentioned, and is language/framework/runtime agnostic. Heck it's even database agnostic to some extent.
- thatha7777 3y ago
  What I've implemented is just (trimmed) serializable continuations for JS. Because ... what you're describing would make supporting migrations (while still abstracting away the RDBMS) very difficult. And because . . . that already exists -- you still have to map your data to views/columnar storage. What I'm describing is ... you _don't_ have to store most data -- your code can just wake up from where it left off