You are viewing a single comment's thread from:

RE: A New API for Hive

in HiveDevslast year (edited)

This is by far the best explanation I've read about the basics of the blockchain and I'm quite sure even someone completely blind about code, making a sketch (I use to sketch when I don't understand completely or it is to much information) of what you've wrote can have a quiet good understand of what is in the base of the blockchain and how it works. Congrats! Impressive document here!
Question: Since hivemind will be in the middle and there is another modular layer of code (HAF), couldn't hivemind rewriten in C++, isn't it C++ must powerful than Python?

Sort:  

making a sketch

I actually thought about adding a drawing or two, but I'm not fast with drawing tools and I was too lazy :-(

couldn't hivemind rewriten in C++, isn't it C++ must powerful than Python?

Hivemind consists of two parts: some python wrapper code and a lot of SQL code. The SQL code is the bulk of hivemind. We could eliminate the python wrapper code, but it just doesn't do a lot, so replacing it wouldn't speed things up too much. Probably the most useful thing to do would be to replace the existing python-based web server with a postgresT-based one: we've already found that this speeded up performance in some of the other HAF apps we've created.

 last year (edited) 

Sketch

The way you've explained it's not necessary, I think.
For someone quiet lost in explanation own sketch might help.
Don't know if you've ever read 100 Years of Solitude from García Márquez, 7 generations of same family, a "ton" of people with the same name and surname, I have to read twice and make a sketch to not get lost lol.


Ok understood, it's mainly SQL.

I haven't read it, but I can imagine the problems that result :-)

My biggest read challenge so far!

Yes, it's mainly SQL, and we did spend a lot of effort speeding up that SQL code (as well as refactoring it for easier future development).

Great!