David Ellis

Hi @tajmone, Thank you very much for such a thorough issue and for your considerate response to the name clash. With respect to #257, we have been implementing syntax highlighters...

I think one of the earlier ideas behind the graph and not-quite-Turing-complete nature of the language is the answer here: we should be computing a rough estimate of the time...

@depombo what about a simple heuristic to choose sequential versions when the number of Tokio tasks crosses some multiple of CPU cores? Not saying we do that now, but it...

> > @depombo what about a simple heuristic to choose sequential versions when the number of Tokio tasks crosses some multiple of CPU cores? > > Not saying we do...

For completeness, the current compiler accomplishes this with interface type aliasing: ```ln fn do(t: any): any { ... } ``` Where both `any`s are implicitly referring to the same internal...

@aguillenv yep, that also needs to be tackled, and was going to be the *next* task I'm going after once this one is done. :) The current strategy of halving...

> I personally think that it might be better to do push - I feel like pull not only requires the extra sockets to grab the data but it also...

> That's solved by an announcement message or a bool flag right? Once a node starts up, it either sets a flag in dns stating it needs all data, or...

> yeah, mostly - the big difference being that the new node is just announcing itself, whereas in the pull method it would poll each of the other nodes for...

> Another alternative is maybe using something like what Riak/Cassandara/ES do which is having a concept/abstraction of partitions? A node can have 0 or multiple partitions and each partition can...