Forum Navigation
Tango future
Posted: 09/30/08 21:02:35I came back from the Tango Conference. It was nice, I could put many faces the names, heard interesting talks, and had some fun :)
At the end of it there was a discussion on the future of tango, but I had not thought much about it, and nobody (larsivi excluded) really knew how the situation with druntime was.
I gave some thought to the situation and I think it might be worthwhile to reopen the discussion.
A common runtime would change the things: at the moment tango has to provide everything, but if std can coexist by side this might change. For example it doesn't for sure make sense to have two stdc wrappers. Tangos ones are probably better, and more portable, but anyway it would make sense to merge them with phobos.
Some things are done differently form phobos and it makes sense to keep both approaches, but for other things this is less clear.
So then the question comes up what tango should be.
It could become the symbol of a part of the standard library that works well and is portable.
could it be more modular? I can imagine these separations
runtime (druntime)
tangoMain: threads messaging/sync (maybe even between computers) collections basic io (sockets)/serialization stdc math (basic) basic text op logging basic formats for config/output
xml json (yaml?)
testing
tango Net: networking http (client/server) mime parsing ... mango?
tango Cluster tina tango distribued objects?
tango Sci: solvers blas/lapack
tango games?
=====
from an organization point of view, one can have persons that are responsible for different segments (but that if allowed can change also other parts?), and a committee (with everybody) in case of problems. I think that most problem should be handled with an ad-hoc case by a committee, it does not make sense to try to fix everything in an automatic and fixed structure...
Fawzi