Monday, March 16, 2009

Why is .NET built as GPLs?

The CLI (basically .NET) has shown that Microsoft has what it takes to integrate several different language so tightly that it just works. You could almost roll dice to pick your language for each function and never notice problems. Given that they can make inter opt that easy, why is Microsoft primarily building GPL's * on top of it? Why don't they leverage that interoperability and start building a pile of very small, very special purpose languages? These language could be tuned for very specific cases and, by avoiding corner-cases and compromises, could be made very clean. The theory would be that points in code that transition from one domain to another are good candidates for functions anyway.

I'm thinking of languages like:

A control structures language
This would provide all the ifs, foreachs, switchs and whatnot.
A math processing language
this would define precise math semantics and advanced math expressions
An systems-of-equations language
Think Maple or EES. This would provide no control structures, or define order of evaluation (not to be confused with order of operations) but would rather compute solutions for a system of equations (math for sure, maybe logical and bit-wise) given a number of inputs
An SQL like language
aside from small cases, why should LINQ even be part of C#?

* Sorry there dosn't seem to be an right now, it redirect to the DSL page :b

No comments:

Post a Comment