
17 Aug
2009
17 Aug
'09
1:29 p.m.
on 17.08.2009 at 17:17 Rutger ter Borg wrote :
Indeed, it should be easy to write. Just as an example, a function "solve" would already cover a large portion of the whole of LAPACK. A DSL for linear algebra does not need to be complex. The more expressiveness, the better, I would say.
just to summarize all the discussion above i say that a clear design goes first e.g. "x = solve(A, b);" can be a design statement telling that "the user want to do it that way!" only THEN comes an implementation wich can be whatever you come up with -- Pavel