18 Jan
2011
18 Jan
'11
12:06 a.m.
On Jan 16, 2011, at 1:15 AM, Hal Finkel wrote:
You might be interested in looking at Martin Erwig's Functional Graph Library (available for Haskell or ML). I don't know Haskell or ML, but I've found the algorithm implementations reasonably easy to understand (there is also a paper), and I've found that using functional programs as a "template" for template metaprograms is often easier than trying to port imperative versions.
Thanks Hal, this is very interesting - I'm not sure if I grasp it yet but I'll be studying it thoroughly in months to come. We have O(1) maps in MPL so life is perhaps a little easier for us than for pure Haskell authors (at least as far as time complexity goes).