Lambda the Ultimate

inactiveTopic C++ Templates as Partial Evaluation
started 11/17/2000; 11:35:44 AM - last post 11/18/2000; 2:12:54 AM
Chris Rathman - C++ Templates as Partial Evaluation  blueArrow
11/17/2000; 11:35:44 AM (reads: 855, responses: 1)
C++ Templates as Partial Evaluation
In a discussion in a parallel universe where I picked up the Perl links, there was a mention of Blitz++ which is a fast numeric processing C++ template library. In looking up this library, I ran across the above link which is an interesting read about how the C++ template mechanism is being used:

Templates were designed to support generic programming but unintentionally provided the ability to write code generators and perform static computations. These features are accidental, and as a result their syntax and semantics are awkward. Despite being unwieldy, these techniques have become somewhat popular because they partially solve an important problem in scientific computing- how to provide libraries of domain-specific abstractions without performance loss.
Posted to "" by Chris Rathman on 11/17/00; 11:38:37 AM

Ehud Lamm - Re: C++ Templates as Partial Evaluation  blueArrow
11/18/2000; 2:12:54 AM (reads: 908, responses: 0)
As a fan of generic programming, it is no surprise that I like this...

Of course, specializing sw components is, in itself, an old idea. A good intuition can come from McIlroy's sw factories paper (which I think is available on the Net somewhere).

I see that C++ Gems appears a few times in the references. I think I looked at it once and decided not to buy it. But I dodn't remeber why, so I am not sure I am not mistaking it with another book.

Anyone read it, and can rate it?