In the documentation, would it be a good idea to discourage the use of C extensions and embedding of Racket (in the old Inside MzScheme senses of "extension" and "embedding")?

I have been discouraging their use for a few years now. I think they made more sense when mzscheme was smaller, was slower, used the CGC, and did not have an FFI. Today, extensions and embedding seem like a last resort, when one has ruled out all other approaches.

This seems like one of those things that is likely to derail potential new Racket adopters: they pick a poor approach that initially (according to documentation) seems viable, they run into difficulty, and they might give up on the entire platform.

Or, if extensions and/or embedding work better today than I thought they do, then my mistake, and I withdraw the suggestion.

Neil V.

____________________
 Racket Users list:
 http://lists.racket-lang.org/users

Reply via email to