holistic code-generation model

[Matt Miller]

On Tue, 13 Oct 1998, RE01 Rice Brian T. EM2 wrote:

> as for the example itself, we should first try to pick one which could
> be implemented with a relatively few number of Tunes semantics-modules
> to factor out complexities that would only result from our having to
> forecast too far into the future development beyond our current
> development thread, since the code-generator definitely affects future
> development.
> 

Wow, that's all one sentence.

Matt