> Right - the latter was my idea.  Sorry Tim - didn't mean to run away
> with the thread there.
> 
> Cheers,
> CY
> 
> --- root <[EMAIL PROTECTED]> wrote:
> 
> > My suggestion was that each implementation include the spec as
> > documentation
> > not that the spec contains a reference implementation.
> > 
> > Tim
> > _______________________________________________
> > Gardeners mailing list
> > [email protected]
> > http://www.lispniks.com/mailman/listinfo/gardeners

The idea of documenting the spec with code is fine and would be a useful
guide to those who want to understand possible implementations. But that
isn't my goal in trying to convince people to adopt literate programming.

The real benefit of literate programming is very hard to see giving the
up-front cost of writing a lot of text "for nothing". In the long run
when the original developers leave, which hasn't happened to most open 
source projects yet, the project will die if the "next generation" is
unable to maintain, modify, and extend the existing code. 

A lisp system that has the program text, the ANSI standard text, some
implementation information, and test cases all in one place would make
it possible to sit and read how "setf (foo bar)" or "(gc)" or even
"(ffi-call baz)" works. Imagine being able to sit and read the
constraints on a hash-table (from the spec), the reason why these
affect the implementation, the impact on time/space by the choices,
the test cases that check for boundaries and an explanation as to why
these are "interesting boundaries", etc. And even check the index for
places where the hash table is used so if you change it you know where
it might break the system.

When that occurs I believe we can begin to consider programming more
of a way to learn and programs will live past their creator/company.
We will have started to document "how and why" things are built rather
than just building them. History will have begun in computer science.

(throw 'please-stop nil)

_______________________________________________
Gardeners mailing list
[email protected]
http://www.lispniks.com/mailman/listinfo/gardeners

Reply via email to