[erlang-questions] Rhetorical structure of code: Anyone interested in collaborating?

Richard A. O'Keefe <>
Tue May 3 02:36:52 CEST 2016



On 2/05/16 11:56 PM, Joe Armstrong wrote:
> I was thinking. There is another problem, which is not addressed.
>
> We just publish final code - but not the process involved to get to the code.

How that takes me back.  There was a PhD student at Edinburgh when I arrived
there working on transformation-based programming.  I've always remembered
his name as Martin Feather, but that may be wrong.  The idea was that you
started with a high level executable specification and then provided a 
script to
transform it to efficient low-level code.  The idea of working via 
transformations
from short clear code to longer code was not new even in 1979, and 
people have
continued to work on it.  What Martin Feather (if that was his name) was 
one of
the first to do was to make the transformation process semi-automatic, 
with a
script saying things like "unfold that function" or "use this law". (It 
tied in with
the Burstall/Darlington work on automatic transformation.)  The idea was 
that
if you changed the high level specification a little, you might be able 
to reuse
most of your transformation script

In principle, a script could even have alternatives, and I *think* his 
system did this.
(Again, this ties in with the work on Edinburgh LCF and the idea of 
controlling a
theorem prover by writing tactics (do a specific thing) and tacticals 
(combinators
for tactics.)

I am going to forward your message to a class because I think you're 
completely right.
> Unfortunately I only ever have a clear description of my problem at the
> very end when the problem is solved.
>
> Chicken or egg first?

The amniote egg is many many millions of years older than birds.



More information about the erlang-questions mailing list