[Axiom-developer] About pamphlets and literate programming: how to identify identification levels?

[David MENTRE]

Hello Tim,

I know that you are offline this weekend, but there is no hurry to reply
to this email, we have a 30 years time frame. :)

I reconsidered what you said at the end of August about pamphlets:
http://mail.nongnu.org/archive/html/axiom-developer/2003-08/msg00071.html

When you describe View 1, top-to-bottom path, and View 2, side-to-side
path, you say you want to respectively descend abstraction levels and
stay at the same abstraction levels.

It raises for me the following question: how would you define a given
"abstraction level"?

For some levels, this is easy and can be deduced from the kind of source
code: Spad, Boot or Common Lisp (and even C if you want to dig into GCL
internals and GMP). However, within for example Spad code, I'm sure you
would like to define some intermediate levels. How would you define
them? From Spad language syntax (domains, categories, operations and
their implementations, ...)? From properties of domains (all domains
having the '*' operator)? From a combination of both (all domains
inheriting from Ring and having the '**' operator)? Would you need a
kind of query language on Axiom's algebra to write properties that would
define a given abstraction level? Something else?

If you have some example or further explanation, it would propably light
up some candles in my brain. For example, could you give a very short
overview of a top-to-bottom example?

Yours,
d.
-- 
David MENTRE <address@hidden> -- http://www.nongnu.org/axiom/