RE: [Axiom-developer] RE: Boot vs. Lisp

[C Y]

--- Bill Page <address@hidden> wrote:

> Concerning lisp style, I think the following article is
> particularly good:
> 
> Tutorial on Good Lisp Programming Style
> 
> http://www.cs.umd.edu/~nau/cmsc421/norvig-lisp-style.pdf

Looks interesting.
 
> Being dated August 10, 1993, perhaps it is not entirely up
> to date with the available tools. And what it has to say about
> documentation seems very much oriented to the typical lisp
> programmer's view rather than Knuth's literate programming
> view, i.e. writing *code* that is intended to be read by a
> programmer as opposed to writing a *document* that contains
> code intended to be processed by the computer.

I'm not entirely sure that's incompatible - I see no reason that human
readibility of code shouldn't be a factor in the code in literate
documents.  After all, it's part of the document!  Obtuse code
associated with a documentation of the theory and reason for it might
technically be literate programming, but I'll be more impressed if the
code itself is also well layed out and organized (in the cases where
there is an irreducibly complex bit of code that it doesn't make sense
to break up and document).

> With regard to BOOT, you should note what Peter Norvig says
> about "Syntactic abstractions". I wish he had written more
> about that.

Oh, I agree that a language suited to its task is a good thing, but
this is true if and only if it makes that job so much easier it is
worth learning the additional syntax rules required to work with it. 
Any new syntax has to demonstrate it does this, or it is detrimental to
the program's maintainability - e.g. it would be simpler to figure out
a slightly more complex program in one language than to learn a new
language to figure out a simpler routine.  Maybe it would help to lay
out:

a)  What problems in Axiom BOOT is used to address.  My understanding
was that its primary task was to enable writing the SPAD/Aldor
languages - is there going to be enough work on these languages that
figuring out lisp code vs. BOOT code will impose a significant drag on
features/maintainability/what have you?

b)  A design document and spec detailing the purposes of BOOT, the
problems it solves, and formalizing the language itself so we can do
things like write advanced Emacs modes and try to adapt the SLIME
development tools to working with it.  

Tim, being as you've been through this once before, can you give us
your thoughts on this?   I'm trying to figure out a way to frame this
question so it can be decided on the merits, but it's a bit diffcult
and I'm probably not really the one who should be trying.

Cheers,
CY


        
                
__________________________________ 
Yahoo! Mail - PC Magazine Editors' Choice 2005 
http://mail.yahoo.com