scsh-users
[Top] [All Lists]

Re: Now it works (was Re:Modules)

To: pax! <pax@dedalo.ws>
Subject: Re: Now it works (was Re:Modules)
From: sperber@informatik.uni-tuebingen.de (Michael Sperber [Mr. Preprocessor])
Date: Tue, 10 Dec 2002 08:32:49 +0100
Cc: <scsh-news@zurich.ai.mit.edu>
>>>>> "pax!" == pax!  <pax@dedalo.ws> writes:

pax!> Now I'd want to do a suggestion. I can't understand why doing this
pax!> relatively simple thing has been so difficult. Loading files in
pax!> runtime is a relatively common task in interpreted languages, and a
pax!> very useful feature. IMHO this particular problem could have been
pax!> addresed with something like:

pax!> (load file (current-package))

pax!> or something similar. Is there something flawed from a scheme point of
pax!> view in my perspective or in my solution?

Yes there is.  In the presence of modules, incrementally loading code
*into a module* after it's been initially defined is a semantically
complex operation: what's the effect on dependents if you change
bindings, for instance?  Worse, what's the effect of redefining a
macro?

The module system of Scheme 48 is fairly unique in that it allows you
to do this at all.  As I suggested, it would have been easier (and
semantically more sound) to just write your runtime-loaded files in
the module language.

-- 
Cheers =8-} Mike
Friede, Völkerverständigung und überhaupt blabla

<Prev in Thread] Current Thread [Next in Thread>