Re: Weirdness in Dylan spec

[moon (David A. Moon)]

> Date: Tue, 07 Jul 92 16:22:24 -0700
> From: nathan@akbar.teleos.com
> ....
> I agree with what you said about this example, but what about my
> original issue.  Namely, is it broken for a 'dynamic'/interactive
> language to *force* the user to forever (where forever is the duration
> of the current process) commit a symbol to a particular use?
> ....
> define-generic-function appears to be intolerant of such changes.  Is
> this the case?  If so why?
> ....
> My hope is that this is all pretty obvious, and this is just an error
> in the spec.

Well, the Dylan book is about the language, not about any particular 
implementation or any particular development environment.  I'm afraid in many 
places the book is not written in a way that makes this clear.  The problem is 
worse because people used to other Lisp-family languages are usually used to 
implementations that don't make that distinction clear, so our minds are 
pre-set to be confused by the book's unclarity.

You assumed when you read something about the language that the development 
environment would not let you change the arguments of a generic function.  
That's jumping to a conclusion and I personally would consider any development 
environment that didn't let you do that to be pretty deficient.