Re: (problem? 'guile)

[Shriram Krishnamurthi <sk@cs.brown.edu>]

Manuel Giraud wrote:

> ... I think I *do* need some pointers on this (variable aliasing,
> implications on compiler's tools, etc.) to fully understand your
> argument.

You could take, as a starting point, Lecture 7 (Assignment and
Mutability) of

  http://www.cs.rice.edu/~shriram/311/Lectures/

or the last question on the midterm exam at

  http://www.cs.brown.edu/courses/cs173/2001/Exams/

A bit of Web searching coughs up a few more references.  See the first 
few paragraphs of

  http://g.oswego.edu/dl/aliasing/aliasing.html

or search for "aliasing" on

  http://www.cs.wustl.edu/~sck/455/current-semester/hws/hw7/

That should get you started thinking about this problem.  Probably
tens of millions of dollars have been poured into C/C++ compilers
simply to undo the effect of this nasty phenomenon.  Actually, just
search for "alias analysis" on Google, and spend the next ten years
reading.  Here's a nice little survey:

  http://liberty.cs.princeton.edu/Memory/Doc/Primer/survey.html

Shriram