DEFVAR and DEFCONST

["Guy L. Steele, Jr." <GLS at MIT-AI>]

    From: David Chapman <ZVONA at MIT-AI>

    I sometimes change the value of a constant in the source and then would like to
    be able to use c-t-c to reset it. GLS's suggestion would not allow this.

    MOON@MIT-MC 04/21/80 20:25:32 Re: DEFVAR and DEFCONST
	I would recommend that DEFCONST not always reinitialize the variable
	involved, but should complain if the variable is bound and is
	not EQUAL to the newly computed value?
    Why?  The point of DEFCONST is to be the one that always reinitializes
    it, so that you can change the value.  What did you have in mind with
    this suggestion?

Having this sort of error-check would guarantee consistency in the case that
a DEFCONST for a given constant appeared in more than one file (which would
be useful for various reasons).  Maybe there should be a third thing --
but it seems to me that a "constant" should be constant.  My suggestion
would allow you to change the constant, but only after an error break.
Maybe the criteria for giving an error should be the same as those for
giving an error in FDEFINE.