Re: struct variable definition (in var.h)

[Ben Pfaff]

John Darrington <address@hidden> writes:

> Currently I'm not too impressed with the system we've been using for
> storing per procedure info which has a rather cumbersome general
> purpose storage area in the definition of the variable type:

[...]

> I think there's a lot of problems with this approach.  To name some:
>
> 1. Doesn't promote reuse between procedures.
> 2. Makes no provision for factored variables.
> 3. There'll end up being a lot of members to that union
> 4. Makes var.h dependent upon almost every procedure definition.
>
> Does anyone object if I rework it?

It's fine with me.  I've been unhappy with this approach for a
long time, and it'd be great to get rid of that ugly union.  The
only reason I didn't change it was that I didn't have an idea I
was really happy with.  What's the plan?

By the way, when it comes to FREQUENCIES I think that this is
useful enough that we should try to make it into a library, as we
do for, say, moments and Levene calculations.
-- 
address@hidden - address@hidden - address@hidden - address@hidden