Am Samstag, den 03.04.2010, 14:32 -0400 schrieb Raymond Toy:
> On 4/3/10 1:32 PM, Dieter Kaiser wrote:
> > Am Samstag, den 03.04.2010, 11:10 -0600 schrieb Robert Dodier:
> >
> >> On 4/3/10, Raymond Toy <toy.raymond at gmail.com> wrote:
> >>
> >>
> >>> Ah, in lmdcls, it's pretty clear that unspecial is not supported on sbcl
> >>> and openmcl. We need to fix this. (I think the best solution is not
> >>> to have to do unspecials at all.)
> >>>
> >> The use of special variables in Maxima could be cleaned up a lot.
> >> I wouldn't be opposed to anybody who wants to take on the task.
> >> I am imagining stuff like: change names to use the *foo* convention,
> >> replace special variables with lexical variables where possible,
> >> perhaps wider use of CL packages to isolate symbols used in
> >> different contexts. (If we did that stuff, I think we could do away
> >> with unspecial entirely.)
> >>
> > Perhaps, some unknown and known bugs will vanish too, if we change the
> > use of a lot of special variables.
> >
> > One of these variables is EXP. It is used a lot and it is declared to be
> > special.
> >
> > Another variable I do not like to use is VAR. Here we have to change the
> > implementation of schatchen or m2 to allow passing the value of VAR as
> > an argument. Furthermore, a lot of test functions use the global value
> > of VAR.
> >
> > I think there are a lot of places, where we should pass the value as an
> > argument of the function.
> >
> I would be happy if we just renamed them as *var* and *exp* to make it
> clear about their special nature. Same with all the other specials. Of
> course $<foo> is always special, or should be if it's not already.
Yes, and if we replace exp with *exp* we will see the many places a
special variable is not necessary.
Dieter Kaiser