On 7/30/13 9:22 AM, Robert Dodier wrote:
> On 2013-07-30, Jaime Villate <villate at fe.up.pt> wrote:
>
>> Well, wxmaxima, sage and any other projects that use Maxima should
>> follow the rules that Maxima developers has been using long before those
>> projects appeared.
> There aren't any such rules, so how could they follow them?
Certainly not laws. However it is a convention that was
there from early times. I remember a discussion I had with Joel
Moses who suggested it. He initially thought %i and %p. I suggested
%pi instead of %p.
>
> There is nothing in the code to force names beginning with % to be
> constants; there is no requirement that constants begin with %; there
> is no documentation to suggest % be used only for constants.
Except that it is used for %i and %pi.
> (And before anybody suggests doing any of that, let me mention that
> I'm against it.) Some existing packages, written by long-standing
> Maxima developers, use % names for things other than constants.
> Do you still want to claim there is a "rule" ?
As you know, the package developers were probably trying, in a partly
ineffective way, to make a separate private namespace by using funny
character prefixes. % is a funny character that is allowed as
the first character of a name, at least with default syntax.
I think we should not require %constant, but also that binding
of a constant should force an error. I think we can follow common
lisp here, defconstant handling. But I have not studied the consequences
ofdoing so.
RJf
>
> best
>
> Robert Dodier
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima