asksign, was: proposal to modify op and args

I haven't thought much recently about the asksign
question, but there the number of answers is
limited to two or maybe three.  So we can at least
think about chosing one of them and continuing.
Or maybe making several choices and pursuing them
in parallel, sort of.

(I think Stavros said essentially this next theme....)

Regarding other interactions with the user, like
"you asked me to do something which I cannot do"
it seems to me that the reasonable thing for us to
do is assume a program will have associated with
it an error handler, and in the case of a program
run in "batch" mode, there may be a default (e.g.
skip to next command) that
is different from interactive (e.g. Maxima debug/break).

Maybe Java's method of handling would be more congenial
that lisp's error handler, but probably less comprehensive.
I don't know much about the Java setup.

RJF




Robert Dodier wrote:

> On 09 May 2006 15:51:06 +0200, Albert Reiner 
> <areiner at tph.tuwien.ac.at> wrote:
>
>> I also dislike that assumption, but the real problem are the
>> questions, not the errors that can be caught easily.  AFAIK there is
>> no way to tell Maxima to simply give up instead of asking questions.
>> Or is there?
>
>
> There is no way to disable all asksign questions.
>
> There is assume_pos and assume_pos_pred which cause asksign to
> assume the expression is positive under some circumstances.
> ? assume_pos_pred shows some examples.
>
> However there is no way to work around asksign, and there is yet a third
> kind of asking mechanism which is used only in the integration code
> to track half-integer values, if I remember correctly.
>
> I made an attempt to revise the ask-stuff so that a conditional
> expression if <positive> then <value-1> elseif <negative> then <value-2>
> else <value-3> could be returned. See:
> http://maxima.sourceforge.net/wiki/index.php/Noninteractive%20mode
> It's no more than a proof of concept -- it is susceptible to stack
> overflows, probably from failing to catch some conditions.
>
> If you have any comment on this topic, I would be interested to hear it.
>
> Robert
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima