Re: set.lisp redefines POWERSET / {} for sets

On Thu, 2003-03-06 at 07:33, C Y wrote:
> --- Martin RUBEY <rubey@labri.fr> wrote:
> > 1. operators in the interactive session should *default* to there
> > most common usage (and it should be easy to change that, it is in
> > maxima I believe)
>
> How about this:  We create a sample init file that makes these
> assumptions and let the user put it in place if they so desire. We
> document it thoroughly, so it becomes both a useful tool for newbies
> and a tutorial for the init file.  That way, people could tweak it to
> their own liking for say a class environment, so it behaves the way
> they want it to.
>
> > 2. a program written in the maxima language should *not* make use of
> > these operators. If it does, it is not "shareable".
>
> This is one of my major reasons for not wanting to make the above init
> file the default configuration - people should have to seek out the
> simplifying assumptions and know they are not standard before using
> them.  Otherwise people will just start writing basic programs with the
> simple assumptions and not understand why they don't work for someone
> else.  We will probably get some of that anyway, but at least this
> would raise the bar a little.

A different argument would be that a regular user (I am thinking of a
freshmen student here) should not need to move an init file into
position to get the expected behaviour. On the other hand it is
reasonable for a developer to know or at least to check what special
configuration have been set in his init file.

>
> Another thing we could do would be to have the basic init file print a
> little warning/info about what it was doing.  Something very brief but
> which lets the user know this isn't the standard configuration.  We
> could make that a standard for when anyone creates this type of
> behavior altering init file.  Not that we could enforce it, of course,
> but at least start out that way.

That does make sense to me.

>=20
> Perhaps, if this becomes a popular thing for class environments, we
> could create a repository of init files which customize behavior for
> physics, engineering, etc.  Of course when things hit that level it
> almost makes as much sense to make a share package that does that job.
>
> What do other people think?
>

I have used Mathematica, ISETL and MuPad in class environments. Both
Mthematica and MuPad required init files to make them work remotely
acceptably for my purposes. I encountered the difficulty that it is easy
within the university environment to ensure that each student has the
appropriate init file, students also started using it at home on
different platforms in which case it was difficult for many of them to
place the init file with the correct name in the correct place. I would
therefore like it if some of these in my mind basic settings were
included by default.

Andreas
--
Andreas J. Guelzow <aguelzow@taliesin.ca>
Taliesin