inconsistent definition of "sort" (and what about stability?)

Oliver Kullmann wrote:
>
> Sounds like the reaction of a current politician to me: when pointing
> out an error, reacting with cynism that the world is not perfect,
> and only a fool could expect otherwise.
>

Which is true


> I always hoped (and still do so) that Maxima is part of mathematics,
> and that whatever happens, when speaking to a mathematician the
> least common denominator is always truth (and its simpler cousine,
> correctness).

I don't think computer science has much to do with mathematics. It may have
to do with some branches of mathematics, like discrete maths or formal logic.
But it is also close to physics, a domain where the "politician" maxim above
is quite appropriate.



> Things relating to the perceived mathematical side can be discussed in
> therms of truth and correctness, and this is widely done on the mailing
list.
> Things relating to the non-mathematical side can and must not be considered
> in terms of truths, perhaps better are not discussed (mentioned) at all.
> One sees a degeneration into a hackish attitude ("read the code" or "read
> that document", without discussion of what's in the document --- where
> programming standards are notoriously vague), and a rather strong
> emotional "disavowal".
>
>

It happens that maxima is free software, a quality which is appreciated by a
number of people. This means that:
- we don't pay for it, and as a consequence we don't have a contractual
relation with the authors, which would allow to ask for correction of bugs
(note that most paying software has so called "licences" such that the authors
deny all responsibility about the bugs in their work).
- but conversely we are allowed to look at the source code, and perhaps fix it
at least for us. This is the "hackish" attitude you are complaining above, but
you have to contrast it with the situation for paying software, e.g.
mathematica, where you don't have the slightest idea what is inside the box,
and have to cope with the bugs without any hope of a fix.

For the case of the documentation, it is even easier, anyone, even without any
competence in programming can play with maxima and document what they are
finding. Personnally i don't like the official documentation, i find it almost
impossible to understand in the sections dealing with pattern matching and
similar advanced stuff. However i am not going to complain about the work
Robert Dodier has invested in these sections, because i am not able to do any
better, and it is not tomorrow that i will know these things inside out like
he does. The biggest problem is that it is hard to discover useful functions
or variables. This is the first complaint i have heard from colleagues
"it is impossible to find anything in the doc". The "Categories" stuff
in the html documentation is a useful tool to help hypernavigate the doc,
but the root problem is the general organization, which is difficult to
get fine. All those problems seem to me far more important than the eventual
correctness of the documentation of some obscure feature.


-- 
Michel Talon