Florent Hivert <florent.hivert@...> writes:
> is the easiest way for the end user. Also due to the fact that there
> are not typed, I don't see a way to let something like
> coerce(p, combinat::permutations::cycles)
> works. Is it true ?
As long as you can write a single function that gets whatever p you
put in and does the right thing, it works. The above call first
tries to ask p::dom::convert_to(p, combinat::permutations::cycles),
which fails because DOM_LIST doesn't know anything about combinat.
Then, combinat::permutations::cycles::convert(p) is called, which
must analyse p and decide what is to be done. On the other hand,
having a single case where this can't be done may already be
sufficient reason not to do it at all.
> I was thinking about one such pages per combinatorial objects which
> contains a list of hyper-links to pages on specialized functions. So
The idea behind having a formal structure for our help pages is to
aid the user. Anything that actually does help the user is ok by me;
consistency is only a means to an end, not a goal in itself.
That being said, I would still love to see as much structure of
either the function help pages or the domain documentations as
possible in the combinatorics package. We have put a lot of thought
into this structure and it seems to work with most of the users most
of the time. Obviously, repetitio non semper delectat and
information that would be repeated on more than 80-90% of the pages
should go into some separate place such as a glossary or introduction
and be referenced, as we did with terms like "arithmetical
expression" (see ?glossary for more examples). It might be
worthwhile including a special section in the help pages entitled
"standard methods/entries" or something similar, just containing a
list of links. Or is all of this nonsense and shows my ignorance on
the structure of the library/package?
> one page on reduced words with
> toReducedWord, toReducedWords, fromReducedWord, toLenght
We try to keep such lists ordered alphabetically unless there is
good reason not to.
> By the way, we are planning to have some user of our package :-).
Oh yes, I thought about having a MuPAD user, too. :-)
> And we need to integrate our doc in the standard system. So that the
> user can type
> ? combinat::permutations::toCycles
> Can you provide a way to do that. As far as i know, for the moment it
Not for 2.5 (which is out of the door). This has been on my
ToDo-list for a while and definitely will not fall off it, but I
can't make any promises about the "when".
Attachment? Nein: http://piology.org/ILOVEYOU-Signature-FAQ.html
Christopher Creutzig, MuPAD group, Tel.: +49-5251-60-5525