On Thu, Nov 19, 2009 at 01:00:11PM +0100, Florent hivert wrote:
> > Is there any particular reason why when one tries to iterate over
> > IntegerVectors(k,n) one gets the vectors represented as lists and not
> > tuples?
Nothing really but a historical choice to port things from
MuPAD-Combinat as quickly as possible. Also, internally the algorithm
uses lists because it needs to modify them.
> > The only difference between the two that I can think of is
> > that lists are mutable and tuples aren't. I cannot see why being
> > mutable is an advantage. On the other hand, if the vectors weren't
> > mutable one could use them as keys to a dictionary e.g.:
> >
> > sage: p = MixedIntegerLinearProgram()
> > sage: v = p.new_variable()
> > sage: p.set_objective(sum(v[i] for i in IntegerVectors(2,3)))
> >
> > Now I need to write
> > sage: p.set_objective(sum(v[tuple(i)] for i in IntegerVectors(2,3)))
>
> Mutable is an advantage in the sense that if you want to modify the vector you
> have to make a copy of it. (This sentence looks like a tautology !). There is
> some plan to make the result type (and parent) configurable. Note that someone
> could find it more natural to get the result not as a double but rather as a
> true vector as we can get using
> sage: v = vector([1,2,3]); v
> (1, 2, 3)
> sage: type(v)
> <type 'sage.modules.vector_integer_dense.Vector_integer_dense'>
> sage: v.parent()
> Ambient free module of rank 3 over the principal ideal domain Integer Ring
> I'm not sure why we (maybe = Mike Hansen) finally chose list. I don't think
> one is better. We should allow both.
We definitely should allow to specify any constructor for list-like
objects (list, tuple, vector). Depending on the context, you may want
the result to be hashable, to have further properties (like being a
vector and doing linear algebra), that is mutable, ...
Now what should be the default? I am not very keen on vector; we are
really returning a raw combinatorial object, that will typically be
used for creating exponent vectors of polynomials.
And this should be done consistently with all other related things
(compositions, ...).
Florent: combinatorial factories are becoming more desirable every day :-)
Cheers,
Nicolas
--
Nicolas M. Thiéry "Isil" <nthi...@users.sf.net>
http://Nicolas.Thiery.name/
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
