Ralf Hemmecke wrote:
>
> On 08/19/2014 09:25 PM, Waldek Hebisch wrote:
> > I am thinking about changing interface to ConstantLinearDependence.
> > Original version has list of derivatives as parameter to the functor.
> > However, after using it a few times I think that nicer interface
> > will put list of derivatives as parameter to each function.
>
> You may have good reasons to change the interface, but I'm somehow not
> yet convinced. Not that I really understand the impact (it's probably
> very low since that package is only used in rdeefx.spad and furthermore
> that constructor only seems to be an auxiliary package that is just
> needed for the Risch DE), but a change currently feels a bit odd to me.
> The reason for this is that the list of derivatives is something that I
> consider to be more or less constant, i.e. I'd expect just a few domains
> created with different lists of derivatives.
>
> Waldek, you are probably currently much more deeply into this
> integration business, maybe you can convince me that this "list of
> derivatives" is a parameter that is more frequently changing.
Well, in current use derivatives correspond to variables. "current
variable" is stays constant during execution of several routines,
but user is free to choose any variable name he/she wishes. And
integration code is free to perform change of variables when
meeded. So it does not matter that variable name changes rarely,
to allow easy modification of the code we allow change of variables
and pass variable name around. Current use pattern is:
c_lindep := ConstantLinearDependence(R, F, [(c : F) : F +-> D(c, x)])
sys1 := reducedSystem(m1, vector([dg])$Vector(F))$c_lindep
That is, variable is a parameter and we need to create domain before
each call. Note: it does not matter that typically 'x' will be the
same as in previous call. It _may_ change and we need to dully
call constructor. For me this is more clumsy than passing derivatives
as an argument to 'reducedSystem'.
--
Waldek Hebisch
[email protected]
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/fricas-devel.
For more options, visit https://groups.google.com/d/optout.
