ufl team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Sat, Apr 25, 2009 at 07:16:54PM +0200, Martin Sandve Alnæs wrote:
> On Sat, Apr 25, 2009 at 4:23 PM, Anders Logg <logg@xxxxxxxxx> wrote:
> > On Sat, Apr 25, 2009 at 03:10:07PM +0100, Garth N. Wells wrote:
> >> On Apr 25 2009, Anders Logg wrote:
> >>
> >>> On Sat, Apr 25, 2009 at 11:40:35AM +0200, Martin Sandve Alnæs wrote:
> >>>> I've verified the curl vs wikipedia and defined rot as the z-component
> >>>> of the curl of the 2D vector operand embedded in 3D. Is that right?
> >>>>
> >>
> >> rot is just a synonym for curl. For simplicity, I would remove rot.
> >
> > I'm used to the following notation (in pseudo-math):
> >
> >  rot:  R^2 --> R
> 
> So this can be defined as
> 
> nabla x (a(x,y), b(x,y), 0)
> =
> |  i  j  k |
> | ,x ,y ,z |
> |  a  b  0 |
> =
> (b,x - a,y) k
>
> rot( (a,b) ) = (b,x - a,y)
> (or curl)

Yes.

> >  curl: R   --> R^2
> 
> How is this defined?
> 
> curl( a(x,y) )
> =
> nabla x (0, 0, a)
> =
> |  i  j  k |
> | ,x ,y ,z |
> |  0  0  a |
> =
> a,y i - a,x j

Yes.

I guess we can let curl cover all three cases and then just make rot
an alias. Then everyone will be happy. The rot-curl notation will work
but it's also possible to just write curl.

-- 
Anders

Attachment: signature.asc
Description: Digital signature