On Thu, 2006-07-09 at 17:11 +0200, Eike Rathke wrote:
> Yes, much clearer. This seems to be the only usage scenario where the
> fallback to a global name actually could make sense. I think all other
> situations just would add to the user's confusion. Imagine the deletion
> of a sheet-local name where also a global name exists: suddenly all
> calculation changes.
Well, in any case if you delete a used name the calculations will
change. If you don't fall back to a global name you would just get
errors.
>
> Btw, does Gnumeric implement that? It doesn't seem to in v1.4.3 that
> I have on Debian/sarge, at least I wasn't able to define a global name
> without sheet name, nor two identical names, one global and one local,
There used to be a bug in the interface used to define names. That has
been fixed in 1.7.0
> and a local name is not copied when duplicating a sheet.
That is still the case. And I am not sure whether local names ought to
be duplicated or not.
>
> Anyway, to me a reference Sheet1.$$Name2 should imply that when there is
> no sheet-local name Name2 defined for sheet Sheet1 it is an error.
Suppose there is a global name Name2 and no local sheets. And Name2 is
defined as [.A1]. Then
On Sheet2, how can you access Name2 as interpreted on Sheet1?
As far as I am concerned $$Name2 should translate as [.A1]
and Sheet1.$$Name2 as [Sheet1.A1]
Don't try this in Gnumeric. Currently we don't let you define [.A1] as
an expression.
Andreas
>
--
Andreas J. Guelzow, Professor
Dept. of Mathematical & Computing Sciences
Concordia University College of Alberta
