Re: [office-formula] Syntax issues

["David A. Wheeler" <dwheeler@dwheeler.com>]

Original text:
> > >> ='Sheet1'.$$Hello            # Look for Sheet-local name first, THEN global.

Eike Rathke said:
> > > I don't get it why a global name should be tried at all in this
> > > constellation. Of course one _can_ do, but writing the sheet prefix plus
> > > global name would be an error, or what am I missing?
> > >   
> > Hmm, this construct could have two different possible semantics.
> > If possible, we should be very specific.
> > 
> > I think we ought to duplicate whatever the typical semantics are
> > for applications that support sheet-local names.  I know Gnumeric and
> > Excel have them - can someone double-check their semantics?

Andreas J. Guelzow said:
> Gnumeric will default to the global name. In fact without that fall back
> there would be little need for sheet-local and global names since then
> sheet-local names could be eumulated by a naming scheme. 
> 
> For example I want to be able to access the database used on Sheet1,
> which may be the global database.

Um, agree, but I don't think that's what Eike was asking about.
Obviously what I wrote was unclear - sorry, let me try again.

Clearly, in this construct:
> > >> ='Sheet1'.$$Hello
You better look at sheet Sheet1 for the variable "Hello".

But let's say that you have a sheet-local reference, but there is
NO sheet-local variable defined (no "Hello"); what should the spec require next?
1. Look for a global name, and return that if found (and error if not).
2. Error out, because there wasn't a sheet-local name?
3. Do either of the above (implementation-defined which one)?

Current text is (1).

--- David A. Wheeler