Section 4.6 Reference: TBDs

[Eike Rathke <erack@sun.com>]

Hi,

Section 4.6 Reference says:

| A reference contains 0 or more areas, in order.

Which order? There is none, is it?

| TBD: Need more details here about internal structures of what references
| point to. This one needs fixing.

I don't get what exactly was meant there.

| Range concatenations like A1:B3:D2
| create a single range. See AREAS(), for example.

Shouldn't we simply define RangeOp there? Or elsewhere?

| What about references to empty ranges of cells?

As in the intersection case already mentioned producing a reference
error?

| Can you distinguish A1:A1 from A1?

Well, most times ;-)  but IMHO it isn't necessary in interpreter
context, a range reference resulting in a single cell should be treated
as a single cell reference. Example: enter =A1:A1 anywhere on a sheet,
you'll get the content of A1 as the result. Do that with =A1:A2 and
you'll get an error anywhere except on rows 1 and 2. Which leads to the
yet missing definition of

- Implicit intersection

  For arguments that represent a single Number, not NumberSequence,
  a given area reference results in a reference to a single cell,
  depending on the cell location of the formula cell. If intersecting
  the sheet/row/column of the formula cell's position with the reference
  results in exactly one single cell, the reference of that cell is
  taken. Else an error condition is generated.


| How do you handle 1:3?

Not possible yet. The ODF definition doesn't allow it as
a CellRangeReference (correct term and section to be looked up in ODF)
and needs to be changed.

| Do array functions change anything here?

Yes, they do. For example, in array formulas there is no implicit
intersection of area references. Instead, arguments are evaluated for
each array element.

It turned out that Jody and I both intended to work on an array
definition, we'll best have to coordinate our work.

  Eike

-- 
Automatic string conversions considered dangerous. They are the GOTO statements
of spreadsheets.  --Robert Weir on the OpenDocument formula subcommittee's list.