RE: fdo69552 backward compatibility with optional arguments in calc functions

Hi Eike,

> No, I proposed (or at least I think I did, that's what I meant anyway ;-)
> to introduce a new / change the UI names, and only the UI names,
> depending on the number of arguments encountered during reading a file.
> There'll be only one CEILING for the file format and never another one.

(Just to make sure, I am a slow learner)
So, version 4.3 will have a CEILING in the UI, no diffence with previous versions and a CEILING_ODF 
in the UI which complies with ODF1.2.
Both will be saved in files as CEILING.
Upon reading a file (with version 4.3) the UI-function name will depend on the presence of argument 
2 (regardless of the presence of argument 3).
And the object of this is that users of version 4.3 will not leave out argument 2 unwittingly. (For 
users with older versions there is difference, the ODF1.2-compliant CEILING will produce an error 
if argument 2 is missing).

The hitch I see with this is that when a user creates a calc document, enters CEILING_ODF, saves 
and reopend, his function will be 'different' in the user's view.

And to be honest (don't confuse with blunt), how big is the gain of all this as compared to making 
CEILING comply with ODF1.2 in version 4.3? Users who don't read release notes and help will not 
know that argument 2 is now optional. By the time they discover this, older version that 4.3 will 
be really old and 4.3 or newer will be commonly used. And should they read release notes or help, 
they will read that this change is not compatible with older version than 4.3.

Winfried