RE: [office-comment] Need more detailed definitions for formulas

["Scott Wiseman" <scott@intercore.net>]

Scott Wiseman
Novell Groupwise Consultant
Microsoft Exchange Consultant
Great Plains, MAS90, QuickBooks Consultant
http://www.avidware.net
http://links.avidware.net
http://www.intercore.net
888.603.6333 ext 89

-----Original Message-----
From: David A. Wheeler [mailto:dwheeler@dwheeler.com] 
Sent: Tuesday, October 26, 2004 10:40 AM
To: Michael.Brauer@Sun.COM
Cc: dwheeler@dwheeler.com; office-comment@lists.oasis-open.org
Subject: Re: [office-comment] Need more detailed definitions for
formulas

> > Michael Brauer - Sun Germany - ham02 - Hamburg wrote:
> What I said is not that it wouldn't be useful or even required to get 
> a common standard for formulas within office application, but that 
> this is outside the scope of this TC, at least for the moment. From 
> the office file format view, formulas are some kind of programming 
> model that is very similar to scripts, and even may make use of 
> scripts. For these programming models, the Open Office TC can only 
> establish methods to uniquely identify programmimng models (that's 
> what the namespaces do), but cannot standardize the programming models
themselves.

I think you're unnecessarily conflating "programming models" with
"formulas".
To be useful, the specification needs to specify a minimum interoperable
format for exchanging formulas that may use a set of standard functions.
This specification does NOT need to define ANY way to define new
functions; you can leave that out for a useful spec.

But as I surf through the specification, the amount of the specification
that REQUIRES formula processing becomes larger, not smaller.  For
example, the word processing portion includes variables that can be
computed.  Er... how are you going to do that, without a way to evaluate
formulas?

Thus, I don't see a conflict at all.  The TC can leave out "programming
models" from its initial specification; that means that there may not be
a standard way to define NEW functions.
But the TC should NOT leave out a standard way to evaluate functions; it
needs to define that, and a common set of functions (like SUM).

--- David A. Wheeler