PROPOSE-DISCUSS
I agree with this recommendation and the analysis about changes in the current working draft of OpenFormula.
It will be important to say something about the preservation of white space as it occurs in the formula when presented for human inspection, since the layout may be an important aspect for human understanding and review/modification of the formula. Not sure how we express this.
I also think we still need to come at the layers.
It appears that we are specifying a reference syntax anchored squarely in Unicode code points. The limitations on what of such code points can actually arise in a concrete encoding and what can be preserved when saved in a concrete encoding is an interesting one that we must acknowledge somehow. Likewise, some concrete encodings have their own rules for treating white space before it will be seen at the OpenFormula level. (They may have escape mechanisms for forcing white space delivery, too.)
per 1/12/09 TC call: Delegated to Formula SC