Re: [office-formula] Whitespace fixups

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

Eike Rathke: 
> Should we add U+00A0 NO-BREAK SPACE to this list? I don't think we'd
> need the other no-break space characters though.

That sounds like a good idea, though we'll need to be clear that it's
the no-break-space character that is whitespace.
(OpenFormula can be embedded in documents that use
other encodings, like Windows-1252).

> > An embedded line break shall be represented by a single newline
> > character (U+000A), not by a carriage return-linefeed pair. When
> > embedded in an XML document the newline character is typically
> > represented as “&#0A;”.
> 
> Using the term typically (remember TC discussion) might need
> clarification on what,when,why,which,... I suggest to avoid it and
> simply remove the last sentence, leaving only the definition to use
> a newline character.

We should replace "is typically" with "may", and avoid the issue entirely.

> So, if CR and LF are both considered whitespace characters, which
> I second, and an embedded line break shall be written as LF, which
> I second as well, this actually means that an application must accept
> a CRLF combination when reading a document, but not write CRLF back to
> a document and transform it to LF instead.
> 
> What about a single CR? Should it be treated as a newline and written
> back as LF? Or should it be preserved, if the application supports that?
> 
> Note: Applications not capable of preserving all whitespace characters
> should be allowed to omit them when writing a document.

I omitted the text afterwards, but what it said was that applications
SHOULD retain the whitespace characters.  My thought is that
if an application intends to write a line break, then it should write \n.
But once it writes whatever whitespace it writes, other apps should
retain it.

Comments?

--- David A. Wheeler