Andreas J. Guelzow wrote:
> I think what one really should want is a specification of the serial
> number matching 1904-1-1 and a statement how days and changes in serial
> numbers are related.
Oh, *now* I see what you mean. Sometimes it takes being hit over the
head with a 2x4 :-). Thanks for persevering.
That's a good plan. Would you mind proposing the specific changed text,
since you noticed this? Robert Weir wants to describe the whole thing
as a separate type, but it'd clearly be easier to simply make this
clarification instead. And even if eventually we described it as a
separate type (which I'm not sure we really should do since ISNUMBER()
etc. reveal they're really numbers), we would STILL want this clarification.
--- David A. Wheeler
