Re: [office-formula] DATEVALUE and TIMEVALUE with full DateTimesupport

Eike Rathke:
> Yes, it does. ISO 8601:2004 section 4.3.2 "Complete representations" (of
> 4.3 "Date and time of day") says that the character 'T' shall be used as
> a time designator, and adds a note that by mutual agreement of partners
> in information interchange the character may be omitted if the
> representation is unambiguous.

Okay, that's good enough for me, let's make that a SHALL and see who complains.

BTW: We need to use SHALL, not MUST, since "SHALL" is the official ISO term.  Yes, I know that MUST is what you use for IETF documents, but we're not submitting this to the IETF. I've tried to fix a few places, but I'm sure more work is needed.

> here are some good public sources:

... Thanks!  I actually know about those, I just wanted to know what the official ISO spec says.  I couldn't check, because:

> Since ISO makes you pay (currently 126.00 CHF for ISO 8601) for copies
> of standards documents,

*Sigh*.

> > We'll need to add tests to check for all this, esp. since we KNOW of implementations that don't do this.
> 
> Could you do so before uploading the version of the document I sent you
> last night?

Yes, happy to. I've added them, so they'll be in the next post.  I'm merging in my work now.

BTW, I'm realizing we have locale-dependent numbers redux in the VALUE function.  Should VALUE("1.151") be accepted everywhere?  VALUE("$1,000.00")?  How about VALUE("3,14159265")?  Excel certainly accepts VALUE("1.151") and VALUE("$1,000.00") in the en_US locale, I'm not sure what others do... or what we SHOULD require.

--- David A. Wheeler