My (electronic) copy of ISO 8601:2004 says (the html tags are added by
me to indicate boldface):
expansion of a representation to allow identification of dates in
calendar years outside the range  till 
The introduction to the standard explains that square brackets are to
distinguish date/time representations from the text of the standard, and
the square brackets are not part of the representation. Since 0000 is
withing the range 0000 till 9999, I would use four digits for the year
zero, not 6. Also, I am at a loss to understand why you would want to
use a negative sign with the year zero.
The text provides further support for this interpretation in section
3.2.1, The Gregorian Calendar. It states:
NOTE In the proleptic Gregorian calendar, the calendar year
 is a leap year.
On 2010-08-05 6:59 PM, Ray Denenberg wrote:
> " Please see page 7 of ISO 8601:2004."
> I don't think our copies are paged alike, could you cite by section instead
> of page?
> Anyway, I think it should have said "The year zero is represented as
> '-000000'." rather than "The year zero is represented as '-000001'."
> And 1 BC would be the same as year zero thus '-000000'. Right?
> -----Original Message-----
> From: Discussion of the Developing Date/Time Standards
> [mailto:[log in to unmask]] On Behalf Of ashtongj
> Sent: Thursday, August 05, 2010 6:43 PM
> To: [log in to unmask]
> Subject: Re: [DATETIME] EDTF Features list
> I believe the consolidated table contains an error. Section 2 subsection
> "Extension and Precision" rightmost column last bullet says "The year zero
> is represented as '-000001'." But of course the notation other than AD/BC or
> CE/BCE that is most widely known is astronomical year numbering, in which
> the year 0 would be represented as 0000 in ISO 8601 notation. It is the year
> 1 BC that would be represented as '-000001'
> when a profile had been agreed to that required six digits for expanded year
> Please see page 7 of ISO 8601:2004.
> Gerard Ashton