I want to be sure we are clear that we cannot use  '?' (question mark)
because it is used for a different purpose, for example,

2004-06? "questionable year-month"
2004-(06)? "year known, month questionable"
(2004)?0611  "questionable year; month, day known"

We need to dedicate a character as a single-character replacement (which '?"
doesn't do in the above). If we use '?' that would make the syntax


> -----Original Message-----
> From: Discussion of the Developing Date/Time Standards
> [mailto:[log in to unmask]] On Behalf Of Edward C. Zimmermann
> Sent: Wednesday, November 24, 2010 5:58 AM
> To: [log in to unmask]
> Subject: Re: [DATETIME] Proposal to change unknown marker from 'u' to
> 'x'
> On Tue, 23 Nov 2010 11:02:48 -0500, Ray Denenberg, Library of Congress
> wrote
> > The problem is that 'x' is often used to mean "any and all digits",
> > so that "8xx" literally means "all values 800 through 899" when what
> > we want it to mean is "one specific value in the range of 800 through
> > 899 (inclusive)".
> On Mon, 22 Nov 2010 19:30:15 +0100, Jakob Voss wrote
> > Wikipedia already uses this notation, which should be argument enough,
> > for instance:
> >
> >
> The said Wikipedia page sets 19xx to REDIRECT to 20th century which is
> Ray's suggestion of "any and all digits".
> 196x in Wikipedia
> is not even defined. So its not all that consistent and not part of
> their
> policy--- as if their policies truly mattered to us--- but a page
> contributed by a young advertising website copyrighter named Nicholas
> Moreau (aka "Zanimum").
>     * (cur | prev)  16:33, 20 March 2007 Zanimum (talk | contribs) (26
> bytes) (←Redirected page to 20th century) (undo)
>     * (cur | prev) 16:32, 20 March 2007 Zanimum (talk | contribs) (19
> bytes) (←Redirected page to 1900s)
> Please note that originally the page redirected to 1900s. That page was
> created in 2007. A stub for 18xx was created in 2005 and 20xx in 2008.
> Note also that other than 18xx, 19xx and 20xx I don't think other pages
> have been created. I found nothing under 196x or 1x66 or x999. When you
> look at 196x or 1x66 or x999 what do you immediately think?
> While our expectations of 19xx when viewed as a date lead us to assume
> that x is some single non-negative decimal integer, I would even argue
> that x has not been specified to be a non-negative single digit decimal
> integer and thus, following the abstract model, could be anything.. an
> argument, I think, that also holds against the 'u' and space ('#') of
> the MARC domain. In many computer textbooks the convention n for non-
> negative decimal, o for octal, h for hexadecimal etc. single digit
> integers are popular. I don't suggest these.
> We want, I suggest, to be able to express both the intent of 19xx as
> representing the 20th century and 19xx as representing a specific year
> date in the 20th century.
> We have, in contrast to the bibliographic date example, two different
> kinds of vagueness as well as states of knowledge:
> 19?? where it can be meant sometime in the 1900s (one specific value
> that just not known) and 1900s which can mean (but does not need to)
> all values "1900 through 1999".
> We need also to distinguish between expressions about the 1900s that
> refer to the 20th century and those that refer to the first decade of
> that century.
> While the colloquial use of the expression "1900s" typically refers to
> the turn of the 19th to 20th century the expression "1300s" more
> typically is used to address the 14th century than its dawning decade.
> In the bibliographic model there is also the "unknown" (marked | and
> not 'u') to denote something that might be known but has not yet been
> collected.
> So for 19?? we have three semantics for '?':
> 1) A specific value currently unknown. Example: 195? to represent a
> book known to be published in the 1950s but with unknown year (I
> actually have a few books that don't have copyright dates but from
> later additions know, more or less, roughly when they were published)
> 2) A specific value not yet collected: to be filled in at a latter time
> (placeholder).
> 3) A implicit expression of lesser precision. The 1950s, for example,
> referring not to a specific year but to a view of dates roughly in
> decades just as 1929 may refer to an explicit event in 1929 (such as
> the 29 October stock market crash) or a series of events such as the
> "Great stock market crash" (spanning from no later than 3 September
> 1929 when the Dow hit its high to no later than 1932 when the market
> bottomed out at under 10% of value), to the whole of the year or to a
> watermark (the Great Depression started in 1929).
>   I think we need to kick around another
> > character, if 'u' is considered unacceptable.  I find it unfortunate,
> > because I think that 'u' is the most natural character to use for
> this
> > purpose.
> I don't find it terribly natural but accept its use case in MARC21
> which is, I think, sufficiently strong to warrant its use here.
> --
> Edward C. Zimmermann, NONMONOTONIC LAB
> Basis Systeme netzwerk, Munich Ges. des buergerl. Rechts Office Leo
> (R&D):
>   Leopoldstrasse 53-55, D-80802 Munich,
>   Federal Republic of Germany
> Telephone:   Voice:=  +49 (89) 385-47074  Corp.Fax:= +49 (89)  692-8150
> Umsatz-St-ID: DE130492967