> The GYMD proposal is ridiculous for several reasons: > > -- It becomes useless after 127 years, as it cannot > represent dates after year 2129. > > -- It cannot represent dates before the year 1080. > > -- Any attempt to represent a year with only two alphanumerics > [A-Z0-9] necessarily limits the range of dates to about 1300 > years, which is too short to be of general interest. > > -- The unit larger than a year is not based on a decimal > multiple. The proposed 30 year unit is downright silly > (except perhaps for geneologists who devised this scheme). > The 25- or 26-year unit (mentioned elsewhere) is hardly better. > New notations should, wherever possible, resemble metric (SI). > > However I do find its compact notation appealing, and > (for data processing reasons) I greatly favor notations > that have fixed-length fields. Unfortunately, ISO 8601 > is far from compact, partly due to leading zeroes in > two fields and also due to required punctuation. > (Futhermore, pre-emption of the dash makes that character > unusable in its customary role as an indicator for ranges > of values -- but that's an issue for another day. ;^> > > If radically different notations are to be considered, > then my preference is to retain both the year and the > week but to either discard or rationalize the month. > While there is very strong cultural attachment to the > seven-day week, people don't care much that months vary > in length and that they rarely coincide with lunar cycles. > > Because day of the month is so irrationally determined, > people tend to be uncertain and inaccurate about whether > Saturday is the 28th or the 29th (actually, it's the 2nd!), > but usually do have a clear concept of how many days it > is from tomorrow, due to the nearly-universal 7-day rhythm > of our lives. Indeed, week numbers are far more sensible > than month numbers, but perhaps a range of 1 to 53 is a > bit too large and should be subdivided. > > Named months could still be retained, but containing > exactly four weeks each (except for the last one) > and there would be thireteen of them. Each month > may be assigned a unique single-character designator > such as the letters [A-M], thus allowing any date > to be designated with a meaningful three character > designator: [A-N][1-4][0-6] e.g. Today is "B41" > (as opposed to the virtually meaningless "Julian" > day-of-year number or a bloated designator with > four-characters plus punctuation > (i.e. 02-25 for today or 12-25 for Xmas). > > (Yes, I know we could make the notation even > more compact by using a single character for > day-of-month, but I'd happily settle for just > 3-characters replacing 4-chars + punctuation.) > > A calendar where the first of a particular month > always falls on the same day of the week would > not only be very attractive, but would also avoid > much error, confusion, and wasted time. > Even better would be a calendar where the first > of every month is always (say) a Sunday. > > Choice of New Years Day is a trivial distraction > from the advantages of a common week-based notation, > as is the fact that the last month of the year will > vary in length. A simple scheme would be to always > end the year on the first Saturday after the solar > year ends. (Some years will have 52 weeks and some > will have a 53rd partial week; same is true now!) > > The following day (traditionally dedicated for > hangovers, football games, and so forth) would always > be Sunday, and in addition to calling it January 1st > or "-01-01" [ISO 8601] it could also be designated > as A11 -- that is, the first (4-week) month, first > week of that month, and first day of that week. > > I realize that this is probably beyond the scope > of ISO 8601, but if the discussion is going to > other schemes then it is fair to consider the > basic underlying calendar system, too. > > Bruce A. Martin > > P.S. In all of these schemes, there is also a > range problem, which I don't know how to address. > While I'm not too worried about Y10K, which may > invalidate ISO 8601 after 9999-12-31, I am unsure > how (or whether) ISO 8601 refers to the date of > Aristotle's birth (much less, the date of the > Permian extinction) > > bam. > > > > --- In [EMAIL PROTECTED], "Stephen GOULD" <[EMAIL PROTECTED]> wrote: > > Ian - we agree with Peter that you should use a standard coding > structure like GYMD for dates especially for Electronic Commerce. > >... > > On 24 Feb 02, at 23:34, [EMAIL PROTECTED] wrote: > > > > > [2001-Feb-24] > > > Many people recognise that the current calendar schemes > > >... > > > I found a completely new proposal that is radically > > > different to any that I have seen before. This is at: > > > <http://www.gtbaddow4.freeserve.co.uk/>. > > >... > > > > > > > > Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ > >
