Re: (!!) Request for clarification; RFC 1123 sect. 5.2.14

[Erik Naggum <[e-mail]>]

|   But the specification 2*4DIGIT aslo permits a year matching the
|   specification 3*DIGIT.  It is unclear what the semantics of a 3*DIGIT
|   year are, and RFC 1123 is silent on this matter.

I think as we approach the 21st century, we may find that some may want to
use a three-digit number if only by continuing the somewhat naive belief
that they can subtract 1900 from the year number (e.g., Unix systems).
this naive belief is certainly better than the other naive belief that
would write "4 May 19105".  (this is thus not an argument for 2*5DIGIT.)

most systems that today accept a two-digit year will naively believe it to
be offset from 1900, even after the year 2000.  Common Lisp is among the
few that specify +/- 50 years of the current year, which makes a date
specification depend on the "current year" which will be somewhat elusive
for a sufficiently long time span (barring natural or not so natural
disasters, I should live to see "95" refer to 2095), which is also why I
think we decided to recommend 4DIGIT.  if we deprecate 2DIGIT, and nobody
uses 3DIGIT, it will be hard to interpret the RFC as recommending 3DIGIT.

but do we really need to act on this now?  can't we just observe those who
talk about (and, as time goes by, _in_) the next century and see if it is
actually a problem?  I don't necessarily think it is an actual problem.  we
may actually be happy that we allowe 3 digits if it becomes a convention.

#<Erik 3008541659>