Re: Issue 163, was: Meaning of invalid but well-formed dates

On 20/05/2009, at 12:35 AM, David Morris wrote:
> On Tue, 19 May 2009, Julian Reschke wrote:
> 
>> -> 50 years from when? Does the semantics of the message depend on when you look at it?
> 
> 50 years from now ... sure ... the problem with 2 digit years is well known and has been for more than 15 years. So this is simply a bad data fix up which has essentially no risk of a bad outcome. In the case where the recipient knows of a bad potential outcome for the wrong interpretation, such a date should be rejected ... (I don't feel a need to
> say this in the spec).

If we're going to disallow producing these dates when HTTPbis publishes, it seems like it would be reasonable to choose a fixed date -- say, Jan 1 2050?

I doubt that any valid HTTP messages were generated before 1950, and I *hope* that any remaining implementations that generate two-digit dates will be gone by 2050...

Cheers,


--
Mark Nottingham     http://www.mnot.net/

Received on Sunday, 23 May 2010 02:35:16 UTC