Date and time format (Was: Re: Euro currency sign)

Jukka Korpela (jkorpela@cc.hut.fi)
Thu, 23 Oct 1997 09:07:31 +0300 (EET DST)


Date: Thu, 23 Oct 1997 09:07:31 +0300 (EET DST)
From: Jukka Korpela <jkorpela@cc.hut.fi>
To: www-html@w3.org, www-international@w3.org
In-Reply-To: <01bcdf09$4ca7db20$371cd9cf@davidp>
Message-ID: <Pine.OSF.3.96.971023083806.19856A-100000@torvi.hut.fi>
Subject: Date and time format (Was: Re: Euro currency sign)

On Wed, 22 Oct 1997, David Perrell wrote:

> Chris Lilley wrote:
> >and there is no difficulty in interpreting the European local form
> >3 February 2004.
> 
> Yes, but 2004-02-03 sorta better.

European local form? Well, in that case "kolmantena helmikuuta
kaksituhattaneljä" is a European local form, too.

The notation from which 2004-02-03 is an example has rarely been used in
actual practice, despite being an ISO standard approved years ago.
I (as well as probably many others) had thought that the standard
should be just forgotten. Well, _I_ have changed my mind, especially
when reading the HTML 4.0 draft.

In that draft, the datetime attribute of INS and DEL is required to be
in "ISO date format", referring to ISO 8601. On the other hand, other
specifications related to the Web, such as HTTP protocols, allow
several date and time formats but _not_ the ISO format. One might expect
to see some remark about the necessity of changing this. Notice that
HTTP headers are mostly generated and processed by programs whereas
INS and DEL can be and often will be written by human authors, so it
is somewhat strange that in the _latter_ case one has to use a format
which looks unnatural to most of us (mainly because of not being widely
known and used, but anyway).

There is, however, a very good reason to use the ISO format within HTML:
it is language-independent (no month names), and among
language-independent notations it is the only one which can reasonably
be assumed to avoid ambiguity.

Moreover, requiring _full_ ISO date and time format, although somewhat
inconvenient to HTML authors, is necessary to resolve ambiguities arising
from the fact that there are different time zones. When 3rd February
begins where I live, people in the US will still be living 2nd February.

Naturally, authors can use whatever format they prefer in the _text_ of
their documents. It might take time before the ISO format is so widely
known that it makes sense to use it in normal texts, too. But for
_metainformation_ there is hardly any reason to specify any other format.

Yucca, http://www.hut.fi/u/jkorpela/