emergency-comment message
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]
Subject: What does a CAP date look like?
- From: "Bob Wyman" <bob@wyman.us>
- To: <cap-list@incident.com>, <emergency-comment@lists.oasis-open.org>
- Date: Sat, 6 Mar 2004 21:05:07 -0500
Title: Message
CAP specifies that dates should be in ISO 8601 format. However, ISO 8601
specifies not just one format for dates but a number of alternative formats.
Unless it is the intention of the CAP authors to require that all implementers
of CAP be prepared to accept any date in any of the numerous formats defined in
ISO 8601, it would be useful to modify the CAP specification to explicitly state
*which* ISO 8601 defined format or formats are intended to be
used.
At the simplest level, while ISO 8601 defines a number of representations
for how timezone is to be represented, it does not *require* that a
timezone be present in a date. This is, of course, an issue that has caused
endless debate in numerous protocol and format definition groups in the past.
There are those that will argue passionately that timezones MUST be provided (I
am among them) and others who will argue that they should be optional. What is
the position taken by CAP? In any case, some consideration should be given to
how a timezone should be represented. According to ISO 8601, all of the
following identify the same point in time:
12:00Z = 13:00+01:00
= 0700-0500
Which representation
is correct for use in CAP?
Also, ISO 8601 supports at least the following alternative date
formats:
20040306, 2004-03-06,
04-03-06, 040306
One can also say
2004-W01, 2004W01, or 04W01 to indicate the "first week" of
2004...
And, of course,
2004-W01-2 would indicate the second day (Tuesday) of the first week of
2004...
Times in ISO 8601 are available in all sorts of forms. Thus, you can say
23:59:59.2234 or 235959.2234. And, when you form a "date/time" the "T" which
separates the "date" from the "time" part is optional... Does CAP require it? If
not, why not?
Simply referring to ISO 8601 unfortunately does not result in an unambiguous
definition of date/time formats unless one intends to require that all CAP
implementers are prepared to implement a date/time specification which is
significantly more complex then CAP itself. It would be very useful if the CAP
specification were to identify the profile of ISO 8601 elements that were
supported. This is the normal thing for protocol or format specifications to
do.
bob wyman
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]