Re: [whatwg] <time>

Andy Mabbett wrote:
> In message <CC3986D1-6DDC-4007-8BBA-42A5D4E398CA@eatyourgreens.org.uk>, 
> Jim O'Donnell <jim@eatyourgreens.org.uk> writes
> 
>> This is already a solved problem in the Text Encoding Intiative  (TEI). 
>> The value of a date/time is encoded in the Gregorian calendar,  using 
>> ISO8601. The calendar attribute is used to indicate the  calendar of 
>> the original, written date enclosed in the tags.

I'm not sure why the original calendar would need to be indicated in the
'calendar' attribute. It does not matter for the 'value' or software in
general, if I've understood correctly. If the 'value' is always in
Proleptic Gregorian calendar, it's all the software needs to know.

>> eg. from the TEI docs for dates and times
>> <date calendar="Julian" value="1732-02-22">Feb. 11, 1731.</date>
>> I suggested that the calendar attribute be adopted in HTML5, as it 
>> would be useful to those of us who mark up historical texts in HTML.
> 
> That's one possible solution - better than none - but I do wonder why 
> we'd force authors to manually convert dates, when we all have machines 
> which can do that.

The author can use machine to convert the date. You cannot expect all
users (the readers of the output of said author) to have machines to
convert any date from any calendar to another calendar.

I think that we can expect the user (reader) to have machine that can
convert from Proleptic Gregorian calendar to user's own calendar system.
(There's still the issue if the day before 0001-01-01 is 0000-12-31 or
-0001-12-31 and it must be defined in the spec if years less than 1 are
ever allowed.)

-- 
Mikko