W3C home > Mailing lists > Public > www-archive@w3.org > March 2009

Re: <time>

From: Leif Halvard Silli <lhs@malform.no>
Date: Sat, 14 Mar 2009 06:44:44 +0100
Message-ID: <49BB444C.8080101@malform.no>
To: Robert J Burns <rob@robburns.com>, Dave Singer <singer@apple.com>, www-archive <www-archive@w3.org>, Charles McCathieNevile <chaals@opera.com>
-public-html
+www-archive

Dave, Rob + Charles (who brought <time> to our attention)

Supposing that the proleptic Gregorian calendar should be used in the 
entire Julian period or so, then, with authors in mind, if Marthin 
Luther's death on the the Julian 18th of February 1546, is recorded in 
@datetime as 18th February, Gregorian - which is incorrect as it should 
be 28th if it is Gregorian, then, "in the wild" we'll see errors in how 
that wrong date is presented/perceived: as 8th of February if that dates 
is "converted" to Julian. As "correct", historic *Julian* date if it is 
*not* converted (even if it is noted wrong). Further, if it is correctly 
noted as 28th February, Gregorian, then  we will se that it isn't 
converted to Julian, thus it will be presented as if he died on the 
28th. This level of confusion is *only* possible for Julian dates. 
Therefore Julian dates should be taken especially care of.

Either authors must be allowed to record the date in another format as 
well as in the Greogorian format, where the other format has the role of 
presentation (the name of the attributes are completely made up, but not 
take completely out of the blue):

    <time julian="0650-05-25" datetime="0650-05-27">25.5.650</time>

And perhaps even *only* in the Julian format - singling out that it _is_ 
Julian:

    <time julian="0650-05-25" >25.5.650</time>

Or they must be allowed to have an attribute that informs that the 
recorded date is not Gregorian

    <time daysBehindGregorian="2" datetime="0650-05-25">25.5.650</time>

Or there must be something that tells that this date should be 
*presented* as a Julian date:

    <time julianPresentation="-2" datetime="0650-05-27">25.5.650</time>

The presentation info could be inside @datetime, or an extension to the 
date format:

    <time datetime="0650-05-27; Julian:-2;">25.5.650</time>
    <time datetime="0650-05-27_-2;">25.5.650</time>

Or else we must decide that for dates from before the Gregorian calendar 
started, the purpose of @datetime is *only* comparison and sorting - 
*not* presentation. (Meaning that authors would have to use e.g. @title 
to convey presentation.)

    <time present-as-ISOdate="no" datetime="0650-05-27">25.5.650</time>  
<!-- the October revolution -->

Because if we do not take care of Julian, then we will end up in the 
dilemma: When is @datetime intended to be presented as a Julian date, 
and when is it supposed to be presented as a Gregorian date? (Because, 
@datetime is supposed to be presented - it isn't only a interchange 
format, like in e-mail.)

Dave, I hope you see that we need to be able to tell how the number is 
supposed to be perceived? The ISO date is supposed to be unambiguous in 
the first place. However, we got to see that it isn't for the times and 
places where the Julian has been used. Authors must be encouraged to 
record dates in a standardized format. And then there must be some 
incentives for them to do so. We should not rely on authors adding the 
ISO date only because they reason that "it is probably good that I do 
so". They must feel and see that they record something that it is 
relevant and accurate to record. There must be immediate benefit, also 
for them.
-- 
leif halvard silli
Received on Saturday, 14 March 2009 05:45:27 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:43:28 UTC