Common usage sometimes creates new meanings.  When I was growing up it was
said that "ain't" ain't in the dictionary.  (I never looked it up at the
time.)  Now ain't is very popular.

Even though the offical Julian version (see web site, thanks Wirt) is
different, the HP Calendar intrinsic returns it as YYDDD.  So 01/31/01 is
101031.  (BTW, the Calendar intrinsic is Y2K27 compliant.  I hope HP fixes
it before the last minute.)

However, even more important to common communication: Look at just about
any desk calendar and you will find a day count based on the day number of
the year.  Most people I know refer to this as the Julian day.  It might
not be 100% correct, but a lot of non-computer people use it.

So if my client says he wants the Julian day on a report, I will
automatically setup a 3 digit field, not a 7 or 8 digit field.



At 09:37 AM 1/31/2001 -0800, you wrote:
>Good question.
>I think all are correct in a loose definition.  Which is to say, a number of
>days since a given date in the past.  It's just that some folks decided to
>start their count much earlier, yet I agree, one representation of the
>julian, is the year + the 3 digit number.  But you cannot use this to add or
>subtract dates.  Therefore, they use the other representation, to find out
>the number of days since a given date, then use an algorithm to convert it
>into a date-readable format such as yyyymmdd.
>
>I look forward to what others have to say about this.
>
>Paul wrote:
>> today someone asked me what the Julian date was for today; I
>> replied 01031, with
>> tomorrow
>> being 01032, etc.   Another person said that's not what a
>> Julian date is at all,
>> and upon looking it
>> up on a web site - one site says it is 2451941 another site
>> defined Julian date
>> as the number
>> of days since Jan 1, 4713 BC and that today is 1757966.
>>
>> Of course neither of these two answers fit my definition of a
>> Julian date.
>>
>> So why does my COBOL manual state that a function can convert
>>  to Julian date
>> form
>> (YYYYDDD) and shows that today would be 2001031 (newer
>> manual, now uses the 4
>> digit year,
>> whereas I in my olden days ways, still think of a Julian date
>> as YYDDDD).
>>
>> Has my COBOL manual been lying to me all these years??
>>
>
Ric
[log in to unmask]