FM weakness to calculate accurate time.

[edudna]

I give an example file below:

FM calculates two consequently years two days less...

In real time a year begin 12 a.m in 1/1/xxxx and end 12 a.m 12/31/xxxx.

No I did not post this thread two times.

Pascal

test_dates.zip

[Søren Dyhr]

If you have 5 days how many full nights between them, without the halves initiating and ending? Filemaker's dateformat knows nothing about time, it's just an integer behind the screens, while the an algorithm establishes what date it correspond to.

To establish what qualifies as a new year from a given date must this conversion be kept in mind (using your fieldnames):

Date ( Month ( DateStart ) ; Day ( DateStart ) ; Year ( DateStart )+1 )

Say you enter 29th of february 2004 will the same day the year after be 1st of march 2005 ...since there is nothing including your midnight distinctions - must even you accept that it's only a representation of an integer value, where it makes absolutely no sense to represent two days as 1,9999999999 something.

--sd

[comment]

I think the answer is much simpler. In the formula:

DateEnd - DateStart

the last day of the period is not included. If an event starts and ends on the same day, the result is zero. To include the last day, use:

DateEnd - DateStart + 1

[Søren Dyhr]

the last day of the period is not included

Is it much different than telling that each day has a single integer value attached, and that two days next to each other differs with just one:

1 next to 2 gives 1 ...if the difference is made up?

--sd

[comment]

Is it much different than telling that each day has a single integer value attached

I don't know - it seems the same thing to me, just from a different perspective. 2 - 1 = 1, but in everyday use, an event that started on Monday and ended on Tuesday is considered to have lasted 2 days (as if it ended on Wednesday 0:00).