Hello Nicholas!
30 Jan 17 08:36, you wrote to me:
NB>>> Wouldn't numeral days of the month (as you describe above) be in
NB>>> the exact same category as hours, minutes, and seconds since they
NB>>> are not an enumerated list of names?
KE>> But a 00 second, a 00 minute and a 00 hour are used during the
KE>> first second, minute or hour. Day zero is not in date/time
KE>> presentation.
NB> Odd that internet RFCs seem to account for a 24th hour, and a 60th minute
NB> and/or second though.
I know they do, the present common practice in the Internet and allow
the use of 0-59 and 1-60 etc.
I think you are documenting common practice in FTN or only Fidonet.
I just expressed my view.
Some of the confusion stems from the lack of a definition of terms.
If you enumerate the months of a year in software you will most likely
use 0-12. If you present a date in numerics 2017-01-17 or 17-01-2017 you
will use 1-12. If you count seconds you may go from 1-60 but in the
presentation of time 17:00:00 would be the same as 16:60:60. The latter
is not used. Strange enough 24:59:59 is.
What you have to separate in the documentation is the presentaion within
the software and visible presentaion. The visible world is more diverse
in its habits. One will prefer 2017-01-17 for today, another 17-01-2017
and again another uses 01-17-2017.
The internal representaion with the software/data should not provide
for these freedoms.
When going back to RFC's en internal and external presentaions,
you will probably find that there is little space to manouver in the
structures presented in e.g. time.h, but he date command can output
0-6 or 1-7 as day of the week, where 0 and 7 represent Sunday.
Your documentaion of the exchange formats must be unabiguous, if practice
is abiguous, then there should be a warning for future implementors.
You cannot change the document because the past is wrong in your/our
current view.
Programmers are stubborn people with strong opinions on how things should
be done. The documentation should keep them at bay to ensure
interconnectivity.
Kees
--- GoldED+/LNX 1.1.5
* Origin: As for me, all I know is that, I know nothing. (2:280/5003.4)
|