On Thu, 7 Jul 2016, Van Snyder wrote:
> Date: Thu, 7 Jul 2016 12:50:22 -0700
> From: Van Snyder <[log in to unmask]>
> Reply-To: Fortran 90 List <[log in to unmask]>
> To: [log in to unmask]
> Subject: Re: Does anybody else care about reliability?
>
> On Thu, 2016-07-07 at 16:24 +0100, Anthony Stone wrote:
>> As a theoretical chemist I am certainly concerned with units, especially
>> for energy, since many different energy units are in common use, but
>> also for length and angle. I deal with it by using a particular unit
>> system internally throughout my programs, and requiring users to specify
>> units for any input they provide and any output the program generates.
>> Conversions are carried out at the point of reading data or writing
>> output, and the output includes the name of the unit used. The external
>> units can be changed at any point in the data files, but the internal
>> units are always the same.
>>
>> This approach may not be sufficient for all applications, but I find it
>> adequate.
>
> This is exactly what we do, and it's not very difficult. It works fine
> until somebody drops a new hydrostatic equilibrium module on you, and
> doesn't bother to tell you that its internal units for, say, height, are
> kilometers, while yours are meters. The processor can't check procedure
> interfaces for this error unless there's language support for it, or you
> use derived types instead of real variables, which is hideously
> expensive both in labor and computer cycles.
>
>> Anthony Stone
>>
It can be worse. There was a time when heights above sea level were in
feet and depths below it were in fathoms.
-- John Harper, School of Mathematics and Statistics
Victoria University, PO Box 600, Wellington 6140, New Zealand
e-mail [log in to unmask] phone (+64)(4)463 5276 fax (+64)(4)463 5045
|