Re: when to use a function or file for information

["Lennon_s_j@xxxxxxxxxxx" <lennon_s_j@xxxxxxxxxxx>]

Yes, it might be hard coding, but it is hard coded in only one place, 
thus easy to change.  And down the road the routine could be changed to 
a file lookup if needed.

As is often said of the best approach, "it depends".  And if the powers 
that be don't commonly do much code reuse, you will probably have a hard 
time convincing them to allow you to "hard code", even if it is only in 
the one place.

But IMHO, a single callable routine is the way to go.  A routine that 
does one thing, and does it very well.  Probably in a service program.

Sam

On 3/24/2010 3:09 AM, David FOXWELL wrote:
> > -----Message d'origine-----
> > [mailto:midrange-l-bounces@xxxxxxxxxxxx] De la part de Loyd Goodbar
> >
> > Like Jonathan said, if you can infer the flag, just put that
> > in a subprocedure and be done with it.
>
> So, if the flag were dependent upon say, the client type, I could have :
>
> clientFlag = getClientFlag ( clientType )
>
> and in getClientFlag :
>
>   If clientType = TYPE_THIS....etc
>   If clientType = TYPE_THAT....etc
>
> But this is seen as 'hard-coding' the client type and seems to be discouraged here. Hence the preference for adding the flag to the client file.