Re: Re: Long Constant Names

On Wed, 29 Dec 2004 14:56:23 +0100, antoine.contal@xxxxxxx
<antoine.contal@xxxxxxx> wrote:
> 
> 
> Alan,
> 
> When you write 'I'm not convinced about the utility of too-long names, or too 
> many of
> them anyway.', you introduce two interesting concepts.
> 
> First, what is too long for a variable?
> 
> This thread's context would suggest the 15-char limit is the boundary between 
> what's good and what's too long. Why is 15 characters the right limit? Why 
> not 14? 16? Or 18 if we want a multiple of 3?
> 
> Secondly, you imply quantity may be factor.
> 
> I hadn't thought about that, but you could be right. When people say they 
> want to prevent long names, they think about programs with many long names. 
> Then, they start talking about a hard limit that each variable has to abide 
> to. Yet, maybe
> we approach this from the wrong angle. What about a rule stating that 90% of 
> variables must be within the 15-char limit but the remaining 10% can be 
> somewhat longer? Wouldn't it be a rule closer to human common sense?
> 

something else to keep in mind is that a constant will 99% of the time
be one of a set of values.
   dayofweek_monday
  dayofweek_tuesday

  statuscode_active
  statuscode_held
  statuscode_deleted

It follows that the constant name should always indicate that
relationship. What better way is there than the first half of the name
is the name of the constant set, the second half is the name of the
constant within that set? You dont have to use the underscore if you
dont want to.

-Steve