| Darren New 2006-12-15, 7:10 pm |
| Neil Madden wrote:
> than the error message (e.g. is "ARITH DIVZERO {divide by zero}" really
> any more use than just "divide by zero"?).
It would be if (a) one had nationalized error messages, or (b) everyone
used them. In this case, you catch an error, it's an ARITH error, so you
know when processing the file it wasn't a "could not open file" error.
> So, while
> producing useful errorCodes should be encouraged, is it fair to say that
> by and large, other methods tend to be used for distinguishing between
> error conditions?
Don't misunderstand. I wasn't saying it's not the best solution for now.
But I just cringe when someone says "Let's add some functionality, and
since the user will often want to distinguish this failure from others,
let's parse the error message, and not consider specifying an
errorCode." I'm simply encouraging all who define new error conditions
to define the errorCode that goes with them.
--
Darren New / San Diego, CA, USA (PST)
Scruffitarianism - Where T-shirt, jeans,
and a three-day beard are "Sunday Best."
|