Re: Bit flipping .. will this work in VB6?
- From: "Michael C" <mike@xxxxxxxxxx>
- Date: Wed, 5 Mar 2008 11:08:18 +1100
"Larry Serflaten" <serflaten@xxxxxxxxxxxxxx> wrote in message
news:uPbSPqffIHA.5296@xxxxxxxxxxxxxxxxxxxxxxx
I've read what you said, and will refute anything to the effect that I
said
error handling should be avoided in the entire application.
Fair enough but you are saying as a general rule that your own routines
should ignore errors?
I merely said
that if you add the err.raise command, somebody (somewhere on the
callstack)
has to have that routine wrapped in an error handler. By adding the Raise
command you force someone to trap that error or let the program crash.
You need to trap the error in every event anyway because just about anything
can raise an error. Even something as simple as x + 1 can raise an error. It
would be a *very* rare event handler that didn't call some code that
couldn't raise an error.
But I think the main point of contention we have is from the statement
above. You claim bugs are inevitable, I'd say they aren't.
As I said, the only place bugs don't exist is in newsgroup fantasy land or
maybe in *very* small apps. If the big players can't eliminate all their
bugs then what hope have we got? It seams the difference between me and a
lot of people on these groups is that I can admit I'm fallible.
Way too many
people use that philosophy to support their 'ship now and fix it later'
schedule.
And I can appreciate that for most 'run-of-the-mill' coders, it would take
too
long to try to write code that won't fail unexpectedly. But it can be
done.
High quality can be achieved but there will just be a smaller number of
bugs.
If you don't attempt to reach that goal, then you'll never attain it.
I suspect you get closer to that goal by ignoring the vast majority of
errors.
That's not to say there aren't going to be problems along the way, errors
that need correction or debugging, but experience gained while striving
toward that goal will help to giude the designs and coding style needed
to avoid writing so many bugs. Working through the pitfalls and dead
ends, and what not, (replacing failures with more stable code) causes
one to gravitate toward writing the more stable code the first time
through....
There's one thing that strikes me as a bit of a contrast in what you're
saying. On one hand you're saying that you can achieve very high quality
code but on the other hand your saying your code isn't robust enough to
contain err.raise without hitting it. If you're that confident that your
code is bug free then you should be able to go in tomorrow and put err.raise
everywhere it's appropriate and release without even testing. Didn't think
so. :-)
Michael
.
- Follow-Ups:
- Re: Bit flipping .. will this work in VB6?
- From: Larry Serflaten
- Re: Bit flipping .. will this work in VB6?
- References:
- Bit flipping .. will this work in VB6?
- From: Ronald Raygun
- Re: Bit flipping .. will this work in VB6?
- From: Karl E. Peterson
- Re: Bit flipping .. will this work in VB6?
- From: Jeff Johnson
- Re: Bit flipping .. will this work in VB6?
- From: Michael C
- Re: Bit flipping .. will this work in VB6?
- From: Larry Serflaten
- Re: Bit flipping .. will this work in VB6?
- From: Michael C
- Re: Bit flipping .. will this work in VB6?
- From: Larry Serflaten
- Re: Bit flipping .. will this work in VB6?
- From: Michael C
- Re: Bit flipping .. will this work in VB6?
- From: Larry Serflaten
- Bit flipping .. will this work in VB6?
- Prev by Date: Re: Scan
- Next by Date: Re: Bit flipping .. will this work in VB6?
- Previous by thread: Re: Bit flipping .. will this work in VB6?
- Next by thread: Re: Bit flipping .. will this work in VB6?
- Index(es):
Relevant Pages
|