|
PureBytes Links
Trading Reference Links
|
At 3:51 PM -0400 8/17/98, Orphelin@xxxxxxx wrote:
<snip>
>It's a bug.
>A bug always comes from a lack of testing.
>One bug in such a piece of software, it's really amazing for a company
>that you report to be lacking in debugging.
Pierre:
I don't know why you persist in always defending everything that Omega
does. It just undermines your credibility and serves to fan the fire,
keeping the useless discussion going on and on - "sound and fury,
signifying nothing."
You seem to think that everyone enjoys bashing Omega. You are wrong. We
would like nothing better than to be happy customers. We are just trying to
get them to change and are frustrated when we can't. Using this case as an
example:
The TS Express Newsletter of Jul/Aug 1996, page 10, talks about
"Multidimensional Arrays". "Multi" means more than two. I dug out that
issue before I tried to use this feature to see what it said. I found the
article useful so I decided to try it. And, son-of-a-gun, the program
compiled perfectly! No error messages! Wow, this is great! As Gary Fritz
pointed out, there are other simple ways to do the same thing so I didn't
NEED this feature, but since it was there, why not use it.
Then, after hours of searching out the screwy behavior of my program, I
discovered that it doesn't work. Two years after it was favorably described
in a respected newsletter that has YOU as the Technical Editor, I find that
the feature doesn't work. Seemingly impossible!
I put together a simple test case to prove it and sent it to Omega's
EasyLanguage Support. Three sets of messages later, they agreed; it doesn't
work.
We are not complaining about a bug. As you say, all software has bugs.
We are complaining about:
> The development process that fails to test even the most basic
operation of an important new feature, described in a popular
newsletter.
> The process that fails to tell customers about it,
> The process that requires three messages before they understand
that there is a problem,
> The process that loses the previous bug report from Sam Tennis,
> The process that fails to disable the compiler feature that allows
the user code to compile without an error message,
The fact that these processes are screwed up causes lots of people,
including me, to question whether we should be trusting our trading capital
to a program with so many "latent defects", built by a company with
apparently no regard for establishing integrity in their internal
operations.
Good software development processes are not all that hard to put in place
in this day and age. They are now well understood and there are many
suppliers of tools to help. I can only conclude that they don't think it is
important.
They are wrong.
We, their customers, know they are wrong and only wish they would wake up
to the fact.
If you would join us in this effort, I think we would all have a lot better
chance of succeeding.
Bob Fulks
|