Re: Software Vendors -- MIDRANGE-L

Why would I want to change my code, when not a single user
reported any problems whatsoever with it?  For six months this
program was distributed to company after company with user
after user using it, and not a single bug report.

A bug report finally did come in, though, and I told them that
was impossible, it was bug free.  They showed me the bug on
the program on our system and it was a bug.  That's when I
went into the code and saw the bug my boss put there inten-
tionally (he admitted to it later).  I fixed it back the way it
was before hand and it was once again bug free.

So what is bug free?  When a program never crashes, when
it never displays data that is incorrect based on the information
that has been entered, what it never provides incorrect results
based on information provided.

The fact that the program with the scroll bars not working with
*ROLLKEY set, is that a bug?  Yes.  It does not work as
advertised.  Now, the fact that his program won't run on a
Linux system, is that a bug?  No.  If a new version of the
operating system comes out and his program doesn't work
anymore, is that a bug?  No, that would be a bug with the
new version of the operating system.

A saying I often use is, "it works as advertised".  If a program
is designed to run on an AS/400 V4R4M0 with a certain PTF
level on a 5250 screen and it does, but gets "broken" on a
different operating level or PTF level is that a bug?  No, that is
how it was advertised.

Saying such, why is it not possible to test every conceivable
condition?  You also need to remember, if your program is
coded right and looks at incoming data with suspicion, 90%
of your possible bugs can be trapped right there.  Yes, I know
that it is not really conceivable to do this, but any data entered
from a user should be checked before it propagates into your
data base.

Now, there is a certain case in program on our system (not
ones I write) that if there is a blank record in the data base
all kinds of horrible things happen (took me a while to track
down that bug).  Now, the fact that horrible things happen
with a blank record in the database, is that a bug?  Yes.  But
where is the bug?  It would be the program that allowed the
blank record to be entered in the first place.

Someone goes into DFU and enters a blank record into the
database, and things crash, is that a bug?  Hard to say. I would
have to say there was a bug, but it was with the user who did
that in the first place, not with the program.

Remember, I never said all my programs are 100% bug free.
This one in particular was, and yes, I was and am very proud
of that.  I do strive for 0 errors, but it can be impossible to
achieve in some situations, but it should still be tried.

Regards,

Jim Langston

John Earl wrote:

> Jim,
>
> <SNIP>
> I mean no disrespect, and in fact your program could
> possibly have been "100% bug free", but you could never
> prove it.

<SNIP>

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---