Denial (millennium problem) LO10872

Sherri Malouf (sherri@maloufinc.com)
Mon, 4 Nov 1996 13:56:14 -0500 (EST)

[Linked arbitrarily to LO10863 by your host...]

John wrote:

> - how could the analysts/programmers be so dumb

Ben wrote:

>Sherri, I've written a lot of code in my day and I can tell you from sad
>experience you never think of every possible scenario; software is never
>finished, it is just evolving. For example, on one project I was near
>completion. I had about 50,000 lines of code written; a minor change in
>the next version of the operating system I was working with caused me
>significant grief (I just hadn't expected things to change the way they
>did); and so I spent another four months adding another 20,000 lines of
>code to make the adjustment. And when the program was ready to release, I
>said to myself: "I'm not scared by what I know about this program; I'm
>scared by what I don't know about it." To me, it was a living thing. . .it
>had a life all its own. I wrote it, but I wasn't sure "I knew it."

First of all I want to do a blanket apology to ALL programmers who felt
like I was accusing them of being dumb. I guess I spoke out before I
understood enough about the issue. I visited two web sites and learned a
lot more. I read an interesting article about the whole silver bullet
controversy which was great. So now I understand the whole issue with a
lot more depth.

So here's my question again.. How do we hold the little pieces (like
making assumptions about how long programming for something as routine as
a date will last) and keep thinking about the bigger picture? This I
believe is what we need to learn. I realize as Ben states -- there are
some things we will never know until it bites us in the butt. OTOH, how
do we choose the assumptions that NEED to be challenged?

Sherri
sherri@maloufinc.com Tel:603-672-0355
LMA, Inc Fax:603-673-7120

-- 

sherri@maloufinc.com (Sherri Malouf)

Learning-org -- An Internet Dialog on Learning Organizations For info: <rkarash@karash.com> -or- <http://world.std.com/~lo/>