| From: | Robert Nesius <nesius@******.COM> |
|---|---|
| Subject: | Re: Real-Life Computing ... |
| Date: | Thu, 28 May 1998 20:34:20 -0700 |
>
>><rant>
>>I whole heatedly agree. I don't know if there is some huge
>>earth-shaking concept I'm missing here but WHY IS THIS SUCH A BIG
>>PROBELM!?
>
When production systems work - systems such as transaction databases -
the trend is to let the box/software doing the work to sit in the corner
of a computing room and quietly suck bits off the network. These old
software packages have been proven to operate CORRECTLY. When faced
with the risk of rolling out a new system that may have bugs when you
all ready have a system that you KNOW doesn't screw things up, people
whose data represents other people's money tend to want to let the old
stuff sit there.
So what's the problem? The problem is this stuff is old. Even if a
corporate entity resopnsible for the code is around, the programmers
who WROTE it are long gone in most cases, and that knowledge base is
usually not documented. So, y2k (that term, btw, has been trademarked
by someone :/ ) - the problem is not hard to understand. But in some
cases, it's damn hard to fix. And the systems that need to be fixed
have exacting standards - transaction databases are one example.
In otherwords, it's not the nodes (like personal computers ) that are
at risk - it's the *infrastructure.*
-Rob
