Re: Casino switches to AIX

Steve,

You don't know what you're talking about regarding unstable legacy code.
You may have lousy code, but you can have old lousy code and new lousy
code, it's just lousy code.

We're running a 24/7 operation, taking bill payments on our system. We
often run 3-4 months between IPLs, and ran a year between IPLs on our
previous system. We do backups, PTFs and IPLs ON OUR SCHEDULE. All of
this is old RPG code, doing the transaction processing with a combination
of dial up and IP connected payment stations, along with our Web site,
hosted on two IBM i partitions.


Steven Morrison
Fidelity Express
903-885-1283 ext. 292




"Steve Richter" <stephenrichter@xxxxxxxxx>
Sent by: midrange-l-bounces@xxxxxxxxxxxx
11/30/2008 10:39 AM
Please respond to
Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>


To
"Midrange Systems Technical Discussion" <midrange-l@xxxxxxxxxxxx>
cc

Subject
Re: Casino switches to AIX






On Sun, Nov 30, 2008 at 10:28 AM, Trevor Perry <trevor@xxxxxxxxxx> wrote:

Steve,

It seems that it is not the i that is causing the problem, but....

1. Poor change management practices

I am making a lot of code changes to a legacy system,
writing and deploying new apps.

2. Poor application development practices

There is a constant battle with system locks.

3. Poor system management practices

The system is way behind on PTFs because we cant afford
the down time ( and lack of knowledge on how to minimize the time
needed in restricted state to apply the PTFs ).

Seems like the complaints you continue to make are all related to bad
practices in your IT shop. For every one of your complaints, there is a
viable and modern solution available. Maybe if you improved your IT
practices, you would stop complaining less about the i.

I would call it understanding the terrain you are working in and
thinking about how to overcome obstacles. Nothing to be afraid of and
it is a good way to learn. Regarding legacy apps, I think it is a
fact that the i is burdened by the old code that runs on it. Kind of
like it is a victim of its past successes. A way to ditch the legacy
view of the system is for IBM to not worry about breaking code in the
next release. Come out with a modern, .NET like version of RPG that
cant compile code with I or C specs, that uses global variables, or
redefines variables in data structs.