What do you think of media and pundits that foster bad blood between IT generations?
Phil Murphy: There is no good that can come from pitting one group of people against another. IT can't get enough done with the folks we have pulling the oars in the same direction. Why in God's name would anybody responsible say "The other half of the people are no good, they're done."
The danger is that CIOs [chief information officer] ignore it and let it smolder, and what's already a bad situation between IT and business becomes less productive.
What kind of training would you recommend for the older IT worker?
Murphy: Everyone thinks service-oriented architecture (SOA) is the right way for the world to evolve, loosely coupled systems. SOA is going to be a big part of future application design, development and implementation.
SOA is technology agnostic. It doesn't care who does what part. What it cares about is that you wrap it in WSDL and SOAP. You could be a PC, a dumb device, a mainframe -- SOA doesn't care.
Mainframers have to learn some of the newer stuff. But the folks who only grew up on Windows have a lot to learn about production applications and mature processes. There is a lot that each side can learn from the other.
Can you give an example of what the younger generation could learn from its veteran counterparts?
Murphy: There is a prevalence of smaller mainframes moving to a Windows platform. The No. 1 complaint of people who do this is a lack of mature operational processes and tooling.
I recently interviewed a large financial firm in New England. They were moving 450 MIPS workload from the mainframe to Wintel. The company is 17 months into it and they are banging their heads against the wall. They have no decent job scheduling capability. Any time there are complex job dependencies, there are issues.
Your mainframes operations teams have a lot to teach the people going to a Wintel environment. Here's how you do mature operational processes. Here's what you need. The trick is to pair them together and get them to work together.
No one is going to write a Web app[lication] with mainframe COBOL. Conversely, Java is terrible at math. They're both tools, and they should both be used for the right thing.
We've written the 'COBOL skills shortage' story again and again. Has that been overblown?
Murphy: Never gets any truer does it? Yeah, it has been overblown. Not that it's completely false, but some of the pundits and trade rags are squeezing all of the fear, uncertainty and doubt out of it that they can. The market for COBOL people wasn't created overnight and it won't disappear overnight.
Nobody retires at 65 anymore. They can't afford it. We trained the last big wave of COBOL workers in the 1980s. It will be 2030 before it really hits.
Here's another acid test for you: go out and look at Monster.com. How many COBOL jobs are out there -- a few hundred? How many people are out there looking for COBOL jobs? Thousands. The oversupply is people, not jobs.
Let us know what you think about the story; e-mail: Matt Stansberry, Site Editor