Randi, along with some of the other folks who sit near me, has been trying to get a project off the ground for several months. The proposal’s important to some of our larger customers, which makes it more directly political than some of our other projects. Most of us have had a hand in the planning, but in the end Randi will likely be the PM.
At this point, we’ve got a pretty good idea how we want things to work out, and Randi has three potential vendors lined up for the project. All three vendors have working relationships with our group, and each is certainly capable of doing the job, but their differing skill sets are likely to impact the project’s shape. Randi and her boss, Mark, have been sorting through the implications of that for the past few weeks. Within the analyst group, Randi’s been bouncing ideas off Alice, and I’ve had a chance to comment on (or snipe at) an occasional memo. She’s about ready to set the proposal to paper, and to officially invite bids.
The system, which will have a browser interface to a server app, needs to interact with our mainframe. Randi, Alice, and some of the managers met with IT staff to discuss the possibilities last week. It went OK, but everyone says it wasn’t pretty.
What Randi needed from the meeting: In discussions with the potential vendors, she’d discovered several possible approaches to the mainframe interaction issues. Were any (or all) of them feasible? Did the mainframe programmers have a preference?
Randi’s meeting included members of the business staff, the mainframe staff, and the client/server staff. The C/S analysts had some familiarity with the project; for the mainframers, this was their first real contact with the proposal. That, I’m told, didn’t go very well; it seems the mainframe group was expecting a formal project plan, while Randi’s object was to get a sense of what was possible so she could begin to develop that plan. This is not a fun starting place for a discussion. After things calmed down, Randi learned that the programmers thought all of the approaches were reasonable, and had no preference for a specific solution. Then they discussed possible timetables, some technical issues, and security concerns.
At one point in the meeting the programmers were asking questions about the delivery formats the project was likely to need, which led one of the client/server representatives to comment “The mainframe’s just a container, right? There shouldn’t be any programming issues; just give us the data.” I’m told that George, the mainframe DBA–who’s spent the last fifteen years working to separate the logic from the data in a system with a 1960s architecture and millions of lines of code–did not take this ignorant suggestion well.