by Linda Dienberg
 
When spending time with new friends at one point or another talk inevitably drifts toward your respective working lives.

My counterpart's firm was working to implement a new system. He's an optimist, and so am I, but I found myself saying, "There's one thing you always say never to … implementing a new system is never easy". He looked a little surprised. I said "It's the surprises that get you."

And we talked about where you can go wrong. The assumptions (all software does that!), theories (it should work), best guesses (we can try it) and simple misunderstandings (oh, you wanted a user interface with that?). All are opportunities for bumps in the road.

Here's another litany of prospective problems that lie in wait. These are some specifics that we didn't talk about: Planning (and budgeting) without reality, designing without proof of concept, coding without testing, testing without broad and representative data, and training without context. Issues in one or more of these areas can net results that range from pain to total failure.

OK, we may know this already. It's nothing profound. But customers expect a vendor to be honest about what is provided. This is just as a vendor expects a customer to be reasonable.

But, of all issues the most daunting to resolve can be a software roadblock. Or, it can be the simplest. When the situation "is what it is" prudence and reason are strongly prescribed. Sometimes there is no getting around the fact that there is a piece that the software simply doesn't provide. The feature could have been changed or even eliminated from a previous version. There can be many reasons why …

In these cases workarounds are sometimes necessary vs. awkward or inadvisable modification or some combination of the two. To avoid even more pain, look at the long haul. Look toward a solution that is relatively efficient (if not perfect) and least costly (if possible) as your mutual objective. Working together toward those goals will usually get the system running, even if it's not ideal.