A context for the Web App effort….
Note: This entry uses “we” a bit loosely, and collapses a several-year effort into a few paragraphs. For about half of the period under discussion, I was under the thumb of a Dragon; I joined the team around the middle of the project. But the story’s worth telling from the start….
Although the nation has fifty states, each with its own body of law, there are many areas where common sense dictates that all states’ laws be similar just so folks know how to behave when they travel, and so enforcement agencies can have common methods and measures. It would not do, for instance, if traffic rules varied radically from state to state; thus standards bodies with acronyms like AAMVA and NCUTLO. The finance industry has a similar need for consistency; so do other commercial groupings and government practice areas. Typically these bodies invest effort in proposing standard legislative prose, devising what are typically called model laws; the expectation and the norm is that the models become the first draft of each state’s legislation in the organization’s area of interest. These models become the focus of lobbying efforts, and the basis of new law.
It’s pretty rare for a model law to get a complete rewrite, but tweaks are common. Nonetheless, sometimes the effects of a change can be pretty dramatic. A few years ago, IACA’s change to a model law invalidated the logic–and with that the procedures and software–in our Uniform Commercial Code (UCC) section. This change, by the way, was largely driven by the clients, not the state agencies; the clients were trying to simplify the process. We, and similar agencies in all other states, suddenly found ourselves in the market for new software.
The legal changes gave us reason and opportunity to rethink and rebuild our processes. First we implemented a jury-rig on the old system; it was clumsy and ignored the spirit of the legal changes, but adequately met the intention of the law. More or less simultaneously we began work on an Invitation to Bid for a replacement system which would meet our “corporate” needs. Much of the planning effort went into improving process flows, of course; another intention was to move much of the activity to the state website, so our customers could interact directly with the system. While adapting the old system to the new law was not considered feasible, it contained millions of data records and microfilm images, so moving the data and the images to the new system would be a major part of the undertaking and was a major cause for concern.
Naturally there were interested vendors. None of these was yet in a position to offer a finished product. Some were already supplying systems based on the old law, others were adapting systems built for other purposes, still others were planning to build from scratch. One state was attempting to form a consortium to build a shared system with its neighbors. We looked over the responses, found a pair of vendors who seemed in tune with our intentions, negotiated with both, and eventually signed a contract.
Implementation of the Production System
The vendor we hired was adapting code from a similar system, and made sales to several states; conceptually we were buying an “off-the-shelf” system with considerable customization to fit our needs. Although it wasn’t an issue to our operation, in some states the vendor’s marketing angle was that their two systems would work well together. The company–basically a small coding shop, though one with a large clientele–was sold twice while we were negotiating, but ended up with the letterhead of a Large National Corporation; the sales weren’t harmful to the purchase effort.
The system’s implementation wasn’t unusual, and went pretty cleanly. Our spec required them to stretch their programmers’ skillsets, which required some detailed planning–an interesting time for their team, and for ours. About six months after the contract was signed the vendor installed the hardware and software, converted the data, and trained our users. It took three months for the users to get comfortable with the new interface, which was a bit of a frustration but had been expected; they’d had thirteen years to get accustomed to the old system and were reluctant to abandon old habits. We had a crisis about a month into the changeover for reasons unrelated to the new system, and survived; a christening by fire, and a success.
The system is reliable, stable, and meets our needs. There are some bothersome details, but we’re quite happy with the results.
The Web Interface
As you’ve seen, implementing the web interface, which was developed by a different vendor in a parallel-but-separate project, has been a little problematical. Our original plan was to bring up the web operation a few weeks after the internal interface–a plan the testing coordinator (that would be me) had the opportunity and responsibility to veto. Things just weren’t Ready.
We finally got to Ready about eighteen months after our original target, and are now cleaning up the loose ends. Had we gotten the online system running fifteen–or even twelve–months ago, the delays would have been acceptable, and the still-outstanding problems would have long since been fixed. As things worked out, Go Live was a mundane event, with a bittersweet aftertaste, so the continuing problems seem like more of the same, rather than finishing touches. Projects are supposed to end; this one staggers on.
A few weeks ago Jack, who authorizes this project’s checks, attended IACA’s annual meeting. He mainly spent the week assessing other states’ experiences their new systems. He learned that folks who’d worked with our vendor had results all over the map, and two states have moved to other systems after implementing this one. He also learned that those who joined the consortium, mentioned above, had the best experiences. He’s become concerned about the longevity of the vendor’s project, which seems a legitimate issue.
He’s also having second thoughts about the purchase process. Those doubts are illegitimate on two fronts:
- With the same information, we’d repeat the purchase. At the time of the purchase, no vendor had a fully-functional product and there was no way to predict the best choice.
- The system we purchased works well–except for the web interface, which is finally under control, and would have had similar issues if we’d gone the consortium route. The successes and failures of other states’ projects are really beside the point.
That doesn’t mean we can’t learn lessons from the experience. But he’s asking the wrong questions, and threatening to learn the wrong lessons. Mel and I told him as much.