Its Friday night, and I'm waiting for Battlestar Galactica to start on SciFi (we already know I'm a geek) and trying to come up with something interesting that's not too contrived to put on this blog. After all, you people have been patiently reading through boring stories of my car burning and pictures of my kids -- and really, only my mother is interested in seeing those on the blogsite -- so I suppose I should throw something out here that has the markings of being business related.

I have a client (hard as that is to believe for some of you), and like many of my clients, they're a marketing oriented bunch. For obvious reasons I won't say who they are, but I will make clear they're a NEW client, not the one you all are already thinking of. Anyway, they had me build a "one-off" site to collect data from a targeted group of people. Now they want to use that data more fully and grow it into slightly more of a fixture in their plans. I'm in the same place I was around ten years ago with another system -- one that I'm still managing to this day. It is completely clear to me that unless we build something designed to handle a kind of data that so far doesn't exist -- then each and every thing I do for them is ultimately going to be tossed away for each next thing. Ultimately, all the power and utility of the platform I use that routinely allows me to underbid by as much as two decimal places will end up not saving anything.

Here's the dilemma -- I know that they're going to grow this into a cornerstone of what will become a fairly important "extranet". This basic "one-off" which was never meant to house any permanent data will frankly never be good in the role of keystone. At the same time, the client has an I.T. department who is working on a "Grand Project" to produce the kind of data management system needed for this kind of data and as long as that's ongoing, I'm never going to convince them that's its worth spending money on a concurrent effort even if its just designed to hold data that it will ultimately not own.

Even though they don't have a data structure designed yet for their big relational database, on the Domino side all I need to know are the kinds of "entity objects" I'm likely to be getting from them when they finally get it done. For this kind of site, that's easy. Its predictable at the macro level even if I have no idea what field names and values I'm ultimately going to be asked to bring over from the as yet unbuilt data store. BUT -- there's no way they'll pay me to build a repository they don't yet need.

If its you, do you spend about twice the effort you need to on these small efforts -- essentially doing work you're not getting paid for -- to build them with way more planning and organization than they would need if they were in fact never going to be more than short term tools? Do you build just what you need to for each task, but then never grow the system out to what it can be because ultimately there won't be any cost savings visible at the medium term?

I'm leaning toward taking the risk, spending the extra time, and over-building the initial pieces at a loss; with the goal of ending up with something compelling enough to be looking back on it 10 years from now and showing another junior admin how it all got started. Can you do that twice in a row, or am I better off just taking the quick hit little jobs and letting the rest of it die of its own foolishness?

-Andrew