Working Around (and around and around)

Your Humble Blogger works with data.

I mean, everybody does, if you think about it, right? Because everything is a datum, and the plural of datum is data. Or something. Anyway, what is actually true is that YHB has some skills in getting data into systems and out again. I’m not a true data jockey, but most of the work I’ve been paid for over the last twenty years or so has involved databases of some kind, and most often I have been working alongside and for people who don’t really get databases, or data at all, really, so I have picked up a lot of useful tricks.

This is unfortunate in many ways.

Oh, it’s not so terribly bad to have the ability to write a report and see whether the report is telling you what it’s telling you it’s telling you. On the other hand, my experience has led me to think almost entirely in terms of workarounds. And while workarounds are useful, and get things done, they also tend to inhibit good data management.

Here, let’s talk specifics. You know how YHB’s employer has gone with a new, brand-new, open-source, southern-hemisphere ILS, and how that has worked out to be very very interesting for all concerned? Well, for two or three months in there, we simply did not bill for missing books. And I’ll go back a bit and talk about our policy—when a book is overdue 45 days, we mark it as lost and bill the patron. If the book is returned later, we credit out a chunk of that bill. That bill is sent through the Bursar’s Office of the University, which means for a student, no course credit, transcripts, a diploma, whatever. It’s our little way of getting attention.

What that means is that once a month or so (on the Bursar’s schedule, not ours) we need a list of those items that (a) are more than 45 days overdue and (2) have not yet been billed. The second one isn’t that big a deal, as we can and do indicate in the system that the bill was sent. The first one should be that difficult either. Except that I lied just up there—what we need is a list of those items and the person who took them out. And our system insists that once a book is missing, it isn’t checked out. Which makes sense. If we know who has it, it isn’t missing, because we know who has it. If we have billed for it, then we are saying the person doesn’t have it, so it isn’t checked out to him. See? It makes perfect sense, except that it does not allow us to use the system with our policies.

We asked for a report that would give us all the information we need, and we got a report that gives us much of the information we need, along with several pieces of information we don’t need. So for the last three months, I have obtained a report that is, oh, say, 491 lines long. Then I can sort and throw out the 317 lines that I can easily eyeball as being wrong. Then sort out another chunk of twenty or so, and then another twenty or so, and I’m left with, what, a hundred and thirty or so? Now, of those, there are still a dozen or so that are not real, but I have to look at each one individually to find out which. But that’s OK, because I have to look at each one individually anyway, to get the rest of the information that wasn’t in the report. Getting the report and sorting it down to 130 (or so) is a matter of fifteen or twenty minutes. Looking up each individual entry takes longer, at a rate of five a minute or so, call it half an hour. So let’s call it an hour, altogether to turn the report into usable information.

Now, an hour a month isn’t actually that much work. It’s a perfectly workable workaround. Getting a report that works could take days. The annoying and repetitive work of getting the information is not actually more annoying than the annoying and (largely) repetitive work of report writing, and a lot less annoying than continuing to pester people for (f’r’ex) a list of the names of the fields (without which report-writing is extremely difficult). And, as I say, I’m all about the useful tricks and workarounds. All it really requires is that I understand why the report is wrong, what is wrong about it, and respond cheerfully to the three hundred false positives, knowing why they are there. This is one example of three (that I can think of off the top of my head) similar reports that don’t-quite-work unless I massage them after-market.

The problem is that it may come to pass that someone else wants information from the system. Perhaps even when I am not around. That me-not-around thing happens quite frequently, in fact. And if I were to sit down with someone and explain how I go about getting the information we need, well, that is likely to be the most annoying part of all. That, then, is my impetus to get a properly working report. Even though I would just as soon not bother.

Tolerabimus quod tolerare debemus,
-Vardibidian.