You did such a good job playing the role of Imelda in Part 1 (nomenclature), are you ready to resume for Part 2 (classification)?
Good. Let’s get started. You’ve more or less solved the problem of talking about your 10,000 pairs of shoes (admirers send new ones every day), but the big headache now is how to find them when you want them.
Then, for your birthday, your husband, Ferdie, adds a new room to the palace with 256 big bins, or “cubbyholes,” as your servants prefer to call them. So you and they get right to work. You soon realize that no single attribute sorts the shoes into 256 neat categories. Color, for example: the “black” cubbyhole would be overflowing onto the floor while the “puce with lime heels” would have only one pair in it. After nine tries, you finally get a scheme that works well enough, so you have it written down as Imelda’s Classification of Footwear (ICF-9). Copies are distributed to the full-time shoe servants—“cubbyholers” they call themselves—and pretty soon they are forming groups like the Cubbyholers of Imelda Meeting Annually (CHIMA) to promote “correct cubbyholing.”
ICF-9 starts by determining what kind of shoe it is (pump, high-heel, sneaker, slipper, etc.). Then, depending on the kind of shoe, it further classifies by some other attribute—color for pump, height for high-heel, intended use for sneaker, material for slipper, and so on. Of course, there are always leftovers at the bottom of the hierarchy requiring the designation of some Not Elsewhere Classified (NEC) bins—informally known as “garbage cubbyholes”—as well as some pairs (your bronzed baby shoes, for example) that refuse to be nicely classified anywhere. Furthermore, arguments arise (exactly what percentage of nylon in a silk/nylon blend disqualifies the shoes from going into the “silk” cubbyhole) that you have to resolve. The decisions are codified in the official Cubbyholing Guidelines or, less formally, in a quarterly publication of cubbyholing advice called Cubbyhole Clinic.
What happened to SNOFOO in all this? You got so caught up in ICF-9 that you forgot all about it. The SNOFOO folks dutifully add the bin number to their system—after all, it is another way of talking about a pair of shoes. Hence CH127, “Hiking boots, brown or tan, hooks above eyelets, laces unspecified” becomes FID 100552, and so on. You have the nagging feeling that ICF-9 ought to be specified in the precise terms of SNOFOO’s FIDs, but the two systems have such different objectives that even thinking about them gives you a headache, so you keep putting it off. SNOFOO tries to accommodate all the ways of talking about shoes, so it is pretty expansive. ICF-9 has to land every shoe in one and only one cubbyhole, so it is pretty restrictive.
Meanwhile, the shoes keep pouring in. The floor of the Cubbyhole Room is sagging perilously. Finally, Ferdie builds you a big new room with 2048 cubbyholes.
What an opportunity! Now you can finally fix the things about ICF-9 that are out of sync with the times. Flip-flops, for example. When you designed the ICF-9 and were limited to 256 cubbyholes, only children and gurus wore them, so they only got one cubbyhole. Now the bin is overflowing. So you design a modern classification system to update the classification based on current shoe terminology and the latest understanding and practice of shoe wearing. You call it ICF-10.
Your cubbyholers are a little intimidated by the new system at first, but they are servants after all, and they will cope or find a new job. It’s not as if they have a choice. They may have to look at a shoe more carefully—flip-flops, for example, now go into eight different cubbyholes depending on sole material and thong material—but the process of cubbyholing a shoe in the new classification system isn’t any different, so they are confident they can do it. No, it is the shoe manufacturers who are giving you grief. It turns out that they use a shoe’s cubbyhole number to price their shoes and track sales.
They whine: “Our databases only have three characters for the cubbyhole number.”
They complain: “Our store shelving policies are all in terms of ICF-9. Who’s going to foot the bill to rewrite them?”
They lobby to stay with ICF-9: “More cubbyholes isn’t going to make shoes any better.”
They try to make fun of your work: “Can you believe ICF-10 has different cubbyholes for fuzzy slippers with animal faces on them versus those with aliens!”
You give them an extra year to get ready, but in the end, it just means another year of whining and complaining.
You decide to try to help. You ask yourself: suppose a pair of shoes has just been stored in an ICF-10 cubbyhole, and that’s all I know about it. I can’t see the shoes themselves or find out anything else about them. What ICF-9 cubbyhole would they have been assigned to if we were still using ICF-9? This could be helpful to people who were keeping track of things by ICF-9 cubbyhole and want to be able to make historical comparisons after the shoes have all been reclassified. So you go through ICF-10 and write down all those correspondences. You call them the ICF-10 to ICF-9 Cubbyhole Equivalence Map – 10-to-9 CHEM for short.
To be thorough, you do the same mental calculation from the other direction. What if all I know is a shoe’s ICF-9 cubbyhole? Where would it go in ICF-10? You write down those correspondences as the 9-to-10 CHEM.
Building the ICF-10 classification was a lot of work. Constructing the CHEMs was a lot of work. But now everyone should have what they need for an orderly transition. Right?
We’ll see about that in part 3.
Ron Mills is a Software Architect for the Clinical & Economic Research department of 3M Health Information Systems.