There’s a very common saying among data modellers to describe the logic they use to layout their diagrams – “Dead crows fly east.” There’s a simple explanation on John Owen’s website.
While this appears to make sense, there is a real problem with the statement “Dead crows fly east.” It isn’t the fact that dead crows only fly if you throw them, or the fact that flying into the sunrise shortens your day, causing data models to darken prematurely. Please bear with me while I unburden myself.
To a townie (I explain this term later in the article), the phrase appears to describe the layout perfectly. As a data modeller (and therefore a professional pedant), I have to protest at the inaccuracy of the statement, and the severe limitations imposed on any entity relationship diagram (ERD) relying on crows for notation.
I don’t claim to be an expert on birds, but I can see a lot of crow-like birds from my house, and they occasionally fly east, but I can’t put them in my data models. There are just too many of them.
This minor part of data modelling terminology must have been dreamt up by a townie. “What?” I hear you say. “Preposterous proposition!” I’m semi-townie myself, which is embarrassing, but I’m used to being embarrassing, so I’ll continue.
As a data modeller, I must define my terms here; I hate to be accused of ambiguity. By ‘townie’ I don’t mean those “normally aged between 11 and 15. Listens to so-called ‘garage’ music such as Blazin’ Squad and So Solid Crew”, or those who “follow whatever trends are in fashion, for about 3 weeks, then go onto the next thing”, though of course both statements may also be true. I suppose there aren’t many teenage data modellers out there, but I could be wrong.
I don’t even mean “A permanent resident of a town, especially a resident of a college town who is academically unaffiliated with the local college or university.” I’m sure there are some gifted data modellers in the back streets of Oxford, though they don’t show themselves when I accompany my family on their shopping fix. Maybe they’re disguised as buskers, for safety.
Nope, my definition of ‘townie’ is nicely stated by the Collins Dictionary – “if someone who lives in the countryside refers to someone from a town or city as a townie, they disapprove of that person because they think they have no knowledge of the countryside or country life.”
Sorry if this a bit of a ramble, but most ramblers will know what I’m on about. Have you heard the rural saying about crows and rooks? “If you see one rook, that be a crow. If you see many crows, them be rooks.”
There we have it. You can’t have more than one crow (or perhaps two) in a data model; if there’s more than one, they surely must be rooks, or the data model breaks one of nature’s own rules of normalisation! The only exception (there had to be one, didn’t there!) is a crow family. The chicks do hang about for a while after fledging, but they by definition are still dependent on both their parents, so they can only be represented as associative entities, and have to appear in the top-left corner of the ERD.
I’m sure no books will be re-written to correct this glaring error; the metadata’s in production now. Pity.