The names are given a score in any applicable category. The standard score is 2, but it can be reduced to 1 if the name only mildly fits the category, or raised to 3 if the name strongly fits the category. For example, Reik is a family name, but it's pretty distant, and by marriage, so it gets only a 1 in the "Family" category. Whereas Robert is the name of a grandfather whom I admire greatly, so that gets a 3 in that category. These values are arbitrary, of course. Any values can be used (scale of 1-10, if you like) provided all categories are scored on the same scale.
The categories themselves are given weights relative to each other, both positive and negative. In this first pass, I've decided that having a pretentious sounding name is only half as bad as having a name that's trendy, which is only half as bad as a name with double "er"s. On the other hand, having a family reference is twice as good as having the coolness of a name with rock star potential.
This is just a first pass at my own weighting and scoring... these probably aren't my final weights. And it doesn't take any of Mary's preferences into account. We'll have to give her a sheet of her own, or agree to the weights and scores by agreement and compromise.
It's fun to build a model like this, but I can almost guarantee that the name we end up using won't be the winner according to any quantifiable scoring system. Not unless we tweak the weights and scores to make it come out the way we like! I've built enough numerical models to know that it almost never works out that way... even though we do have both negative and positive "subjectivity" categories to use as catch-all for areas that aren't otherwise scored.
No comments:
Post a Comment