Awards would be a good rating scale, but I have reservations.
This is why I suggested just counting the occurrences of "won" and "nominated" in the awards listing (if that's possible). It's not going to be a very "accurate" measure of anything anyway. A win is obviously more meaningful than a nomination.
There's also an issue with movie ratings...it's not reasonable to get at them.
This is disappointing. I realized weighting the movies by their rating would only be feasible if it could be accessed in the database. It would take far too long to retrieve them again from IMDb (which for some would not be original source anyway). But this is not a big deal since such weightings are going to average out unless applied in some non-linear fashion. Even then, this would just be a way of fine tuning the number of items in the filmography as a measure of importance. That is, being in a lot of bad movies should not make someone more important than another in a fewer number of better movies.
But what really concerns me is this implies it's also not possible for the script to distinguish the movies that are in the collection. This is the best measure of what is important. I doesn't matter if the movies are good or bad, if the person is showing up in a lot of them—in any capacity—I'm going to interested in them. And there's no online source that's going to replace this measure.
Anyway, it needs more thought.
It would seem so. If there really is no way to access movies in the database, how about the option of looking up the movie in a user-supplied list of "movies of interest"? The list could even include weightings, which could be ratings—or anything else. While more cumbersome, this would allow another interesting element of control in how the people ratings are determined. The list could include all movies in the database, or any subset of it. If I don't care about dead actors I could exclude old movies. If I don't care about movies I don't own, I could exclude those. The list would be created, of course, simply by running an export template. I don't know how the script would work, but I guess it would somehow create a table from the list, and getting the information from it would be very fast.
Maybe I'm onto something here... How about another user supplied list of "people of interest"? This, too, could include weightings. This might be compiled (by the user) from other sources like lists of top actors, most beautiful women, etc. Including weightings would make it easier to combine various lists (e.g., beauty is obviously more important than talent). The existence of a person on the list would then be used as another factor (which could be assigned a weight) in the determination of their rating.
So the model becomes...
In people list: weight per list (0-10) x 50 points
Named role in movie: 1 point
Named role in movie in list: weight per list (0-10) x 1
Series: 1 x # of episodes
Series in list: 3 x # of episodes
Years worked: # years
Then rating = total points/100
Again, this is just an illustration. I think it would be much better if users had complete control over the weight assigned to each factor.
And a final thought—about keeping this whole idea in perspective. For me, the primary objective is to separate the relatively small number of people I'm "interested" in from those I'm not so interested in. In the end, I'm going to do something like select the 2,000 people who rated 8 and above and try to get good photos for them all. It really doesn't matter if they scored 8 or 10. I'm just going to be happy I narrowed the scope from 25,000. And those that scored between 7 and 8 simply missed the cut—I may look at them later. And the more factors used in the scoring, the more difficult it will be to understand why a particular rating ends up being what it is.
