Daniel M

Thanks. Always happy to answer questions about it. A lot of what I learned early on I learned from just picking people's brains. Always worked better than tutorials and the like.

Yeah, basic SQL. I think you'd be pretty at home with this software (or newer software otherwise a lot like it) from that.

Yeah, it's definitely been a learning experience. Not a lot of formal learning, just a lot of "learn about that part of the car when it breaks" learning. But after a long enough time that covers most of it anyway.

Yeah, it's a lot easier now I think. Shoulda done it years ago, really.

Dynamic stuff can go too far. When I started programming my data wasn't normalized at all. Then I learned about joins and grabbing things on the fly and did that for everything...until it reached a certain level of traffic and/or complexity, at which point it can be necessary to go back to caching, to repeating data, in select circumstances. Kinda comes full circle like that. But in this case it goes very fast without caching.

Assuming I'm understanding correctly, it already does that. It logs every single film on every ballot as individual rows in a database. It doesn't iterate anything, though, it just counts them dynamically as needed (ditto on point totals, which aren't stored anywhere but are simply calculated from the rankings on request).

The ballot table contains the userid, the TMDB id, the film's rank on the ballot, the total number of films on the ballot it was part of (so we can de-weight shorter ballots), and the time submitted.

Before wiping the table clean for the next countdown I copy and rename it, so it's all preserved at the same level of detail, theoretically allowing us to count all sorts of stuff across countdowns, at least from the point we started doing it this way.

As for next list, not sure. Presumably another genre list. War films was a close second last time, I think?