1 comment

[ 1.3 ms ] story [ 24.0 ms ] thread
It looks like he's drawing them from a pool of 1,000,000 integers which is fine so long as there are much fewer than 1000 items. With >1000 items you will have duplicates most of the time. See

https://en.wikipedia.org/wiki/Birthday_problem

There is something to say for generating a (React) key from the underlying data, possibly applying a hash function, for instance

sha384("geography:earth:jp:jp-17")

In the semantic web we call it Skolemization

https://en.wikipedia.org/wiki/Skolem_normal_form

but it's actually pretty simple.

As for the 300kb one might think at first that that's not very much, but it is so easy to have 30 things that are like that and it all adds up to a funky situation.