On: Hashmaps in Factor are faster than in Zig

[benatkin]

Of course.

I was just saying that a lot of the time hashmaps don't get very many entries.

There is serialization and deserialization where Cap'n Proto, Flatbuffers, and Protocol Buffers have a savings over JSON by not repeating key names of a list of maps.

[murkt → benatkin]

I fail to see how is this relevant to the article. HashMap/dict can as well store a mapping from an integer to an integer. Or from a token to an integer, or whatever, and it has nothing to do with serialization, JSON inefficiency or lists of maps.

[asplake → benatkin]

Use the structure that best fits the problem. The saving to which you refer applies only to serialisation, not to the use of the structure by an algorithm.