Hacker News Reader

For my current project I decided to try out 120-bit/15-byte identifiers. 120 is evenly divisible by 4 (hex), 5 (base32), 6 (base64), and 8. Neither base32 nor base64 encodings need padding. You can usually leave padding out regardless, but with 120 bits it's not even a choice you have to worry about, and there are no wasted bits.

View on HN

Having it be evenly divisible by all those bases is neat!

But I wouldn't say "no wasted bits" unless 120 bits is also exactly the right length for a particular use case. For example, if my use case would be served better by 64 bits, I'd prefer losing some bits to padding than extending it to 120 bits.

For URLs you can also do something else. If you know that your primary format is UUID v4 you can encode your UUIDs into nice URLs in base62 format. They come out as 23 alphanumeric characters and you can reconstruct the 128 bits of UUID perfectly from them.

If it’s served by 64 bits it’s probably served by 60 right? 64 bits is too few to generate randomly.

Reply on news.ycombinator.com