Would it make sense to extend the ULID to include an optional checksum?

[mofosyne]

This would be useful in the case that the ULID is communicated over the phone or by handwriting... Standardising the form would make it easier to recognise when it's being used and throw it away upon insertion into a database system... as well as how to regenerate it again and verify it's integrity when transmitted manually.

Looks like Crockford's Base32 which you are using does have some provision for appending an optional check symbol... however it does not provide a way to disambiguate between a checksummed data vs a non checksummed data... but for our purpose since the ULID is a fixed length... you could assume any extra data appended is the check symbol... anyway best to clarify that in the spec as an implementation recommendation