Make default lowercase everywhere

[ad-si]

https://endtimes.dev/why-lowercase-letters-save-data/

[TheInnerLight]

The Crockford spec says "When encoding, only upper case letters are used." so I wouldn't expect this to be the default behaviour.

[ad-si]

It also says:

When decoding, upper and lower case letters are accepted,

Which I read as "pick whatever style you prefer". His is upper-case, mine is lower-case.

[ad-si]

He also says that it should "be human readable".
I think lowercase is much more readable, because characters have different heights and some even extend beyond the baseline and therefore it's much easier to keep track of where you are.

[TheInnerLight]

It also says:

When decoding, upper and lower case letters are accepted,

Which I read as "pick whatever style you prefer". His is upper-case, mine is lower-case.

The spec is very clear that you should:

• Use your preference when decoding.
• Use upper case when encoding.

When decoding, the library already calls toUpper (https://github.com/haskell-github-trust/ulid/blob/b80ae2c2f2cb7b0c9c29973c762135623df23413/src/Data/ULID/Base32.hs#L36C31-L36C38) so lower case ULIDs would be decoded correctly.

[ad-si]

You can not have a spec where any casing can be decoded, but then request that it is encoded in a certain way. Either you allow any casing or you don't. Therefore I'm considering it a personal preference of him. One I don't agree on that is.

[TheInnerLight]

You can not have a spec where any casing can be decoded, but then request that it is encoded in a certain way. Either you allow any casing or you don't. Therefore I'm considering it a personal preference of him. One I don't agree on that is.

You're absolutely welcome to disagree and have whatever personal preference you like but if you want to diverge from the spec in the implementation, I would suggest you should either:

• Convince Douglas Crockford to update the Crockford spec to prefer encoding in lower case.
• Convince the authors of the ULID spec to switch from using the Crockford spec to a modified version that prefers lower case.
• Introduce a newtype wrapper e.g. LC ULID so that you have a choice of whether to opt in to non-standard behaviour.

FWIW, this is not unusual. In something as ubiquitous as the URI syntax spec (RFC 3986), the section on % encoding says (emphasis mine):

The uppercase hexadecimal digits 'A' through 'F' are equivalent to the lowercase digits 'a' through 'f', respectively. If two URIs differ only in the case of hexadecimal digits used in percent-encoded octets, they are equivalent. For consistency, URI producers and normalizers should use uppercase hexadecimal digits for all percent- encodings.