Fujii Masao <[email protected]> writes:
> On Wed, Jan 21, 2026 at 5:18 PM Steven Niu <[email protected]> wrote:
>> I've noticed an inconsistent naming convention in the patch:
>> we use a capitalized Boolean for the boolean type, but a lowercase integer
>> when referring to the value type.
>> Is there any reason why we follow this mixed style?
> I don't have a strong opinion on this. "Boolean value" has been used in
> command
> error messages from before, and I couldn't find any messages that use the term
> "Integer". So I committed the patch with the original wording.
For context, the reason that "Boolean" is often capitalized is
that the term refers to an identifiable person, George Boole [1].
You don't see "Newtonian" or "Einsteinian" without caps either.
So this message is not wrong in isolation. But I don't think that
we have a consistent project style about whether to capitalize
"Boolean" or not. It's not a correct spelling of the SQL type name,
for sure, but it's commonplace when being used as an adjective.
So does staying consistent with the lower-case type name justify
spelling "Boolean algebra" without caps? I dunno.
If anyone wants to investigate this some more and recommend that
we adopt a project style rule (and create a patch to fix the
stragglers), have at it.
regards, tom lane
[1] https://en.wikipedia.org/wiki/George_Boole
