Re: About bcp47

I think the situation is a bit more complicated that just a distinction 
between living and frozen standard. A standard such as HTTP (whatever 
version you choose) often consists of a core and some extension points. 
The core doesn't really change, but there may be change at the extension 
points where appropriate.

For example, even in the earliest versions of HTTP that used 
Accept-Language and Content-Language, it was assumed that additional 
language tags would be registered with IANA, and user agents, servers, 
and proxies would deal with these to the extent that was necessary for 
their content and their users. The updates to BCP 47 over the years are 
just a continuation of this. I'd also note that 
https://www.rfc-editor.org/rfc/rfc9110.html#section-12.5.4 
(Accept-Language) is very carefully written to allow both older and 
newer implementations.

Regards,   Martin.

On 2025-01-11 02:38, Eduardo Costa wrote:
> Hi,
> 
> I sent a message to `ietf-languages' asking if a laguage-tag should be
> accepted as part of the `Accept-Language' and `Content-Language''
> headers
> in both, HTTP1.0 and HTTP1.1.
> 
> The answer I got is I should accept both, while I'm being pointed to RFC9110.
> 
> I asked then why is it RFC9110 still applies to HTTP1.1, as if that's
> indeed the case, it's implied that HTTP1.1 is a living standard rather
> than a freezed one, where modern-day changes would still apply (I
> guess, so a long as these are backwards compatible. i.e: no change
> introduced after modifies the core protocol itself), while being
> derived to this list. Is this  the case? I ask this just out of
> curiosity.
> 
> This raises the doubt, nonetheless, on whether language-tags should be
> taken as valid in HTTP1.0 requests or not.
> 
> The original HTTP1.1 spec does specify these, so I guess those are
> valid for such version of the protocol.
> 
> Thanks,
> 
>

Received on Saturday, 11 January 2025 06:48:50 UTC