Re: [Webpush] Status Code for Negative Acknowledgements #49

[BCC'ing the HTTP WG]

Status codes must be potentially applicable to *all* resources, not use case specific (as this seems to be). See:
  http://httpwg.github.io/specs/rfc7231.html#considerations.for.new.status.codes

I'm especially concerned about suggestions in the issue to allocate "a range of status codes" for this purpose. 

I strongly suspect you don't need a status code, in that the semantics you're surfacing don't need to be available to HTTP implementations (e.g., generic servers and clients, intermediaries, caches). You have a very constrained use case and control over the payload; why can't this be a header or part of the body?

Cheers,

> On 15 Oct 2015, at 2:15 pm, Brian Raymor <Brian.Raymor@microsoft.com> wrote:
> 
> 
> For - https://github.com/webpush-wg/webpush-protocol/issues/49 - I've submitted a pull request to propose a 512 (Expired Resource) status code - https://github.com/webpush-wg/webpush-protocol/pull/50
> 
> Based on the Hypertext Transfer Protocol (HTTP) Status Code Registry -  http://www.iana.org/assignments/http-status-codes/http-status-codes.xhtml - 509 is the next unassigned 5XX status code, but there may be a conflict with unregistered use of 509:
> 
> https://en.wikipedia.org/wiki/List_of_HTTP_status_codes#5xx_Server_Error
> 
>  509 Bandwidth Limit Exceeded (Apache bw/limited extension)
>  This status code is not specified in any RFCs. Its use is unknown.
> 
> 510 and 511 are registered, skipping 509 which suggests an earlier attempt to avoid potential conflicts. 
> 
> 
> 
> 
> 
> 
> ...Brian
> 
> 
> _______________________________________________
> Webpush mailing list
> Webpush@ietf.org
> https://www.ietf.org/mailman/listinfo/webpush

--
Mark Nottingham   https://www.mnot.net/