Re: Reviving discussion on error code 451

On Wed, Dec 17, 2014 at 01:13:00AM -0800, Cory Benfield wrote:
> On 17 December 2014 at 00:54, Stephane Bortzmeyer <bortzmeyer@nic.fr> wrote:
> > On Wed, Dec 17, 2014 at 09:46:13AM +0100,
> >  Willy Tarreau <w@1wt.eu> wrote
> >  a message of 24 lines which said:
> >
> >> I'd turn the question the other way around : why pick a random code
> >
> > Is it a real question? Because 451 is not random
> >
> >     The novel Fahrenheit 451 refers to the temperature in Fahrenheit that author Ray Bradbury understood to be the autoignition point of paper.

Wow, amazingly in relation with the HTTP protocol... Had the paper
been found to ignite at 551, someone would have argued that it was
a server error in order to match that number then I guess ? I'm
sorry but now I have the feeling that the use case was completely
made up for the code instead of a code being picked for a valid
use case.

> Yes, 451 is not 'random' in the sense of 'chosen without meaning', but
> it is 'random' in the sense of 'not following the previous pattern in
> this context'. We're overloading the word random here pretty heavily.
> 
> My understanding of Willy's point (tell me if I'm off base here) is
> that the fun literature in-joke of the status code being 451 don't
> really justify choosing it.

Absolutely. And I have never heard of that literature probably like
many people. Using private jokes to build standards is not the best
way to see this WG's work look serious outside.

> For the most part, HTTP status codes are
> assigned in ascending order, and I see no reason to change that here.

Thanks, I don't feel alone now :-)

> Of course, for my part I don't care, because I treat the entire 4XX
> block the same in my code, but I suspect others don't. Those users
> would probably like us to keep some degree of rationality.

Also what if we finally need two codes ? For example, if this draft
covers a really useful case (tell the end user that the contents he
wanted to see were blocked for legal reasons), then I think it is as
much important to have two distinct codes : one for the origin server
(I'd say the authoritative server) and one for proxies. Indeed, it's
absolutely not the same to have contents blocked by the content provider
and by the local policies. And are we going to declare this useless
just because it suddenly defeats the fun game of using this specific
status code 451 ? I have a feeling that it's going to be the case...

In the end, I think that :
  - if there is any use to this status code, then the needs have to be
    identified and explained *first*, and only *then* one or multiple
    status codes are registered ;

  - if the purpose is only to register this status code, then its use
    case is made up and probably incomplete and/or poorly analyzed,
    then it does not make sense to release it.

Best regards,
Willy

Received on Wednesday, 17 December 2014 12:42:25 UTC