- From: Julian Reschke <julian.reschke@gmx.de>
- Date: Wed, 17 Dec 2014 11:34:08 +0100
- To: Willy Tarreau <w@1wt.eu>, Stephane Bortzmeyer <bortzmeyer@nic.fr>
- CC: Niels ten Oever <lists@digitaldissidents.org>, ietf-http-wg@w3.org
On 2014-12-17 09:46, Willy Tarreau wrote: > On Wed, Dec 17, 2014 at 09:15:39AM +0100, Stephane Bortzmeyer wrote: >> On Wed, Dec 17, 2014 at 08:27:53AM +0100, >> Willy Tarreau <w@1wt.eu> wrote >> a message of 34 lines which said: >> >>> Let's pick 425 and fill one hole instead of increasing fragmentation. >> >> What is the problem with "fragmentation"? We never aggregate (handle >> together) status codes and there is no "range" of codes who could be >> treated as an aggregate in >> <http://www.iana.org/assignments/http-status-codes/http-status-codes.xhtml#http-status-codes-1>. > > I'd turn the question the other way around : why pick a random code in > the middle of the 4xx range ? It's much easier for implementations to > keep clean and maintainable code when things are a bit tidy than when > it's completely random . For instance, when dealing with messages > associated to error codes, it's easier to know that you covered all > those you implemented when you see that 400-431 all have a message > associated than when you have to carefully check them all one at a > time against a list. Of course there's nothing critical, it's just > that it doesn't seem to make much sense to start with a random value > far away from the previously assigned ones. I agree that we shouldn't pick random codes, and that in general we should use the first available one. That being said, "451" was *not* randomly chosen (*), and changing the code now, while it already has support, seems to be unnecessary to me. Best regards, Julian (*) <https://en.wikipedia.org/wiki/Fahrenheit_451>
Received on Wednesday, 17 December 2014 10:34:44 UTC