Re: Pre-Denied: Status Code (#3410)

HTTP Working Group,

I would like to share some empirical data points that may help resolve the
discussion regarding the selection of a new status code.

The End of Term Archive, a public dataset capturing the US federal
government web, recently archived over 3.6 billion captures of 2.8 billion
unique URLs from over a million hosts. Of the nearly half a billion
captures that returned HTTP 4xx status codes, the distribution for the
codes under discussion was as follows:

* 419: 62 captures
* 420: 11,080 captures
* 421: 540 captures

Given these figures, 419 appears to have significantly less collisions in
practice compared to 420 or 421.

PS: This is not a true representation of the public web as the scope of
crawling was focused to .gov and .mil domains along with some other related
domains.


Regards,

--
Dr. Sawood Alam
Research Lead, Wayback Machine
Internet Archive



On Fri, Apr 3, 2026 at 6:30 PM Rory Hewitt <rory.hewitt@gmail.com> wrote:

> Using 419 sounds fair, if they don't have an RFC...
>
> On Fri, Apr 3, 2026 at 2:49 PM Mark Nottingham <mnot@mnot.net> wrote:
>
>> I'll be more specific.
>>
>> Does anyone see a reason not to use 419?
>>
>> I'd argue that if we allow a single framework to squat on a code
>> successfully, we're going to be running out of codes very very soon.
>>
>> Cheers,
>
>