I was about to complain the misuse of headers and the bad choice of not using an X-* header... then I discovered RFC 6648 [0].

Wow.

[0]: https://www.rfc-editor.org/rfc/rfc6648

When I come to HN, this is what I'm always hoping to see. Lots are here for startup news, FAANG, etc. Nope, I just want to see really, really cool fun things.

Great work.

You should probably spend more time on the Show HN section then:

https://news.ycombinator.com/show

Agreed. It really takes me back to the old days of the Internet, with fun little easter eggs like this!

Thank you for the kind words! I feel the same.

This is really cool!

For posterity, below is the output when the command is run:

  $ curl --head https://ja.cob.land/http2-204
  HTTP/2 204 
  date: one of those frigid Saturdays in November
  age: just turned 27
  location: Messina's Trailer Park on Southside Drive
  trailer: 1967 Elcona single-wide (not as bad as it sounds)
  server: at the Neptune Diner until I can get published
  content-disposition: in the late autumn of my discontent
  accept-patch: if it contains nicotine
  tk: oh shit I should put something here
  expires: in my sleep I hope

One project I have had in mind for many years is a direct two-way, user-to-user communication protocol or user convention that only uses HTTP request headers. One benefit would be easier-to-write clients and logging servers. There would be no need for the complexity of, and hence third party control over, a "web browser", nor the need to learn skills (HTML, CSS, Javascript, etc.) to develop and maintain a website. Users could communicate and transfer small data directly via sending and logging HTTP requests.

If one is a avid reader of HTTP server logs, one might conclude there is a susbtantial amount of text being transfered over the internet via HTTP headers. However, almost all of it is intended for use by persons other than users. To me, as a user, this seems like a waste.

Others have demonstrated how any bits can be transferred via a variety of methods, e.g., https://web.archive.org/web/20001207014600if_/http//decss.zo... That included NNTP or e-mail headers, but did not mention HTTP headers.

Poetry via HTTP response headers. It's a start!

Fix missing colon in URL:

https://web.archive.org/web/20001207014600if_/http://decss.z...

This is awesome!

On a similar note, I added a unique poem to each of our error messages a while ago:

https://readme-error-codes.herokuapp.com/

This is so fun! Thank you for sharing it.

My older, but somewhat similar:

> docker pull elek/herbstag && docker history elek/herbstag

But it's just publishing an old (but good) poem, not as cool as writing a new one ;-)

This is lovely! A long time ago in a galaxy far, far away, much interesting content on the web was only found in source, but this is another level of elegance.

Normally one will get the default headers, plus any other added. In this case, the default headers are all gone. How did you do it?

I guess it depends on what you consider a default header. (I’m not too familiar with standards here; there may be some IETF definition you’re referring to?)

It turns out if you have enough control over your server you can make it do a lot of things! (Which may or may not conform to a spec.)

Right, but how did you do it? What I was asking for was more details. Now, ever since I wrote my comment, I had the time for a little research and:

$/> curl --head https://ferengi.one

Very cool! Not sure if that's your site but if so, nice work. :)

Very nice! By the way, I love that you explain everything, including how console commands can be run! :D

Thank you! My hope was to make it accessible to non-technical readers while still being fun for people who are familiar with HTTP innards.

Happy National Poetry Month (USA), HN!

Love it! Very clever.

This is prose!