Somebody (I think it was ESR) called this pointless, wasteful, and stupid. It would be, were it for the benefit of the dead.
But it's not. It's very much for the benefit of the living. It's a way of remembering a man who was so very important to us, but for me, it goes beyond that.
Somewhere out there, there's a kid who's examining some net traffic. Maybe out of curiosity, maybe to debug some kind of problem. And he'll see an undocumented header, repeated across protocols.
Legends don't come from nowhere: Pratchett fans, of all people, should know this.
And a man whose name the network whispers to itself is a pretty good legend, don't you think?
As a long time Pratchett fan, I inserted this in my Web Server headers when it first became a thing after he died. I mostly forget it's there until I'm debugging something and look at a response header and go... 'oh yeah'
It's mostly pointless, but us Pratchett fans are a loyal bunch.
It's pointless until we're able to reverse entropy on a massive scale. Then everything throughout history that made it long enough to be recorded will have mattered.
"There is only one god, and its name is entropy. And there is only one thing we say entropy:
I've had the Chrome extension for this running in my browser for a while now.
The most interesting discovery I've found is that virginamerica.com is sending the "X-Clacks-Overhead: GNU Terry Pratchett" meta tag on all of their pages.
Too bad chrome's page actions are now to the right of the url/search box instead of dynamically appearing on the box. I'd like to install this, but I don't want to have the button always be there.
Eek, I misunderstood your previous comment. You can still make extensions that use the omnibox instead of the bar, but the things each can do are very different. It's possible the creator of the extension couldn't do what they needed in the omnibox, or didn't know they could do it :)
I wonder how people feel about adding other names to the Clacks header? It feels a tad lonely for his name to be the only one passed around in this manner. It seems quite like a virtual graveyard, but with only one gravestone. As much as I respect Pratchett, many others are already commemorating his name, and I would feel compelled to add another. If people put their own loved ones into the Clacks header, perhaps this idea could grow, and become a more permanent part of the internet.
What do you recommend as the format? Comma-delimited? My personal website has been serving this since he died, but I'd like to also add Iain M. Banks, as someone else here suggested.
Comma-delimited is probably the way to go. RFC 7230 §3.2.2 establishes that you may combine multiple header fields with the same name by joining the values with commas (noting that as a special case for historical reasons, it isn't safe to do this with Set-Cookie).
That is,
X-Clacks-Overhead: GNU Terry Pratchett
X-Clacks-Overhead: GNU Lain M. Banks
should be equivalent to
X-Clacks-Overhead: GNU Terry Pratchett, GNU Lain M. Banks
Yeah, the interents needs a Banks memorial meme. Maybe not a clacks overhead though, it wouldn't make terrible sense.
I'm trying to think of something fitting from the books- perhaps there's something civs leave behind when they sublime? Something reminding those crazy conversations between ships in Excession that could be added to IRC channels?
I once accidently did this on a university computer system. Something was misconfigured with sendmail routing; when I telneted to an unexpected computer in the lab and sent an email from there it started bouncing in an infinite loop between that and another department server. About two days later the system operator noticed when log files filled the drive. My email had been making trips up and down the building a few hundred times a second. I guess I missed the N tag though!
I wonder if it's possible to craft a mail to be infinite-bounced like this. There have to be servers that are appropriately (mis-)configured of course. But that would let you have a lot more than just one name, and a lot more information about each person. An epitaph of sorts.
The trick is to make it so the mail isn't replicated as it's infinite bounced, so we don't get infinite copies (don't clog the tubes!). It's something I've been struggling with in a piece of software I've been writing, actually. My solution is to simply not allow for loops: There is only one canonical path between any two server (netsplits, yay!)
I would rather not add any additional identifying information to my browser, but I am definitely configuring Thunderbird to do this. It's harmless, quirky, geeky, and it might someday make some kindred spirit smile.
I wonder if this header shows up in the statistics of the big email providers.
A proper implementation of this would require a packet with an infinite TTL and no destination. I tried to figure out a way to do that, but did not succeed.
Because of the way email works, if you got it working, but screwed up (which is what would probably happen), you'd end up with a recreation of The Great Worm (Well, to be more accurate, you'd actually get a recreation of Bedlam-DL3: https://blogs.technet.microsoft.com/exchange/2004/04/08/me-t.... But it would have the same effect). Do you want to see what happens when millions of poorly configured email servers round the net scream out in agony?
> Some humans would do anything to see if it was possible to do it. If you put a large switch in some cave somewhere, with a sign on it saying 'End-of-the-World Switch. PLEASE DO NOT TOUCH', the paint wouldn't even have time to dry. - Terry Pratchett
What a wonderful idea. I have just updated out frontend server config to include the header.
On thing though: According to RFC6648, we should not be using the X- prefix any more for custom headers (https://tools.ietf.org/html/rfc6648), but I guess it's too late to change this now.
> 1. Deprecates the "X-" convention for newly defined parameters in
application protocols, including new parameters for established
protocols. This change applies even where the "X-" convention
was only implicit, and not explicitly provided, such as was done
for email in [RFC822].
It says for application protocols. I would argue in this case that it's not any part of an application protocol as the header is never used for anything.
RFC6648 is from 2012 though, but Sir Terry Pratchett died in 2015, which is when this initiative started. As such it's violating RFC6648 because it continues to use the X- prefix.
But it's not. It's very much for the benefit of the living. It's a way of remembering a man who was so very important to us, but for me, it goes beyond that.
Somewhere out there, there's a kid who's examining some net traffic. Maybe out of curiosity, maybe to debug some kind of problem. And he'll see an undocumented header, repeated across protocols.
Legends don't come from nowhere: Pratchett fans, of all people, should know this.
And a man whose name the network whispers to itself is a pretty good legend, don't you think?