Keeping hard copies are only a limited solution. Imagine, for example, imagine you’re in charge of the online encyclopedia Wikipedia. You decide to print out the whole thing to preserve it in just such a rare and bizarre event. But even just two years ago, printing it would have required almost 7,500 volumes and then it’s only a snapshot of Wikipedia at one moment in time. It doesn’t capture history or any subsequent edits. And the printed copy is only accessible in one location.
More on a printed wiki: Updating a Printed Wikipedia 
There are already nuclear war style procedures for the root DNS servers. There is a very select and obscure people that can control them in case of catastrophe. None of them can do anything own their own but they can collectively.
The biggest risk I see is the internet trunk lines (undersea cables). A terrorist group could cut them at any point along their multi-thousand mile mile range and cause major disruption to huge cities. There aren’t very many of them in total and maps of them are available on the web. Individual fiber optic truck lines support massive amounts of data and support parts of whole continents but they are also physical objects that aren’t very well protected even once they reach land. If you cut through a couple of them in the right places, you could cause massive supply chain disruptions.
VoIP relies on the internet: Voice over Internet Provider. Fishermen in 3rd world countries deciding which port to sail into or which fish to keep based on data they get from their phones rely on the internet. ATMs rely on the internet. Companies taking minutes rather than weeks to close each month’s accounting are using the internet to transmit the data around.
The WWW is a minute part of the internet. What you’re doing is equivalent to equating “regular airlines” and “transportation”: regular airlines are transportation, but a small part of it and mostly limited to transportation of people, the immense majority of transportation takes place by other means.
[Here’s something I searched for last night but couldn’t find](Cities and the Wealth of Nations - Jane Jacobs - Google Books nails iron mountain&f=false). It’s from Jane Jacobs’ book “Cities and the Wealth of Nations”, quoting Henry Grady, speaking in 1880 about a funeral in Pickens Co. Georgia, some years before:
Here’s a long review of the book which engages with the main idea: Rural areas are economically passive, making nothing of their own, simply exporting raw materials and importing finished goods. My point is that this is a good thing, economically, because it’s simply wasteful to have a somewhat poor factory in Havre in addition to a better factory in Pudong, Shanghai, when Havre could focus on growing wheat and cattle and Pudong could focus on making clothing and toys.
All that is gone is the www ------- I would say no lives lost or very few (maybe 100 world-wide) and society would readjust in a couple months. The biggest stress would be to libraries who would probably have to disconnect their phones to get people to stop calling them with constant questions.
Go the other extreme and blow all tech (cell phones etc) back to 1980 say and then I expect some bodies; enough to really matter. Also a lot more stress to average people and services.
Except that’s logically inconsistent. It is like saying all Fords die, but only Fords die: All of the Priuses, Chevys, Kias, GMCs, Opels, Bentleys, and everything else on the road keeps driving fine. There’s no possible way it could happen.
So if the only thing which dies is the Web, the obvious answer is “Shesha coils back and the Universe ceases to exist” because absolutely nothing rational or predictable could produce that outcome.
Read posts 1 and 3; logical or not that seems to be the opinions sought.
Then read the rest of my post.
That, or the correct answer is “Your question is meaningless.”
Not really. In those posts the OP keeps alternately using the terms www and internet as if they were synonyms for the layer of software and services that sits atop the hardware.
As best I can parse it, he posits the hardware lives on yet 100% of the software that runs on it dies. That logically includes, although he doesn’t understand this, the routing, DNS, all the protocols from HTTP down to IP itself, etc. None of that infrastructure is hardware.
Even if we somehow imagine just having HTTP uninvent itself while TCP, IP, DNS, and the rest of the alphabet soup continues along just fine that’s still an epic disaster. We could even allow POP, IMAP, & SMTP to live on so email still works and it’d still be a hefty disaster. Permitting the underlying telephony protocols to keep working would reduce the scope some as at least we’d have some real-time coordination capability.
So much data is interchanged between so many servers over so many protocols built on top of just HTTP that commerce would still crash before an alternate infrastructure could be built up. SOAP, JSON, and all the rest would simply stop. One heck of a lot of line of business apps are now delivered as websites communicating over HTTP. Likewise a lot of factory or industrial devices expose their control UI as an HTTP-based website.
The OP may well have *meant *“What would happen if every consumer-facing website and only those websites stopped serving their ordinary human-readable pages and magically couldn’t be restarted?” But that’s a far cry IMO from what he asked.
I do not understand the question though i have seen the OP try to clarify it.
What happens if we destroy the internet
EXCEPT
We dont destroy it?
If all the hardware works still and is not destroyed, you have accomplished nothing.
Networks are restored, data is reinstalled from the backups, things go back up and running.
I mean technically all one guy has to do is turn on a PC, connect it to the network, and launch a web server and poof, the internet has come back.
As far as people dying?
Hospitals have telephones and people that hopefully possess brains
I get to stay home from work in this case. I work in the maintenance dept of international food corp. I swipe my badge to clock in and they register it one nation north of here. The machines get their database instructions from a corporate data center one state south of me. The software that runs the refrigerant compressors is at the same place. All the freezers and coolers shut down in that case. The software I need to override them is there also. I don’t want to be near when the ammonia pipes start to detonate if the release vents fail.
A disaster as in “The Universe is now out to lunch, call back in a half-hour”, sure. A disaster as in “We can never replicate that wonderful software”, eh, not really. The complexity in HTTP is mostly not in HTTP, in that it’s really complexity in IP and TCP and the general practice of writing robust applications which remain robust even when multiple thousands of people are banging on them at the same time. HTTP in and of itself is mostly just a list of verbs and status codes, and most certainly does not include file formats, for example.
SOAP is nowhere near as vital as all that, JSON is a data format and so independent of all of this, and, again, the businesses would be much more disrupted by the Universe taking a left turn into Wackyland.
The OP wasn’t clueful enough to have meant anything at all. They could string words together but they were asking for the circumference of a square circle. Not a square, but a circle, only square.
You’re right the OP is in effect techno-word salad and we can each choose whatever combo of failures we want to talk about and be equally right ref the OP. You’ve probably got the best bottom line on the thread right there.
I agree that HTTP is technologically pretty obvious from where we stand today. There are plenty of experts who could code up a functional clone of HTTP from collective memory even if all the documentation, RFCs, backups, and source code magically vanished as well.
If by the terms of the magic we can’t clone HTTP from human memory and instead need to build something totally new but functionally similar from zero, then I’m pretty sure an alternative system could not be designed, standardized, coded, propagated, and all the critical application software that depends on that layer modified to use it in the 10 days before the groceries run out and anarchy sets in.