The problem with the scenario of "what will this look like 200 years for now" is the underlying assumption that we'll still be using the same addressing system. I see no basis for this - look back at the history of telecom, and we update addressing every few decades, or at least the UX of it.
Let us start with phone numbers as an example - at one point we gave a human operator a short codeword and number. Then we had 7 digit dialing, then 10. Now, we still have those digits, but rarely key them in directly... we use names that we enter info a contacts app. We are halfway to a future where the actual digits are masked in higher layers of abstraction. Domains are similar. We don't key in IPs, we have domains as an abstraction layers. To think that we would not have moved up one more layer of abstraction in 200 years is hard to believe.
So I'm not worried about a 200+ year lifespan of a domain. I'm more curious what comes next, as the whole concept of domain names is already starting to crack.
> Let us start with phone numbers as an example - at one point we gave a human operator a short codeword and number. Then we had 7 digit dialing,
It's an aside, but an interesting one: The "codeword" dialing was also 7 digit dialing, the word just encoded the first two numbers of the Exchange (the first three digits) because AT&T research showed that people were good at remembering roughly 5 +/- 2 things at a time, so 7 was the "upper limit" of digits and because 5 was the sweet spot, they used the "named exchanges" pattern for a bunch of years before they got too many exchanges to effectively name and they decided more people were comfortable enough with 7 digits.
The old exchange names are still a fascinating list and can still be used today to remember some phone numbers. Some were mnemonic like UNiversity was often given to the biggest University campus in a city. You'll still find many Universities still have lots of 7 digit numbers that start 86 (UN) (and/or neighbors 85/87). Most of the others were chosen somewhat like the NATO alphabet to encode two digits of numbers such as KLondike (55) or PEnnsylvania (73). Because these exchanges define/encode neighborhood boundaries (and you would associate the phone numbers of that part of town with that part of town) you can see neighborhoods and streets in cities throughout the US still named after these exchanges, even if they don't remember why AT&T's abstraction of the first two digits of a phone number left a lasting legacy in the very shapes and names of neighborhoods. Neighborhoods and streets named Klondike or Pennsylvania not for any relation to the regions those names but because AT&T thought those would be easy to remember phone numbers.
Perhaps to bring things back towards the topic: the interesting thing in the drop of "codeword" dialing was that it was a loss of abstraction. The bandaid was ripped off and people were expected to just learn 7 digits rather than "vaguely what neighborhood plus five digits". Levels of abstraction aren't always a "ratchet" from lower to higher. Sometimes dropping a layer happens or is important. Per Postel's law it's also a fun reminder that these things will stick around as long as they need to and compatibility will generally be maintained in some way or another. I don't know what the replacement to DNS will be, but I can imagine it will still reflect some things about DNS, even 200 years from now. You can still call Pennsylvania 6-5000 or Klondike 29401. If it is still a landline you might even still have a rough idea where it will dial on a map. In the New York City area code the one immortalized in swing music still even calls the Hotel Pennsylvania all these decades later.
The problem with the scenario of "what will this look like 200 years for now" is the underlying assumption that we'll still be using the same addressing system. I see no basis for this - look back at the history of telecom, and we update addressing every few decades, or at least the UX of it.
Let us start with phone numbers as an example - at one point we gave a human operator a short codeword and number. Then we had 7 digit dialing, then 10. Now, we still have those digits, but rarely key them in directly... we use names that we enter info a contacts app. We are halfway to a future where the actual digits are masked in higher layers of abstraction. Domains are similar. We don't key in IPs, we have domains as an abstraction layers. To think that we would not have moved up one more layer of abstraction in 200 years is hard to believe.
So I'm not worried about a 200+ year lifespan of a domain. I'm more curious what comes next, as the whole concept of domain names is already starting to crack.
> Let us start with phone numbers as an example - at one point we gave a human operator a short codeword and number. Then we had 7 digit dialing,
It's an aside, but an interesting one: The "codeword" dialing was also 7 digit dialing, the word just encoded the first two numbers of the Exchange (the first three digits) because AT&T research showed that people were good at remembering roughly 5 +/- 2 things at a time, so 7 was the "upper limit" of digits and because 5 was the sweet spot, they used the "named exchanges" pattern for a bunch of years before they got too many exchanges to effectively name and they decided more people were comfortable enough with 7 digits.
The old exchange names are still a fascinating list and can still be used today to remember some phone numbers. Some were mnemonic like UNiversity was often given to the biggest University campus in a city. You'll still find many Universities still have lots of 7 digit numbers that start 86 (UN) (and/or neighbors 85/87). Most of the others were chosen somewhat like the NATO alphabet to encode two digits of numbers such as KLondike (55) or PEnnsylvania (73). Because these exchanges define/encode neighborhood boundaries (and you would associate the phone numbers of that part of town with that part of town) you can see neighborhoods and streets in cities throughout the US still named after these exchanges, even if they don't remember why AT&T's abstraction of the first two digits of a phone number left a lasting legacy in the very shapes and names of neighborhoods. Neighborhoods and streets named Klondike or Pennsylvania not for any relation to the regions those names but because AT&T thought those would be easy to remember phone numbers.
Perhaps to bring things back towards the topic: the interesting thing in the drop of "codeword" dialing was that it was a loss of abstraction. The bandaid was ripped off and people were expected to just learn 7 digits rather than "vaguely what neighborhood plus five digits". Levels of abstraction aren't always a "ratchet" from lower to higher. Sometimes dropping a layer happens or is important. Per Postel's law it's also a fun reminder that these things will stick around as long as they need to and compatibility will generally be maintained in some way or another. I don't know what the replacement to DNS will be, but I can imagine it will still reflect some things about DNS, even 200 years from now. You can still call Pennsylvania 6-5000 or Klondike 29401. If it is still a landline you might even still have a rough idea where it will dial on a map. In the New York City area code the one immortalized in swing music still even calls the Hotel Pennsylvania all these decades later.