Internet, telegraph ... the main thing is the tail!
I read about the history of the late telegraph and the early Internet. In general, a typical picture of evolution emerges - nothing appeared at once, everything evolved gradually, there were many dead-end branches, some branches interbred.
To begin with, the late telegraph was generally suspiciously close to the Internet.
One of the fathers of ARPANET , the forerunner of the Internet, Leonard Kleinrock , wrote his dissertation on the telegraph system Plan 55 . Here's what the telegram looked like:
What is not an IP packet? Title, addresses - everything is in place. The principal topological difference was the absence of the end-to-end principle (that is why the sequence numbers fit each station for each next station). Intermediate stations were responsible for the reliability of the transmission - five nines and all that - which made the system more expensive, required a lot of centralized planning and put a limit on scaling. On the Internet, however, healthy pofigism reigns: “Lost a package? This is a sender care! Will send again! »Therefore in TCP sequence numbers are put down by the initial sender.
Plan 55 switching centers resemble modern Ethernet switches, only on punched tapes. The telegram came over the outer wire, was preserved on punched tape. Then the automatic switch read the tape and sent a telegram to the appropriate output device, also via wire, where it was also stored on punched tape. And from the output tape it already went on a long wire, so to speak to the world. If something broke somewhere, the tape was rewound and resold.
There were later telegraph forms of the military, which generally resembled e-mail strongly - with monitors, keyboards, and so forth.
In the late 60s ARPANET was launched, but the end-to-end principle was also absent there. The first link ARPANET connected the labs of Kleinrock and Engelbart . The theory of packet routing by the time was written by theorist Paul Baran, originally from Grodno, by the way.
In the early 70s the group, incl. including Vinta Cerfa, worked on the bugs and, along the way, took a couple of ideas from there and from here, including from the French system CYCLADES, they have already designed the Internet itself, based on the TCP / IP protocols, which by the early 80s had become official and popular.
With the growth of the load, further improvements were made more than once by a file, incl. in the late 80s, Van Jacobson distinguished himself, who is now roiling for Content-Centric Networking. This is in the sense that the Internet was originally intended for remote access to large computers, and today it is used to distribute content, which disparity causes a lot of problems. I totally agree.
And in the early 1990s, Tim Berners Lee blinded quickly, otherwise you can’t tell the World Wide Web, which dark people call the Internet. The project simply crossed the already well-known and proven hypertext technology with the already well-known and proven Internet technology, in the cheapest and simplest way. Coding took a month. Note that he defeated the rival hypertext protocol Gopher, which was not only less pofigistic, but also burdened with some kind of non-free license. Ecological niche for two is not enough.
I note that the changes at each stage were incremental. Attempts to jump over two steps ended either in a crash, or by putting the results on the shelf for ten to twenty to thirty years. The most farsighted of these persons was, perhaps, Engelbart, who invented everything at all, up to the video chat, but did not sell anything.
')
Morality? Progress is a continuation of evolution by other means.
To begin with, the late telegraph was generally suspiciously close to the Internet.
One of the fathers of ARPANET , the forerunner of the Internet, Leonard Kleinrock , wrote his dissertation on the telegraph system Plan 55 . Here's what the telegram looked like:
What is not an IP packet? Title, addresses - everything is in place. The principal topological difference was the absence of the end-to-end principle (that is why the sequence numbers fit each station for each next station). Intermediate stations were responsible for the reliability of the transmission - five nines and all that - which made the system more expensive, required a lot of centralized planning and put a limit on scaling. On the Internet, however, healthy pofigism reigns: “Lost a package? This is a sender care! Will send again! »Therefore in TCP sequence numbers are put down by the initial sender.
Plan 55 switching centers resemble modern Ethernet switches, only on punched tapes. The telegram came over the outer wire, was preserved on punched tape. Then the automatic switch read the tape and sent a telegram to the appropriate output device, also via wire, where it was also stored on punched tape. And from the output tape it already went on a long wire, so to speak to the world. If something broke somewhere, the tape was rewound and resold.
There were later telegraph forms of the military, which generally resembled e-mail strongly - with monitors, keyboards, and so forth.
In the late 60s ARPANET was launched, but the end-to-end principle was also absent there. The first link ARPANET connected the labs of Kleinrock and Engelbart . The theory of packet routing by the time was written by theorist Paul Baran, originally from Grodno, by the way.
In the early 70s the group, incl. including Vinta Cerfa, worked on the bugs and, along the way, took a couple of ideas from there and from here, including from the French system CYCLADES, they have already designed the Internet itself, based on the TCP / IP protocols, which by the early 80s had become official and popular.
With the growth of the load, further improvements were made more than once by a file, incl. in the late 80s, Van Jacobson distinguished himself, who is now roiling for Content-Centric Networking. This is in the sense that the Internet was originally intended for remote access to large computers, and today it is used to distribute content, which disparity causes a lot of problems. I totally agree.
And in the early 1990s, Tim Berners Lee blinded quickly, otherwise you can’t tell the World Wide Web, which dark people call the Internet. The project simply crossed the already well-known and proven hypertext technology with the already well-known and proven Internet technology, in the cheapest and simplest way. Coding took a month. Note that he defeated the rival hypertext protocol Gopher, which was not only less pofigistic, but also burdened with some kind of non-free license. Ecological niche for two is not enough.
I note that the changes at each stage were incremental. Attempts to jump over two steps ended either in a crash, or by putting the results on the shelf for ten to twenty to thirty years. The most farsighted of these persons was, perhaps, Engelbart, who invented everything at all, up to the video chat, but did not sell anything.
')
Morality? Progress is a continuation of evolution by other means.
Source: https://habr.com/ru/post/74156/
All Articles