Why was the Web a success? Rough notes by Daphne Preston-Kendal , 2013-06-10 What is the "Web"? Some people are very pedantic about the difference between "the Web" and "the Internet" -- what *specific* technologies constitute the difference? Answer: What TimBL invented in his _annus mirabilis_ of 1990: the URI, HTTP, and HTML. So, which of those are important to the early success of the web? HTTP? No, not really. The web was initially based on FTP, and HTTP was just added to make it faster. HTML? Maybe, but really it was just a document format -- historically, a subset of CERN SGML (http://cds.cern.ch/record/997909/files/) with the addition of the tag for making links. In fact there were hypertext document formats before HTML, many of which were also based on SGML. HyTime, for instance, was an ISO standard for hypertext systems, and its documents were based on SGML. So, it must be the URI. Yes: The URI, initially "UDI" (Universal *Document* Identifier, not "Resource" identifier), or more specifically the URL, is a brief, declarative, machine-followable, secure instruction, telling you where in the vast existing maze of the Internet one might find a file to download, and more specifically *how* to find it. Before the URI: Connect to the FTP server at ftp.example.org and in the public/software/bin folder you'll find xyzzy.tar Instructions must be followed manually. There's no cross-platform way to just make a button or some simple command that you can use to download it automatically. After: Download ftp://ftp.example.org/public/software/bin/xyzzy.tar You can just feed that URL to a program and it will figure the rest out. Best of all, when combined with HTML (remember, the most important feature of HTML was the tag) you can just make a clickable thing that will do that for you. It's brief, so it's copy-and-pastable; it's declarative, so you can use it as an identifier as well as a locator; a machine can follow it automatically; and it's secure: unlike if you told someone to type a shell command into their machine to download the file, it couldn't run a malicious command like `rm -rf ~`, so it's safe to build a system that follows URLs from arbitrary sources. Best of all, it integrated with things already on the Internet, and wasn't just dependent on HTTP or FTP. For instance, the old WAIS system for distributing text documents could be accessed through URLs with the "wais:" scheme; you can create a link to an existing email message with the "mid:" scheme; you can create a URL that points to your inbox with the "mailto:" scheme; etc. The URL, unlike most existing hypertext link systems, worked with the existing internet as well as with the new hypertext system. So that accounts for its very early success: the fact that the URI was technically a good idea. What about the later days? Why did the web cause a financial boom from the mid 90s to the early 2000s? Well, you'd have to ask Marc Andreessen that. As the URI made the web interesting in the hackerly/technical circles he worked in at the time, his creation of Mosaic then Netscape (presumably because of the intellectual interest in the project) brought the web to a mass audience. From there as the technical features of the web evolved -- the invention of web forms, for instance -- the web became suitable as an application platform. For most of the people who got online in the 1990s, URLs were not interesting pieces of technology. All they cared about -- and all the people making web pages and early web apps cared about -- was that the web could make interesting things with hypertext. But the URI ultimately enabled that, both technically and (early on) socially.