Or are they an accident of happenstance much like "http:// ?
Or are they an accident of happenstance much like "http:// ?
Someone correct me if there's a juicier anecdote.
Here is what I mourn: that we write subdomains from right to left, instead of right to left. We see it in programming namespaces, but I still wish we had com.foo.bar.baz instead of baz.bar.foo.com
Combine the two concepts, and we end up with a hierarchy of names. And because all of the computers in the world are connected to this global hierarchy, TLDs were introduced to add some order to the chaos. By introducing TLDs, we were able to start to categorize companies from universities from network infrastructure providers, etc... and thus everyone could find their place in the global name hierarchy. Even the TLDs are part of a single root object in the hierarchy(.).
It wasn't until things started to scale that we had issues with domains.
Playing shell games with new TLDs will in no way make this less confusing for users.
Am I misunderstanding your question? There needed to be a way to separate, say harvard the university from harvard the ISP from harvard the non-profit to harvard the commercial company. Etc.
Although now (with the exception of .edu) .com .net .org can be used by anyone for any reason at the start they were a way to differentiate different types of sites so many people could use one string. And of course there were country code domains, (say harvard.it which is owned by a Italian company) etc.
With no TLD's there would just be http://harvard - one user of the string harvard.