If I really did it, probably the thing I did that became the most famous was being the first to suggest that internet addresses be in the form site "dot" toplevel-domain.

I recently read an article claiming that "dot" as in "dot-com" was the "most useful word of the year" as chosen by the American Dialect Society -- the most widespread addition to the language. And it certainly has become a convention that literally the whole educated world has come to know.

It sparked a memory that long, long ago I had been in arguments with people in various areas about how multi-level names should be written. Suggestions included user@site@domain, user%site@domain and the leader, based on a proposed standards document, user.site@domain. People only thought in terms of adding a second level in those days, and the "domain" was thought of as a "forwarder" -- a top level site that would know the sites underneath it and handle their mail.

(Before this addresses had been one level -- user@site where the "user" was sometimes a magic string that implied mail forwarding to other sites, notably UUCP ones.)

For valid reasons, I thought it made sense to have the user part on the left of the at-sign and the computer part on the right hand side, and that the levels on the right hand side should be divided by dots. So I said that user.site@forwarder was bad but user@site.forwarder would be better.

I searched and found the earliest suggestion of mine to this effect in the 12th issue of the TCP/IP digest -- the ARPANET mailing list where the TCP/IP protocol and the future convetions of the internet were being discussed. (My userid in this digest is watmath!bstempleton under old UUCP conventions. I was an undergraduate at the University of Waterloo at the time and had in fact been the one who got it connected to the net, just so I could get involved in such discussions.)

Now while there are aesthetic arguments as to why this is better than other suggestions, the truth is it's pretty arbitrary. The choice of this form was no great act of creativity or insight. And indeed I don't even know if the people who finally did code up the systems did it that way because of my suggestions, or they thought it up on their own. If somebody finds reference to somebody bringing it up earlier I will gladly retract this claim.

Indeed if it was me, it was simply by virtue of the fact that having been around at the beginning of these things, and taking an interest in these issues. Being in the right place at the right time. But it's simultaneously mind-boggling, conceit-building and humbling to think that what I said might have sparked something that became so universal.

Of course, Jon Postel, who wrote up most of the specs later, has probably the most significant place in the chain of origin of all this.