[jslabovitz]

Perhaps...

But if you do that, it would also be interesting to trace free software back to the point where the GPL became popular. (GPL was published in 1983, but in my recollection it wasn't until the early/mid-90s that it surged into general usage.) Before that, 'free' software generally meant public domain: no licenses, no limitation on use -- at all. We had two or three decades of free software of that type, mostly distributed over Usenet, BBSes, FTP/Gopher, etc.

I'm probably being too oblique, but I believe it's short-sighted to believe free software must only be ideologically free. There was, in fact, an earlier period, where developers release software for free because, well, why not?

Okay, time for this retired grumpy programmer to retreat back to his hut in the forest now...

[PeterisP]

My memory of that period (mid-90s-299ish) is that in a majority of cases the widely used 'public domain' software was distributed without restrictions but also without source code, and thus did not really enable modification and redistribution of improved versions even if the licence (or lack of it) didn't prohibit me to do so; so back then the 'ideologically free' movement towards open source software did result in an actual increase of practical freedoms.

[jslabovitz]

Ah, I did not explain clearly. My mistake. I wasn't talking about binary-only shareware distribution, but rather source code (primarily Unix/C-based on comp.sources.*). I'm thinking about the Usenet software itself (B News, C News, rn, tin, vnews), sendmail/bind, early Pine/Mutt/procmail (?), scripting languages like Perl/Python/Ruby, and so on. And a pile of early microcomputer code for Z80/8080/6500/6800/etc.