I completely agree that the foundations were laid long before 1998. I'm sure we can both name lots of milestones. You cite 1989 for the GPL v1. I could cite 1985 for the invention of the patch utility and the writing of the GNU manifesto. You could cite 1978 for the first release of BSD. And so it goes. Open source software could not have had the breakout year it did in 1998 without being well along a path to success before then.
However I believe that 1998 was the critical year for the widespread acceptance and adoption of open source technology in the web industry. And a critical enabler for that was the fact that Microsoft got slapped with the anti-trust lawsuit, following which all of the major competitors realized that they had a limited time window to take down Microsoft, and therefore all jumped on the open source bandwagon. Do you remember how much it changed the game to have IBM and Oracle putting their reputation behind Linux and Apache? Obviously Microsoft was fully aware at that point that open source was a huge threat, however they couldn't jump on it because their hands were tied by the lawsuit.
Without the lawsuit the major vendors wouldn't have dared to get behind open source. That would have slowed adoption and acceptance significantly. Furthermore Microsoft would have been free use their control of the client to leverage control of the server. Which they were already starting to do, and which they had done to very good effect against Novell earlier. Could they have succeeded? I sure wouldn't have wanted to bet against them!
Thus I think that Microsoft's blinking at the lawsuit in 1998 was more important for the success of open source on the web than their missing the web before 1995.
The fact that 'open source' got named in 1998 gave it a lot of publicity, but Linux and a whole bunch of other stuff would not have existed in their current form if not for the 1989 version of the GPL.
Open source/Free Software was a movement long before it had a name. MS took too long to take it serious, and the effect that universal net access would have on accelerating open source was very large.