Java is on a very short list of the world's most used programming languages. This strikes me as another case of "people overestimate a new technology's effect over the next five years and underestimate its effect over the next 10 years."
Java did not replace existing desktop software. But it took over most enterprise web server-side development and in the process helped to make most enterprise desktop software development irrelevant.
So, perhaps the initial hype did not play out exactly as expected. But the success of Java does suggest the level of hype was justified, if misdirected.
When I started using Java in 1996 - people looked at me funny and asked why I wasn't using C++. Everyone said Java was a fad and that it wouldn't scale.
It was. And it didn't. Very few applications have been created in Java. Java stuck in the isolated server environment, but everything survives there, since servers are like zoos: easily controlled environment where animals get exactly what they need without losing body parts in the wild. You can build your own toy language and it will work perfectly fine on your server.
Java doesn't scale? Really? A lot of work has been put into java since the beginning to make it fast. Jruby for example is faster than C Ruby. Java might not be adequate for web development (I never liked the initial model and by the time better frameworks came out I had given up) but there are many applications where it really shines.
Jesus... "Java is faster than C" argument again. Inferiority complex aside, I did not mean scaling in a technical way. Java did not scale out as a general purpose programming language - it remained foreign and inadequate on all major platforms and, as a result, no software (in comparison to C family) has been done in it, with a notable exception of zoo-like server environments. Go into Debian packages repository and see for yourself. Or, for Windows users, into your C:\Program Files folder.
I hope nobody will be stupid enough to start listing (one by one, literally) Java software projects, like Eclipse, Lucene, etc.
Java did not replace existing desktop software. But it took over most enterprise web server-side development and in the process helped to make most enterprise desktop software development irrelevant.
So, perhaps the initial hype did not play out exactly as expected. But the success of Java does suggest the level of hype was justified, if misdirected.