It does not have to have a problem to be a fad. Programming languages all require a big time investment in learning them - so those that exist tend to stay around. People only change languages when there are very clear benefits. Clojure seems to have pulled up to the same level as other much older and better known programming languages - there are no benefits (at the moment) to it that would make the mass of programmers switch to it.
Ruby is a good language, but it only became popular when it offered a clear benefit.
For a language to enter the mainstream and stay in the mainstream, it has to do a lot more than just not be bad. It has to be better than others in some way.
(By the way, being a functional language does not make it better, it makes it worse, since most programmers have trouble grasping functional concepts)
At the moment, clojure may not have any negatives, but it does not have any strong positives, so I don't see why it would catch on.
Suppose your arguments are true. Here's why clojure is destined to be big.
Clear benefit: you have Java at your disposal, while being "new and cool." From a library standpoint it is Java + extras.
Functional does not make it better: if there are two types of people, one that prefers functional, one imperative, go ahead! Let the functional use clj, the imperative use java. Now you can compile to .class and use either paradigm!
I don't code lisp regularly, but I'm convinced clojure is the NBT and so I'm preparing to jump ship after it matures :D
Can you really compile Clojure into .class files? I was wondering that the other day, and the conclusion that I reached was that it doesn't work that way.
Frameworks are big in Java-land because you need an enormous amount of scaffolding to make any non-trivial application work. That's less true in other languages. Libraries and interoperability with other languages are more important, and where LISP has traditionally suffered. Clojure has a lot of potential, but there are a lot of languages poised in the wings right now, all jockeying to be the next big thing. F# for example is functional (it's descended from OCaml) and has full access to the .NET libraries, which are as comprehensive as Java's, and has the backing of a major vendor. Now if Sun were to pick up Clojure and make it the official functional language on the Java platform, that would be something.
Ruby is a good language, but it only became popular when it offered a clear benefit.
For a language to enter the mainstream and stay in the mainstream, it has to do a lot more than just not be bad. It has to be better than others in some way.
(By the way, being a functional language does not make it better, it makes it worse, since most programmers have trouble grasping functional concepts)
At the moment, clojure may not have any negatives, but it does not have any strong positives, so I don't see why it would catch on.