>>21
For reasons I personally consider to be obvious to the mere lot. From your question, though, it either might not be as obvious, or you wholly disagree with them, which is very acceptable.
Lisp is a functional language. Grossly speaking, functional languages combines elements in the language to form
relations (or functions) between such elements, in contrast with
actively modifying or calculating things with these elements. In Lisp, there's no such thing as 'incrementing a variable' (disregard momentarily 'setq' and others); instead, you unary relate a variable to return a new element, which happens to be the successor of such variable. In Lisp, there's no 'remove a node from a tree'; instead, a tree, a node and another tree are tied together in a relationship called 'remove'. Lisp semantics are of
correlating symbols to form another symbols. Strictly speaking, there wouldn't even be any output from a Lisp program if it were purely functional (that is, side-effect expressions weren't introduced).
I believe you're fully aware of this explanation.
Some applications are much better expressed in terms of what the computer should
do, instead of which relations should the computer perform. Webservers and device drivers are examples of such applications: it's more natural, it's easier to express the program in a list of
do this, do that instead of trying to combine symbols in lambda-relations. Databases, on the other hand, are not -- that's a reason why there are some models of database systems, and different languages to access them. It's easier to express a database access through another set of primitives of thought; in other words, it's easier to express database access in another
paradigm.
Computer languages are all about expression, but are affected by other issues in practice, such as performance. In this case, though, writing a webserver in Lisp is against the language's expression. It's not a matter of performance or other implementation issues. Writing imperative programs in Lisp is inexpressive.
>>23
I'm not aware what would a 'continuation-based' web browser be, although I believe this is the exact same case as above. Scripting support inside webpages, or content-generation, however, is a different matter, and maybe Lisp has some sane horizon inside web deployment.