Just a few observations. 1. Any argument that appeals to simplicity in an argument against Perl is using so narrow a definition of "simple" that the meaning is lost. It isn't unlike the way that "speed" is raised as a concern in scripting languages versus compiled ones. Is it faster to write, or to run? Is it more simple to write, or to run (or to read)? Perl's popularity is undoubtedly due in part tk the large number of newbie programs that ran properly the first time, much to the surprise and delight of the newbie. If you argue that the ease of programming in a language has the corollary of requiring an "expert in Perl" to be able to read it, then you are begging the question. It's true that idiomatic syntaxes introduce language complexity. But to many, idioms are what make a language interesting, and in many cases, simple.
2. Extensibility is one area where Perl has nothing to prove. It may be true that not everyone is an XS fan, but XS is not as difficult as some portray it. Nor is it the only way to extend (extend in the C sense, not the language sense Nat pointed out), given SWIG and Inline. 3. Embedding Perl is not fun, despite the mod_perl example (a good example, mind you). Tcl has it all over Perl here. Before Perl 6 was begun, this was the main complaint about Perl (and remains my own main complaint). Remember it was the reason behind Topaz, and other discussions about rewriting the internals? Although this is one of the goals of Parrot, I hear a lot more about the language design than I do about the rewrite, and that's not necessarily a bad thing. But my interest in Perl 6 is much heavier on its embedding API than the language changes. I was so disappointed when I learned that "embperl" wasn't something to run on my Palm. ;-) - Dan
