Ovid mentions a 1/4 million line Perl project he worked on which certinally shows that Perl can be used for BIG projects in the right hands.

My argument is more to do with "stricter" languages picking up egregious mistakes earlier when they cost less (by whatever measure) to find and fix. As projects get larger the cost of finding and fixing such mistakes gets higher, almost exponentially.

To put this all slightly in perspective: in many ways I'm a sloppy programmer who depends too much on my programming tools to find silly errors for me. For large projects such errors have a better chance of being detected early with C++ than with Perl.

I don't have enough experience with a wide range of languages to say which might be "The Best". But from my experience, at one extreame, cost per line of assembly language is much higher (and does much less) than anything else I've used and it doesn't scale well at all. C++ and Perl are definately toward the other end of the scale. I've not written code for a BIG project in Perl and I suspect that for me and my style of programming, Perl wouldn't scale at all well. I am involved in a BIG C++ project and know that, despite my sloppy coding, the project is progressing in a saticfactory fashion with about 6 people working on different parts of it and that it is maintainable.