On Oct 12, 2006, at 22:13 UTC, Massimo Valle wrote: > > One of the primary reasons, as I understand it, for RS needing to > > have spent time on the compiler was for maintainability. And I > > think the results of that is very, very obvious. We have multiple > > new targets in an amazingly short amount of time. > > This seems like propaganda to me.
Call it whatever you will, I guess, but RB as a language is way, way beyond what it ever could have been with the original compiler. There would be no Linux target, probably no Mach-O target, which of course means no Intel target. There would be no "for each" statement, no basic optimizations (e.g. constant folding), and so on. Instead, there would probably still be a slew of bugs (anybody here remember the weird problems with floating-point math on Windows way back when?) which were either intractable, or prohibitively expensive to fix. > Maintainability of the compiler is not something you can call a > feature to advertise. True. I don't believe they ever did. It IS, however, a necessary prerequisite for many features people were clamoring for. > But, if you talk about a new compiler, I'd expect to see a smart > optimized compiler and linker. Good, you got that. There are a lot more optimizations done in the new compiler than in the old one, and the linker is a lot smarter, too (which is why, for example, RB can now generate Mach-O as well as PEF). Could more optimizations be done? Sure -- thanks to the clean new compiler design. (You'd never have seen those in the old compiler, either.) > Not something that still force me to > store the argument of a for...next cycle into a variable to avoid > calculation at every loop. What?!? That has nothing to do with optimization; it is part of the language definition. It is a feature that the bound of the for-statement is reevaluated on each pass of the loop, since it means that if that boundary changes, your loop terminates (as a natural reading of the code certainly indicates that it should). If you don't like this feature, then fine, don't use it. But don't blame its presence on some lack of optimization. > Not something that fails to determine correct plugin dependencies and > bloat my applications with unnecessary code. If there's such a bug, I'm sure they'll fix it. Bugs happen. Reporting them well helps them get fixed; whining doesn't. Best, - Joe P.S. All opinions expressed herein are my own personal opinions as an RB user, and don't reflect the opinion of, well, anybody else. -- Joe Strout -- [EMAIL PROTECTED] Verified Express, LLC "Making the Internet a Better Place" http://www.verex.com/ _______________________________________________ Unsubscribe or switch delivery mode: <http://www.realsoftware.com/support/listmanager/> Search the archives of this list here: <http://support.realsoftware.com/listarchives/lists.html>
