On 3/20/06, Levi Pearson <[EMAIL PROTECTED]> wrote: > On Mar 20, 2006, at 12:24 PM, Bryan Sant wrote: > Although I agree that the amount of time that it takes to enter type > annotations is not really significant, that's not really the point. > For some tasks, working around the type system takes a significant > amount of extra logic. Witness Design Patterns, a significant number > of which are methods of doing things in statically-typed languages > that require no extra thought in dynamically-typed languages. IDEs > don't really do anything to alleviate this.
Not that I don't believe you, but please supply and example. How would a dynamic language make a singleton, prototype, strategy, flyweight, facade, or adapter pattern unneeded? I can see how some of these patterns would require fewer charaters to be typed (not need to cast a value frome one type to another or create super-classes/interfaces in some cases). But again, we'er back to my original point that A) with an IDE, I'm just as fast and B) sacraficing type-safety costs in unit tests. > And another point: statically-typed languages in the ML and Haskell > families are often more compact than dynamically-typed languages. > Types are (mostly) inferred, so there are very few type annotations > in the code. There is also special syntax for pattern matching that > makes for very succinct code in situations where it's useful. I learn something new ever day. I'm switching to Haskell on my next project. > The verbosity of a language is not inherently tied to its type > system. It only seems that way because of the current crop of > languages in the mainstream. The choice of dynamic vs. static typing > ought to be made on its own merits, and both are appropriate in > different situations. The ability to create powerful abstractions > (and thus write succinct code) is almost always a win, though. Untill you have to write the documentation that explains what your code does. Not that you don't have to do this with Java, but the language is far more self-documententing than most other languages. I'm not saying that smaller code isn't better. I do think so. But in the totallity of what a good developer is expected to do (write unit tests, and document your code) Java (or C#) isn't nearly as bad off as some try to paint it. -Bryan /* PLUG: http://plug.org, #utah on irc.freenode.net Unsubscribe: http://plug.org/mailman/options/plug Don't fear the penguin. */
