From: MikkelFJ Date: 2002-09-27T01:35:18+09:00 Subject: Re: adding overload to ruby "William Djaja Tjokroaminata" wrote in message news:amv5dg$97f$2@grapevine.wam.umd.edu... > Hi Dave, > > I think you know that "large scale software development" usually implies a > big team of people. For a single-person development, no matter how large > the code is, Ruby will be fine, even without private variables, if the > person is really good and disciplined in writing code. > > For a large team, the story is different. I don't want the whole project > goes awry just because a person has a pointer error. (In this regard Java > is much better, because there is no pointer). Do you really think that we > can assemble 100 people now to write a single, large, Ruby application, > and not expecting performance and bug problems? Pointer errors in C++ can easily delay the release of a large product because one key component suddenly stops working as intended. I don't see how Ruby would be worse in this respect. In large projects there is a lot of abstractions at a level above simple variable types - there are so many dynamic things that can go wrong that typesafety on the variable level really isn't the big issue. For example: Adding a COM object to a COM container. It has the right type, the right interface version - everything is right except the object does not belong to the same session, so the internal references in the object are all wrong when added to the container. Large projects will divide the code into components. There is nothing wrong in having many Ruby components. Private variables aren't really going to solve anything - if you have programmers that think it is OK to circument all modularization just because they can, you've got a problem at a different level. It's no so much private variables as the possibility to inject a new method into a class that worries me in this respect - but normally you don't drive in the wrong side of the road just because you can. Mikkel