home *** CD-ROM | disk | FTP | other *** search
- Path: news.polymtl.ca!news
- From: Pierre Ferland <pierre@meca.polymtl.ca>
- Newsgroups: comp.lang.c++
- Subject: Re: C++ experts only
- Date: Thu, 11 Apr 1996 10:37:32 -0400
- Organization: Center for Applied Research on Polymers
- Message-ID: <316D192C.2781@meca.polymtl.ca>
- References: <4jhrb4$gk2@melon.myna.com> <3164559A.41C6@meca.polymtl.ca> <4k9blh$8a@news2.ios.com>
- NNTP-Posting-Host: pegase.meca.polymtl.ca
- Mime-Version: 1.0
- Content-Type: text/plain; charset=us-ascii
- Content-Transfer-Encoding: 7bit
- X-Mailer: Mozilla 2.01 (X11; I; IRIX 5.3 IP20)
-
- Vlastimil Adamovsky wrote:
-
- > Not having a standard garbage collection is an advantage. If you ever
- > need one, buy it. Generally you don't need it. If you think you need
- > it and you think you cannot buy your favorite garbage collector
- > anywhere, then use Smalltalk. It has a garbage collector.
- > Simply, don't create garbage in your programs and you will not to need
- > to collect it.
-
- Buying a garbage collector would be another way of patching C++ lacks,
- as when we have to purchase/test/reject API packages and other
- stuff. We spend too much time looking for tools to help us
- in the development process. Each tool has its bugs, non portability
- problems, costs, documentation and philosophy. The key problem
- is that we now have to develop high level applications (that include
- graphics, multi-threads, sounds, networks, GUI) while we use a
- low-level language like C++.
-
- When I talk about garbage collection, the point is not really the
- objects destruction: people should delete their objects when no longer
- needed, I agree. But the C++ concept of a pointer really isn't
- sufficient. How many references are made to a specific object ?
- Which client is allowed to do what on that object ? Is this
- a valid pointer ? All we can do is wrap our pointers into
- more intelligent classes, and... deal with the libraries.
-
-
- > We can call it an evolution, which is to preffer to revolution.
-
- Depends on how long it takes to change, and on how many libraries
- will need to be changed, since, remember your point, C++ can
- do virtually anything as long as you build the appropriate library.
- If the long term efficiency of C++ is based upon a good
- set of libraries and tools, you'd better not have to touch
- the basic ones!
-
-
- > >The issue should not be C versus C++, forget about C.
- > >But don't trust C++ neither.
-
- > In god we trust.
-
- Yup! I also used to trust spring, but I recently lost my
- innocence on that point!
-
-
- Pierre Ferland
-