home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.lang.c++:18838 comp.object:4737
- Newsgroups: comp.lang.c++,comp.object
- Path: sparky!uunet!wupost!spool.mu.edu!news.cs.indiana.edu!noose.ecn.purdue.edu!author.ecn.purdue.edu!roger
- From: roger@author.ecn.purdue.edu (Roger J Cass)
- Subject: Re: Pros and cons of C++
- Message-ID: <roger.726503589@author.ecn.purdue.edu>
- Sender: news@noose.ecn.purdue.edu (USENET news)
- Organization: Purdue University Engineering Computer Network
- References: <C0Hp1n.vp@gpu.utcs.utoronto.ca>
- Date: 8 Jan 93 14:33:09 GMT
- Lines: 51
-
- I am an engineer (well, actually still a grad student), so my approach to
- using a language is maybe a little different. I was involved in a finite-
- element mesh generation package development group as part of my graduate work.
- We chose C++ for a number of reasons. based on the technology as it stood
- about 2 years ago.
-
- First, like some of the other articles have stated, C++ offered the widest
- platform of hardware. At Sandia Nat'l Labs, where my research was sponsored,
- the software had to be able to run on a number of high performance
- workstations. We were engineers turned programmers, not the other way. We
- knew exactly what the software had to do, because we were inventing the
- solution as we went. We had to pick some sort of standard that would allow
- us to concentrate on getting better at mesh generation rather that worrying
- about having to write the entire package.
-
- In our search for modules that could take care of as much of our program as
- possible, we discovered that OO technology was really the best *idea*, if
- we could just find an existing solution.
-
- We chose C++ because we needed a solid modeller that we could build on, that
- would become a module in our program, not the program itself. The ACIS
- modeller by Spatial Technologies was the deciding factor. We also needed
- a GUI, and X seemed the best alternative since it was built in to all the
- workstations. Here again, the match between C++ and C was the clincher.
-
- In the course of our program development, I ran into limitations in C++ very
- quickly, but most were able to be worked around. If our original concept
- was purely OO, then our final product (which is available as part of DOE
- technology transfer, contact me via email) is a network of patches to
- make it a program. The original OO framework has been lashed together
- and short-circuited. I am the first to admit that if I had been a better
- C++ programmer, it wouldn't have been so bad, but overall I think we did
- the best we could.
-
- SUMMARY-- In my observations at 2 universities (BYU and Purdue) and at a
- National Laboratory (Sandia) not very
- many engineers have embraced the OO paradigm for modelling physical
- systems and solving engineering analyses. C++ is a good way to move into
- OOP, but does not allow the OO paradigm to be realized in full-- which is
- important to an engineer/programmer who does not care to become a guru just
- to solve a problem. Many engineers of still using **hack** FORTRA(SH)
- (sorry, can't help it), and wouldn't care too much which OO language to
- move into, except that it must be widely available.
-
- I am still waiting for the successor to C++ as the OO programming language
- of choice so I can get on with OO problem solving instead of programming.
- --
- Roger Cass ! No success can compensate
- roger@ecn.purdue.edu ! For failure in the home
- Purdue University CADLAB ! David O. Mackay
- (317)494-5944 !
-