home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.lang.fortran
- Path: sparky!uunet!spool.mu.edu!agate!dog.ee.lbl.gov!hellgate.utah.edu!lanl!cochiti.lanl.gov!jlg
- From: jlg@cochiti.lanl.gov (J. Giles)
- Subject: Design of C++ et. al. (was: Is FORTRAN a viable language?)
- Message-ID: <1993Jan5.184413.3927@newshost.lanl.gov>
- Followup-To: comp.lang.misc
- Sender: news@newshost.lanl.gov
- Organization: Los Alamos National Laboratory
- References: <1hqnn7INNfac@shelley.u.washington.edu> <1993Jan4.201229.12073@newshost.lanl.gov> <29250@oasys.dt.navy.mil>
- Date: Tue, 5 Jan 1993 18:44:13 GMT
- Lines: 40
-
- In article <29250@oasys.dt.navy.mil>, roth@oasys.dt.navy.mil (Pete Roth) writes:
- |> [...]
- |> A third possibility: the concepts of the language are new, and so the
- |> terms "easy to learn/use/describe" are not yet applicable. When we all
- |> understand what virtual methods, templates, overloading, private|protected|
- |> public, friend, class, etc. mean, *then* the issues will be well defined.
-
- That could explain the traffic on comp.lang.c++ but not that on
- comp.lang.c. Further, there are other object oriented languages
- which don't seem to present such difficulty or cause so much
- confusion.
-
- To be sure, C++ is better than C, and Objective C is better still
- (though not as popular, so its development is falling behind) - but
- the main problem seems to me to be the commitment to backward
- compatibility to C.
-
- C++ also shares the C characteristic of haphazardness: more and
- more features are "thrown in" rather than a few well designed
- features with greater power. Hence the presence of all that new
- jargon when a subset should be sufficient.
-
- Finally, if the new concepts are not easy to learn/use/describe,
- perhaps the concepts are not worth the effort. There are other
- ways to provide reuse of code, encapsulation, and the other abstract
- objectives which object oriented concepts supposedly promote which may
- be easier to learn/use/describe. Functional languages are a possibility
- in this regard.
-
- This discussion only has indirect relevance to Fortran. Eventually,
- Fortran implementations will require extensions to provide some
- of these features and the observation of the mistakes mad in other
- languages will be of importance. As far as I can tell, it is too
- early yet to state conclusively whether OOP is even a good idea
- or not (as compared to alternatives like generic procedures or
- functional extensions). In any case, followups are directed to
- comp.lang.misc.
-
- --
- J. Giles
-