From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: X-Spam-Status: No, score=-1.3 required=5.0 tests=BAYES_00,INVALID_MSGID autolearn=no autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 103376,df854b5838c3e14 X-Google-Attributes: gid103376,public From: pitre@n5160d.nrl.navy.mil (Richard Pitre) Subject: Re: C/C++ knocks the crap out of Ada Date: 1996/03/20 Message-ID: <4ipjef$2ml@ra.nrl.navy.mil>#1/1 X-Deja-AN: 143416121 references: <4ipbdb$6j7@tpd.dsccc.com> organization: Naval Research Laboratory newsgroups: comp.lang.ada Date: 1996-03-20T00:00:00+00:00 List-Id: In article <4ipbdb$6j7@tpd.dsccc.com> kcline@sun132.spd.dsccc.com (Kevin Cline) writes: > There is much to like about the Ada language, but it just isn't practical > for development of medium-scale (50-100K SLOC) UNIX or PC applications > with a significant system interface. The high startup cost and portablility > problems overwhelm the advantages of more stringent compile-time > and run-time checking. Please say more about why this is so. Is it primarily the X Window interface that caused you problems? Do you have any opinion about Ada-Tcl. Any more detailed experiences with relative costs of implementation, interfacing, maintenance etc. in C++ vs. Ada would be very useful. All of these things are things that I'm trying to weigh in moving away from pure C coding. Email is fine. I'm encouraged to see that GNAT is available on so many systems. C++ is available everywhere but I have a concern that the features that I like in C++ are not uniformly available. What are your observations on this issue? I've had a little luck with Borland C++ but I don't believe that the feature set of Borland C++ is portable. The Gnu g++ compiler is a potential solution but it did asked for a bug report on the very first file, that I gave it. Dismay. Perhaps g++ it could use some DoD support like GNAT had. > For 1M SLOC projects the advantages of Ada > may outweigh the disadvantages. > -- > Kevin Cline How, in your view, does GNAT affect this final assessment? My application area is sonar simulation for research purposes. The codes are usually not large but they have many interacting components i.e. they get fairly complex and they can get to be of moderate size. Its necessary to interface with old and new Fortran libraries. I'm definitely leaning toward GNAT and am experimenting with it now. Based on a limited experience with the overhead of making GUI's, I confine myself to file IO and use commercial packages like Mathematica, Matlab etc. on the front and back ends. A coworker is looking into the practicality of Tcl/Tk but its not likely. The closest thing that I have seen to a GUI builder that is implementationally cost viable in my setting is NeXTStep Interface Builder. Nevertheless, it is useless because of the non portability. I guess there are X-Window tools that surpass it now. If so I would like to hear about them. richard