home *** CD-ROM | disk | FTP | other *** search
- Path: ra.nrl.navy.mil!usenet
- From: pitre@n5160d.nrl.navy.mil (Richard Pitre)
- Newsgroups: comp.lang.c++
- Subject: rant
- Date: 8 Feb 1996 18:32:07 GMT
- Organization: Naval Research Laboratory
- Message-ID: <4fdfj7$o5n@ra.nrl.navy.mil>
- NNTP-Posting-Host: n5160d.nrl.navy.mil
-
- Two of the last three compilers that I've used have failed before
- I even completed all of the project modules.
-
- The first compiler was a Fortran90 style compiler. I was using it because the
- builtins were a match to my project. I foolishly thought that this would make
- my code easier to manage and debug. You are probably thinking that I got what I
- deserved and perhaps you are right. I should have stuck to my intuition that
- Fortran is a dinosaur, updated by a committee, and that I might not make the
- retrograde adjustment. In any case it explicitly broke. I made a bug report and
- then I gave up on that sack of shite. I'm using the C compiler because the
- system doesn't have a C++ compiler yet.
-
- The second compiler that broke on me was a popular C++ compiler which failed on
- the first compile. Given the complexity of the language this failure is more
- understandable than the failure of the Fortran compiler but for me personally
- its a lot more upsetting. It happened a few hours ago and I'm writing this to
- unwind. I still respect the people who made this compiler and, all things
- considered, I'm not too upset with them in particular. Maybe we need a
- platitude that says something inanely obvious like : A language isn't too
- useful if theres almost noone with human ancestors who can implement it
- correctly. C++ is complicated to use and probably much more complicated to
- implement. The grammer and semantics are so intertwined that the thing needs at
- least a meta meta grammer specification in addition to that mess at the back of
- the reference manual. As it stands future generations of programmers may turn
- into the equivalent of biblical scholars who try to translate the text along
- with what they know of some dead cultures into something meaningful to
- religious fanatics of the day. At this rate Stroustrup will be disintered and
- C++ implementers will carry around little bags with his finger bones.
-
- In both the case of the Fortran compiler and the C++ compiler I am guessing
- that it was the type/use checking that caused them to fail. I don't ever expect
- to trust my compiler as much as I trust my text editor but it would be nice if
- I could. Even so I'm sure that my text editor sits on an os that is infested
- with bugs and that the os in turn sits on a bug infested cpu and associated
- support chips. Maybe there are only a few bugs at each level. Today its almost
- realistic to ask if the number of bugs per code layer goes directly or
- inversely with the layer depth. I guess I'm just being negative but where in
- the hell is this industry headed when the development tools and the hardware
- they rest on are becoming more and more unreliable. If the objective is
- animated singing icons that are "real cheap" and easy to use then I guess there
- really isn't a problem and I can start being Mr. Positive ;-) again.
-
- richard
-
- Feel the FARCE Kook.
-
-
-