We need more ITT Tech programmers
-
Ugh, my wife was the director of placement at an ITT Tech a bunch of years ago (late 90s). One day she came home and said I "had to" look at this great website one of the graduates had created for his employer. It was all jpg images! That's what they were being taught. I couldn't read a word of it. Dreadful dreadful dreadful. Then again, the C code I wrote when just out of college has warts all its own. New programmers really have to work under the supervision of experienced developers.
PIEBALDconsult wrote:
Then again, the C code I wrote when just out of college has warts all its own. New programmers really have to work under the supervision of experienced developers.
I have to agree. No matter how smart a person is they need to work with someone who really knows the business. That is why in the old days trades had apprentices and journeymen that were taught by masters of their craft. Nowdays, anyone with a couple programming classes thinks they are a programmer. Bill W
-
I inherited a Borland Win32 application that has over 60 global variables interspersed throughout it. Many of the globals are multilevel struct instantiations. Nearly every class definition has a line similar to this after it: "extern PACKAGE TTree *Tree;" The actual declarations of these variables are either in the code for the "about page", above the program's "main" function, or in "globals.cpp" There is a file called "structs.h" which contains the definitions for every struct used in the application. structs that have to do with the database, visual controls, file contents, etc are all in the same file. The file is 6,000 lines long. There is a file called "globals.h" that is 3,000 lines long that is nothing but #defines. The #defines are related to, you guessed it: database, individual visual controls, different file formats, and other processing specific code. The file is included in every .cpp file. Very few of the #defines have meaning in more two .cpp files. 8 of the 28 .cpp files in the project are over 10,000 lines long. The longest is 14,433 lines. Most of the code in these mammoths has nothing to do with what the file name would have you think they are doing. Instead of using one of the hundred different XML processing libraries out there, this guy decides to write his own in order to make a way of saving data. The produced files aren't even valid XML so I can't replace it with a more efficient XML processor. I have to maintain his version of XML syntax for backward compatibility. There isn't a single pass by reference in the entire application. It is all pointers. There is no function polymorphism being performed. There are a ton of functions with parameters that specify the kind of data being handled with switch-cases that handle different kinds of data. Every struct declaration looks like this: "struct CALIB_STRUCT Calib;" Always all caps, always has the "struct" keyword, and always ends with "_STRUCT". There is no real naming scheme. Sometimes constants are all caps, sometimes they aren't, sometimes they are structs. He declares every variable that could possibly be used by the function at the top of the function. This includes those huge structs that never get used. Get this guy a copy of the C++ standard please! Many functions are over 500 lines long. Thank goodness for split-screened editors. The only container that is used are statically defined arrays (usually of structs) even when there is no way to know how many elements will be