IBM Mainframe Discussion List <[email protected]> wrote on 12/13/2007 03:33:02 PM: > > Creating good test data is a skill that many programmers have failedto master. > > At my first application programming job, most of the new programmers were > given an assignment of a training program. I never had that assignment, but > from what I understand of it, the specifications called for card > input and it was > expected that all input would be validated with appropriate output produced. > When the programmer thought that the program was ready, the boss would > give them an object deck for input. Most of the time there were S0C7 abends.
The favorite assignment of professor Jack Hollingsworth (at RPI in the 1970s) for his Assembler Language Programming class was the Data Validation project. The students had to write an assembler program to validate fields on an 80 byte card image, and provide (and be graded on) test data for their program. As the teaching assistant for the class, I had to grade the projects and the test data. To validate their programs, I ran them with my test data. To grade the test data, I wrote a program for grading test data in XPL (which we happened to be using in the Systems Programming class I was taking for an assignment to write an assembler for the HP 2100 minicomputer). Prof. Hollingsworth thought the test data grading program was innovative, but I have to admit that it was just laziness. Grading test data is rather tedious. I also recall Prof. Hollingsworth saying "OS/360. Now that's a program that will never be completely debugged." Little did I suspect that 30 years later I would still be debugging it. Jim Mulder z/OS System Test IBM Corp. Poughkeepsie, NY ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
