David Crayford wrote: >I'm not as dogmatic as Elardus WRT comments.
I'm not dogmatic. Did you see my smileys and my serious note that comments are life savers? I was just replying to Phil's tongue in cheek comment about starting a 'war'. I still remember the past 'war' about acronym USS. >In fact I'm quite the opposite. I believe that if you need lots of comments to >understand code then that is an indication that the code is poor and needs to >be refactored. Indeed. What you wrote is very true. Especially if you need to debug an Assembler or COBOL program written by others. Where possible, I insert or correct comments to explain changes to source code applied. I have once a bad case of a COBOL program producing 0Cx abends. A programmer wrote the program and called programs with few comments. Next programmer added features, but wrote comments describing the logic incorrectly. Yet another programmer 'fixed' the program based on that faulty comment. Result - I had to debug the thing and fix the comments also. >One of the compelling reasons for using an OO language is to avoid conditional >logic with language features such as polymorphism. In the case of your REXX >class I would refactor your select statement into methods for each function >type. You can then chain the method calls to perform all three functions in >one statement. Very interesting approach. >FWIW, for this kind of stuff on Linux I usually just write a bash script. Could you please be kind to provide any example(s)? Groete / Greetings Elardus Engelbrecht ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
