Thanks for the comments. My desire is to still have the 30 questions, although your comments are valid.
One note on the GOTO: yes I agree that in theory the goto should never be used; in practice however a little goes a long way.
For example, Let's say that you are validating screen fields for errors. There are 6 fields each of which could be in error. In the error validation subroutine, an error is encountered in the first field so the message is retrieved, etc... The 6th field is in error as well. Same logic applies. With the PC and RI attributes on each field with different indicators, when the program returns control to the screen both fields are RI, but the PC is on field 6. The message indicates that field 6 is in error, so the user corrects it, and presses enter, which returns another error, with the correct message now on field 1.
My preference in a situation such as this is to use the GOTO and tag features offered in RPG because once an error is encountered, why bother checking all of the other fields? While the processing time may not be that much, it is non value added. Additionally, once the error is encountered, it is dealt with and the screen now has control so the user can correct.
If you can come up with 30 good questions, I will review and comment, and you could possibly wind up with 500 points. As a reminder, the question states that ya gotta think about this one. Your comments are good, but with no questions, I have to let this one remain open. I do want to know your opinion on the scenario I played above concerning screen field validation. With 14 years as a programmer, you definitely have the goods, and I am confident you could go into the wayback machine and get some questions from prior interviews you have been through.
jjjjjjj
Main Topics
Browse All Topics
by: mmcmillenPosted on 2002-12-09 at 16:13:18ID: 7557123
I have always found asking questions about specific navigation through applications as unfair and deceiving. Just because a user doesn't know a command and all its parameters by memory does not mean they don't know how to use it and don't have the skills. For example, I have been on the S/36 and As400 for 14 years and I have never typed in wrkmbrpdm. I just typed it in to see what it was. I recognized the screen and also that I get there from the strpdm menu. 10 different people will have 10 different ways to get to the same place.
If these interviews are going to be actual tests I would set up something on the system for them to work with. You could print out a program that does a few things, some obvious , some not so obvious and have them flow chart it or tell you what the result would be from given sets of input.
Design a program on the AS400 that has some bugs in it. First, have some that keep it from compiling. Then after they correct that, have some logic errors.
Test them on subfiles. Test them on database theory. Ask them about when to use goto's(NEVER!). That type of thing.
Give them an abstract situation. Ask them to give you a solution.
Don't expect them to know every opcode syntax verbatim. Don't expect them to know what position field 2 is in an rpgIV program. There are people that know that stuff and make good coders, but in my experience they can't think out of the box.