Computer Programming Is Hard!

Thank you, sir!

Well, ask him the details on how IBM loosened their very strict dress code. I believe this was all post-PC, so sometime in the late 80s, maybe? (On the other hand, I was doing some work with IBM SDLC systems, so it's possible I was aware of IBM culture earlier than that, but I really do think post-84.)

I wish I did have some of my old card deques. JCL-wrapped PL/I or FORTRAN with the diagonal magic marker stripe down the side to help sort them after you've dropped them....

I do still have some of my very first student programs on 7-bit punched paper tape! :-)

Welcome by, this is good. I have working programmers in the Computers thread and working artists in the Art thread! :-)

Heh, yeah, sometimes it's like doing this jigsaw puzzle with many thousands of pieces and some of the pieces you can't see and have to kind of fit blindly.

Doncha just love it when the docs for an API don't bother mentioning lots of "unimportant" details, so you have to start throwing random values into the thing to see what happens.

To some extent, it depends on the nature of the problem. Sometimes it's obvious where the problem has to be, other times it can be very mysterious.

In some cases, like troubleshooting a radio receiver, there's a progression, so you can jump in the middle and see if things work in the middle. If they do, the problem is "down stream;" if not, the problem is "up stream." Jumping in the middle of a progression cuts the possible trouble locations in half.

Sometimes programmers pepper the code with "print" statements that write to a file. This gives the programmer a log describing what's going on moment to moment. It's as if a piece of paper could keep a log of its course through your press. The log lists conditions along the way. Or sometimes the log just dies, so you know that's when the killer jumped out of the bushes.

Modern development systems provide programmers with a wonderful tool, called a debugger. A debugger lets you step through the code inch by inch. It's as if you could step a piece of paper through your press inch by inch but at the same time maintain all conditions of full speed. But at each step of the way, you can inspect all parameters before continuing.

But in the end, just as with debugging a press problem, it's really just detective work. It's based on you knowing how your press is supposed to work, so you need to find where and why it's not. And just like with presses, the more experienced you are, and the more you know about printing and printing presses, the more likely your response to a bug is, "Oh, that."