I didn't say you don't have to think when using ICE. But you certainly have to think less. You can easily single step until you see your program do some branch you weren't expecting and then you think about only that. If you don't have ICE, you go through your program and really think hard about each instruction. When you're a newbie, you go back and verify that you really know what each instruction is doing, what flags it is setting, etc.

When you use ICE, you may end up doing that--but only for the instruction that unexpectedly branched, or the flag-setting instructions immediately preceding it.

ICE definitely makes you more efficient. No doubt about that. But I think a newbie benefits from actually suffering through their errors and not just thinking about the actual instruction that turns out to be wrong, but along the way studying and confirming their understanding of many instructions that were actually right.

You learn through failure, not success. ICE can make a failure an awfully short experience which means you're going to learn less from it. When a failure makes you question your entire program, the newbie ultimately learns and benefits from that.

IMHO.

Regards,
Craig Steiner