Analysis has to precede design. If the customer hasn't done his requirements analysis, then I can't help him, unless, of course, he wants me to do it.

Most of the customers I work with have worked with us before, hence, I know how to communicate with them. I haven't had to "sell" work in over 20 years. People come to me because they know I'll produce what they want. The ones who don't, themselves, know what they want don't seem to come to me. I'm fine with that! Figuring out what's required is the biggest and most difficult part of the task. It has to be done before any sort of design can begin. After all, how can one produce useful work product if one doesn't know what it has to do.

It's true, one can say, "I want a device that does this ..." and then, "I want it to do that, too ..." and, "Well, it doesn't do that fast enough ..." but that always seems to lead to production of rubbish at very high cost and long delay.

I don't consider design and coding to be "artsy" disciplines. You can't do them by "feel." Engineering is the task of specifying a precise solution to a clearly specified problem, with existing technology. You don't have to invent anything. You have to know and understand the task and figure out how to accomplish it. You can't do that if you don't have a clear objective specification. You can't engineer an enigma.

Guessing at what has to be done produces rubbish like that popular cellphone that was on rev 4 before it was functional.

RE