>As you, I would not "wing" it, if we were dealing with something "significant"; but for 10 or 20K, I have no problem with taking the "Trading Spaces" approach. With many clients, and a smaller budget, you can rapidly consume the budget and never produce anything if you take the "classic" approach. Assuming that boredom or change doesn't kill everything first.
>
>I've worked from specs that were nothing more than: "We need a Work Orders system that can produce these x reports". "Fine, I can do it for $20,000. Sign here".

Is still very risky, but if you are comfortable with it...

>
>(That one evolved into a $300,000, 7 year part-time ERP project because they liked what I did with "Work Orders" ... and I still never produced a single spec).

If you had spec'd it, and had some meetings with the customer up front, could the project have been the 300k project from the start? And while you say you "never produced a single spec", was there any spec anywhere? I can't believe any company would put that kind of money into anything without some sort of idea of what they want.

There's a lot of difference in billing a client for development based on their own internal specification, and determining what amount of work is involved in a particular project based on the deliverables of the project.

I guess we will have to agree to disagree on this as I will never do any project without a clear definition of what is to be delivered. This protects both me and my client. My client knows what he is to receive upon delivery, I know what to deliver.