How Much Time Can I Save Not Writing Tests?

Is there a conclusive argument for or against TDD or Test Driven Development?

One Example: Bug 1109

I just closed bug 1109 after 5 weeks of working on it, little bit at a time between other more serious issues.

It fixed in one line of code. I've probably got a dozen or more hours in it. It was pretty obvious, when I found it, but it took me at least12 hours, maybe more, of searching to find it.

If I had written it in TDD style, it would have been a 10 minute find, but it probably wouldn't have happened because I probably would have found it before I even finished the code.

We Didn't Use TDD Because We Didn't Have The Time

Have you ever heard that line of thinking? That's the excuse I used for not employing TDD on this project. We didn't have the time.

I'm thinking we might have had the time. I'm thinking, maybe that was just a hollow excuse.

This Blog - Updated One Month Later:

Filed bug 1243 actually 25 days after writing this blog. 

No, of course I had never really thoroughly tested the fix. I tested it enough to know that it at least did the minimum of what I needed. Bug 1243 was a different bug that came up because ... wait a minute. My fix for 1109 created a side effect that added tens, may hundreds of thousands of "extra" records to a table.