I see that we are in agreement. My approach to the answer has more to do with my history than anything said in this thread. A fair number of painful projects have been the result of the following three step process:

• Customer or Boss: Can we solve problem X with technology Y?
• Me: Probably, if the problem is partitioned correctly and if Y works with our infrastructure, ...
• Customer or Boss: I told the CEO we could have technology Y in place next week because you said it is easy.

Since some people stop listening at yes or probably or whatever, I have gotten really paranoid and toss in the risks up front.

To be fair, I was responsible for the underestimating work comment, but not the other two. I intended the comment more to set expectations than to completely discourage.

As a result of my experience, I tend to focus more on the risks than the benefits. It's been my experience that programmers, in general, are extremely optimistic about what they are capable of doing. (Hubris, maybe?<grin/>) So, I don't worry about that part.

Finally, the other part of my driving analogy that you missed (or glossed over), is that making a mistake whilst learning to drive can easily turn out to be fatal. Mistakes in programming rarely are.

They may not be fatal to the programmer, but programming mistakes can definitely impact your company, or your paycheck in a negative way. I don't want to sound too dire about it, but at least your mistake wasn't fatal is no help if you are fired for a project that failed.

I'm not saying that using threads would get someone fired. I just feel that it's important to take off the rose-colored glasses before making (potentially) big changes to something you are working on.

If the problem can be help by threads and the programmer is aware of places where she should watch her step, I am all for threads.

As for the comment about inexperience with iThreads, I'm guilty as charged. However, my experience with several threading systems in different systems and languages, with multi-processing in multiple OSes, and home-built cooperative multi-tasking systems, leads me to believe that it is easy to underestimate the difference in approach.

I would never suggest that I know more about this topic than you do, but caution is sometimes warranted.