[spoondan]

On the other hand, I'd much rather a client pitch a great idea that's technologically infeasible than just come up with something spur of the moment that they deem possible. If you start the conversation at, "This mechanic may not be possible, but the rest of the game is great," then you can discuss how to change the mechanic or maybe an engineer will have some brilliant insight about how to do the impossible. At the very least, you're starting with your best and working towards the technological compromise rather than starting with something that's already a compromise and may not reflect your best thinking.

[SoftwareMaven]

Any pitch that starts with "this may not be possible" (even if that is inferred) is going to quickly get turned off by anybody who listens to pitches all the time.

This is why engineers generally don't make good sales people: the desire to tell the whole truth is usually too strong. The world of software is an incredibly dynamic space, so sales and marketing can use a little of that to their advantage. :)

[MartinCron]

And this is why we can't have nice things :)

It's impossible to know with 100% accuracy if a great idea will be possible or not. This is one of the reasons why I like the Lean Startup stuff so much, where you can be honest (with yourself, at least) about doing just enough to prove if something is possible/marketable/whatever before investing years and millions into it.

[scott_s]

Games are built around mechanics. If your mechanic isn't possible, then the rest of the game doesn't matter.

[ctdonath]

Or, as the author notes, sometimes you just realize the idea isn't going to work, isn't worth trying to retrofit into a sub-par result, and you're about to lose a rare opportunity to talk to the make-it-happen person. He came up with a better idea on the spot, and it worked.