You need much, MUCH more information, to produce a good questionaire. (eg: Is someone using the stable GNOME? The unstable one? The Ximian/Helix variant? Something they've rolled themselves?)
By using broad, sweeping generalizations, you can get the general trend, yes. But you've no means of finding out the fine detail.
You also need to validate your answers better. 1901 seems, well, a little unlikely. 1999 is definitely out, unless someone cloned Mozart. (But even he didn't start producing masterpieces until the age of 9.)
Many "formal" surveys use questions that overlap, and then verify for correctness by eliminating those overlaps which don't agree. That tends to make for a big questionnaire, though. (You'd be looking at 5 times the size, for accurate results.)
In lieu of such rigorous cross-checking, some bounds-checking would be useful. I strongly believe that =active= participation in 99 projects is a claim someone would need to prove, before I'd believe it.
One idea would be a text box, asking for a list of all the projects the person is involved with. If the sizeof(list) != entered(number), then one of those two entries is invalid.
Also, if a person states they are unemployed, they can't then be earning a salary. And the list of basic cross-checks goes on.
Otherwise, this is a great idea!