But when did I say opening the source should be mandatory?

I'm just stating a simple truth: For the customer, a closed-source program is a product of lower value than identical in all other respects open-source alternative - strictly because of levels of maintainablity and control from the customer side. If you keep your source to yourself, you may be securing your profits, but simultaneously you release a product less competetive and less valuable to the customer. It's your choice and your right to do so. Just don't be surprised the customer isn't quite satisfied with your product... finding it unmaintainable.

As for me releasing my code, I do so on regular basis, the only criteria is general quality and usefulness of the code (I try to make it maintainable, readable and at least somewhat foolproof before releasing), and chances are you will stumble upon pieces of my code sometime when you need it. Probably you won't associate it with my name, because I often release code anonymously, to protect myself from lawsuits or hate. You see, when the customer can download my program which uses plain old RS232 level converter cable for connecting with a device, instead of software provided by the manufacturer, which works only with "authentic, brand, special" 30 Euro cables, enhanced so the program could identify them as such (and besides that being just plain RS232 level converters), I'm definitely costing the manufacturer of the "special cable" some profit, while saving them for the customers - and myself. But now how do I prove I wrote it all myself, and not stole their sources? See SCO vs IBM case. Now I can't afford $1mln/month for lawyers so I prefer to just let the code float around the web without author's name attached to it.