It is hard to not be insulted by your suggestions that I don't understand the topics, that I am unwilling to accept new ideas, and that I need to read more books. I have said it before on this list, and I will say it to you: I would suggest that you do not jump to conclusions about my (or anyone else's) lack of knowledge on a subject. I am far more familiar with generic programming than you probably realize. This kind of innuendo accomplishes little, except aggravation. It also makes the Boost mailing list a very inhospitable place.
Hi, Cristopher I do believe this discussion became too ... personal. But IMO you are partially responcible for that either. If you bring an idea that kind of contradict to common accepted rules you should expect people to attack it from all possible angles. More specifically: you presented an idea to use raw function pointer in places where we now used to use different means. There is no particular reason why it wouldn't work in some specific cases, but *our* experience tells us not to do it. In long term you lose more then you win. Now we (and by 'we' and 'our' I mean people that disagree that usage of raw function pointers has an advantages) couldn't share our experience though mailing list, so we are trying to communicate it though references to known C++ experts. It's not the best way I admit, but this is not exact math either where I could present you with formal proof. As for " why not provide that variant either?", my position is that we (boost community) shouldn't promote solutions that in our opinion deemed to fail this way or another at some point, especially if there is generally accepted alternative. Regards, Gennadiy