On Sep 28, 4:02 am, maldun <[email protected]> wrote: > Ok first I would thank everyone for their thoughts. I think the best > is to do a design document, or at least start a ticket where we come > to a common agreement how to handle the input because there are > several possiblities to do that. And it is important that we are more > or less happy about the solution. > I'm a little bit short on time in the next weeks, but I hope to get > back into it soon =) > > @rjf Of course from a mathematical point of view you are correct, but > from a more pragmatic point of view you have to consider that about > 80%-90% of the input satisfy these assumptions. So one can prevent > about 80%-90% of the mistakes. For the others it is important to give > the user a good control about the choice of the method.
Huh? All you have to do is say that the results of your integration program may be in error by an unknown amount, instead of saying "rigorous estimate". Would you be happy if I gave you a cosine routine that gave you a correct answer to the last bit except if the input was too close to a rational multiple of pi, in which case the error could be anything? Would you argue that a rigorous estimate was "correct to the last bit"? > It remembers me somehow of the arguments people used against > seatbelts, because there are SOME cases where the seatbelt could kill > you even when it saves your life in 80% of the cases... Yes, by this analogy you are saying that there is no need for a seatbelt because most trips in a car do not involve a crash. > > >For example, you could solve World Hunger if you assume > >(a) there is enough free food for everyone and (b) there is > >enough free transportation to distribute it. > > Ever watched "We feed the world"? Just curious :) not that I recall. -- To post to this group, send an email to [email protected] To unsubscribe from this group, send an email to [email protected] For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
