says nati linial and he is probably right. the math learning theory conference was a place of proof that after all the hickhack in the 60s, 70s, computer science and mathematicians are on the way to find a common path again. learning theory serves as uniting force, some say; but actually i am not sure how many of those theoretical guys really like to give up their one and only truth - something a mathematical formular stands for - in favor of heuristic, computer machine learning theory implies. but at least in math theory people have learned to play around with the unpredictable, a field applied mathematics are not really good at, imho. still, for some it can turn out to be simply a matter of surviving; a polish group of mathematicians fear that parts of their tradition are in danger, especially the stefan banach space theory from the late 1920s. mathematicians are still used to think in longer periods of time.