while i agree with you (and the importance of gould's point), at least for reasons of parsimony, it seems that many biologists do seek to understand current behaviours and traits in terms of selection pressures. the thing with altruism is: a) it is not just some odd behavioural trait but directly relates to fitness, and b) it *decreases* the fitness of an individual. point (b) is IIRC wilson/sober's definition of altruism: behaviour of an individual that decreases its fitness relative to another individual or group.
given the significance of altruism as a trait, i think it is important to demonstrate either that a) it did not evolve in response to selection pressures (or is a side effect of other traits that did) or b) it can be reconciled with the model (perhaps genotype propagation) in some way: wilson/sober, for instance, use a group selection model to demonstrate the feasibility of altruistic traits and their continued occurence.
of course we have to be wary of the reductionist tendency of defining away altruism.
--ravi