What do we want to achieve with the translation of a Sanskrit term? Who are our target-readers? This seems to me to be the chief question while deciding about a translation. In fact, in many cases a translation ends up being understandable only to specialists.
A good example is, in my opinion, the translation of vyāpti, the invariable concomitance holding between the elements of an inference, e.g., between fire and smoke.
Vyāpti is a nomen actionis from the root vyāp-, which literally means 'to pervade'. Hence, many (most, I would say) authors translate vyāpti with 'pervasion'. This has almost become a terminus technicus in the works dedicated to Sanskrit logic. But is it a good choice? Is not it only understandable by an elite, which hence runs the risk to appear non-interested in communicating with any scholar outside itself?
To elaborate, 'pervasion' seems to me to be not-understandable for non-Sanskritists. It does not correspond to any logical term (as, instead, probans for hetu) in Western logic, nor is it intuitively understandable (as 'invariable concomitance' for vyāpti, which at least describes what is at stake). 'Pervasion' is just a literal translation of the Sanskrit term, which tries to reproduce the metaphor in English. Personally, I (and I suspect many others) only understand it, because I automatically translate it back into Sanskrit.
However, against 'invariable concomitance' Michael Williams (Manchester) made me aware of the fact that it does not point out that a vyāpti is not necessarily a commutable relationship. For instance, wherever there is smoke, there is necessarily fire, but it is not the case that wherever there is fire there is smoke. Indeed, according to the Ancient Indian Physics, there is fire in a piece of melting iron, though there is no smoke accompanying it.
Hence, one could use a paraphrase, such as "Smoke is invariably concomitant with fire", thus implying that the opposite is not necessarily the case.
Moreover, 'pervasion' is delusory also as for the 'direction' of the relationship. In the standard example, the point is that there is no smoke without fire. But, if one says that "Fire pervades smoke", does the listener understand that fire is a larger set than fire? Or does not s/he imagine fire 'permeating' (i.e., becoming diffused within) smoke, thus implying that smoke is a smaller set?