I won't say "duh, pointless question", since I understand that this has been a matter of discussion in the S&I community pretty much since the game was released, so let me put it this way: The object of the game, as far as I remember (haven't played it for more than a year, will give it a try again as soon as I get back to Denmark), is capturing your opponent's scientists. Scientists produce research that gives you better weapons and various other goodies(?). I don't remember the winning conditions, but I think it's directly related to either scientists as well, or fighting (and scientists, or rather the advanced weapons they give you, are pretty instrumental to fighting well, correct?). In short, the side with many scientists wins, correct? So you steal your opponent's scientist. Why should killing them be wrong if stealing them is fine? Scientists will respawn after a time, so it's not like a single attacker is completely crippling you. Furthermore, iirc, your scientists will become "frightened" and work slower when you start killing enemy scientists, so there's a bad side to this deal as well; you're not depriving the enemy of his scientists for free. And besides, if you can kill the opponent's scientists, doesn't that mean you're in the lab? Unless I'm much mistaken, the enemy is supposed to protect that room, no? So they ain't doin' their job. Their bad. Their fault if something happens to their scientists I say.
Don't get me wrong: I wasn't a scientist killer back then, I ain't one now. It's honor-based (dang non-combatants. Arm yourselves so I can kill you, damn scientists!!). But I consider killing scientists a valid tactic: The game allows killing scientists right out of the box, so it certainly isn't an exploit. It has mechanics in place to avoid rampant scientist killing, so it's not something the Developers didn't think of. Heck, they could've made it impossible to kill the scientists, couldn't they?
Instead is killing the scientist influenced by rigid ratio of your own code of reputation: Because those are not killed you yourself, you must have reputation and grasshopper. As for that how or to you who how are decided in order to meet to cowardice of courage and justice it is in order to stoop in level. When you worry victory, with the return fill up the lake with the blood of the murder and the scientist. But if your center purity, your mind is large, from the flow your center of your reputation freely, like the snake and rat it is not the imitation, those of their strike shames are not copied and, to also their dishonorable maneuvers to lower with your purity those in order in spite.
Translation: Whether killing scientists in return depends on the rigidity of your own code of honor: Since you don't kill them yourself, you must have some honor, grasshopper. It is up to you to decide whether to stoop to their level, or whether to meet cowardice with valor and righteousness. If you care about winning, kill in return, and fill the lake with scientist blood. But if your heart is pure, your spirit is great, and your honor flows freely from your heart and is not fake like the serpent and the rat, you will not copy them in their shame, and you will strike them down in your purity regardless of their dishonorable tactics.
I'm not aware of any house rules on this topic, so if those exist, forgive my ignorance. If we don't have any, I'm willing to discuss the topic. Again, I haven't played it recently, so I don't know how much of a nuisance scientist killing is (blah, what is it with all these non-explosive scientists nowadays? Scientists should go "Freeman!" and "boom", in that order. Harumph.). If it runs rampant, I suggest a plugin: Kill one scientist and receive a warning, kill two scientists and receive a slap, three strikes and you're out (tempban). I'm strictly opposed to a complete ban for doing it, though.
NB: This entire reply is based on the assumption that it was an opponent who killed your scis. Killing your own team's scis is, of course, utter laming and a bannable offense. Nothing different from teamkilling, really.