I think that period should be a comma.
There are two aspects to consider in the errors: how much they change the point of a paper (how they affect a final result), and how much they indicate that the person making the error is a nincompoop, unreliable and unworthy to remain in the debate.
I think either error are understandable in terms of the vast amount of computation occuring and that neither participant needs to hide his head in shame and refrain from battling on the areas of content. I have a hard time thinking that men who can use words like eigenvector and know what it means are unaware of the difference between grads and radians or sqrt versus not sqrt.
Even an error that radically shifts the final answer does not mean that the person making it should stop engaging in work in the field. It could just be luck to make an error in an area where the effect is large versus one where it is small (might be an identical sort of error even). Although of course the effect is more unfortunate. And we can rationally criicise the inability to "second check the work for reasonableness".
I don't think that Mc's error on the radians was in the nature of not knowing the difference. I don't think it is even as bad as making such a mistake when working on a calculator (cause there you have that little symbol in the display telling you the mode). ACtually I can see how it would be pretty easy mistake to make if the wording of the input asked for lattitude rather than "angle". Who thinks of lattitude in angle? Is the source data commonly kept in degrees (yes). And why would one expect a manual conversion outside the program? Anyway...it's not that big a deal.
Can you talk to Steve? His spambot ate a bunch more of my posts...either that or I'll plague you...