Radical? It's only logical.

Things can happen in three ways:
* Without my having any conscious control over them;
* As an intended effect of something I consciously decide to do;
* Or as an unintended effect of something I consciously decide to do.

Of these, the first category can clearly not be reasonably blamed on me
(e.g. a mudslide in Peking or whatever) and the second does not contain
any seriously harmful actions (I am not a criminal; I don't
intentionally do serious harm!)

That leaves the side-effects. Of those, there are two kinds:
* Reasonably foreseeable based on the knowledge I have at the time,
which contains (but is not limited to) a basic Western education, Java
knowledge (relevant specifically in the context of this newsgroup), and
anything self-evident to a person of normal intelligence (or indeed of
substantially above-normal intelligence...)
* The rest.

Call these, respectively, "foreseeable" and "unforeseeable" for
brevity. It is clear that the latter also cannot be blamed on me. For
example, if I flap my arms and it causes a deadly hurricane in Taiwan
six months later, the hurricane is not my fault -- it was just as
likely that my *not* flapping my arms would have that consequence.
Chaos theory effectively means the causal connection is encrypted and
thus unavailable to foresight with the level of modeling technology
currently available (to wit, human brains and
thus-far-fairly-rudimentary artificial computers). Which means these
consequences are just as unmanageable as truly spontaneous natural
disasters or whatever, at least for the foreseeable future (about 20
years, give or take). Blaming someone for the hurricane would make no
sense, because nobody (yet) has the means to affect that sort of long
range chaotic effect in any way that doesn't manifest as pure noise.

The foreseeable but unintended consequences are left. Of course, if
those include a reasonably likely significant harm to someone, then the
action is irresponsible and I won't take it. The rest are harmless.

There's some added complexity regarding lesser-of-evils situations,
including punitive or retaliative action that may cause harm but
prevent other harm, but the above argumentation is the nuts and bolts
of it.

To the extent that a person is capable of reasonable knowledge and
foresight, lacks purely malicious intent, and avoids acting
irresponsibly (basically "is of sound mind and judgement" and "is not
criminal or negligent") they should not be considered guilty of
anything, ever. (Me, or anyone else in the above category.)

Of course, good intentions alone don't guarantee a good outcome; that
is why bad but foreseeable side effects are important to consider, and
why it's necessary to be rational and ground decisions in empiricism.
Killing the planet's whole population to end suffering kind of misses
the point, for example, as do other classical "good intentions, great
evil" scenarios. (Hitler's plans, for one. A case can be made that he
meant well, but obviously he wasn't of anything remotely resembling
sound judgment, and pursued irrational, non-empirically-based ideals
with blood. Of course a case can be made that he was simply a greedy
and murderous bastard, too.) Actually the most common source of evil of
the "good intentions" variety in fact and in fiction seems to be having
some Brilliant Idea(tm) and then imposing it by fiat. I weight any
abrogation of someone's self-determination (including abrogating their
choice not to die just yet, i.e. murder) extremely negatively versus
anything else besides protecting someone *else*'s self-determination.
This leads to a calculation whereby any kind of system that doesn't
maximize freedom for most people is clearly and foreseeably bad.
Ideally, self-determination is limited *solely* to the extent that
someone's determination would stuff up someone else's freedom of
choice. In practise, it's also necessary sometimes to enforce arbitrary
conventions to break symmetry deterministically, e.g. permit travel
only on the left (or only on the right) side of a contiguous road
system, for reasons of safety. (Actually, this isn't really an
exception; the choice to drive on either side is a minor thing to lose
compared to the choice to not be in a serious crash. Preventing a
substantial number of crashes at the cost of some enforced
symmetry-breaking without any other real consequences is clearly
justifiable. In fact, *failing* to use such a simple fix, given that
problem, is unconscionable.)

On the other hand, given how people behave (irrational and downright
nasty) much of the time, it isn't beyond my imagining that a lot of
people do actually hold others responsible routinely for things they
had no chance to influence even with reasonable foresight and good
intentions subject to the overriding goal of preserving widespread
self-determination. It seems that blaming people for accidental
collisions that they took all reasonable measures to avoid is
commonplace, for example, as is blaming people for not knowing things
that they never had any reason to find out -- no-one told them and it
had no obvious importance to them, or perhaps no-one told them and they
never even knew it existed to *be* found. I've also seen numerous cases
of people being faulted for everything from their genetics(!) to
unforeseeable events (an involuntary twitch of their body that damages
something -- it might as well have been a meteor impact for all it was
under their control!) and so on...