Good/bad does not apply. Violence is more fundamental than either. It is a property of the universe that precedes humans and even life itself. Calling it good or bad is like calling the law of gravity good or bad. Violence is a necessary consequence of unavoidable ignorance on the part of systems that lack infinite wisdom. You cannot avoid it, or enlighten yourself out of it.

I'll give you a philosophical answer first and a human answer second.

Philosophical Answer

If you like Aha! moments, try to guess where I am going based on this picture, known as Spreng's triangle.

To quote myself from an old blog post about this triangle:

The idea there is that to accomplish any given task, you can use a mix
of three “pure” capabilities: unlimited knowledge and no time or energy,
unlimited time, but zero knowledge or energy, or unlimited energy and
zero time and knowledge. You can also have various mixes of the three.
The pure cases are, approximately, the archetypes of muscular caveman,
great philosopher and immortal lazy idiot.

Hint: Non-violence is effectively impossible for roughly the same reason processes that cause no increase in entropy are effectively impossible. It defies the laws of physics. You'd need an infinite amount of time to expand a gas isentropically (with no increase in entropy). To reverse entropy, you need shaky constructs like Maxwell's demon (one interpretation of the demon is that it is infinitely knowledgeable: http://en.wikipedia.org/wiki/Max...).

I won't get into the details, but think of the general philosophical idea as the following: violence is a function of the energy you put into an action that compensates for your lack of perfect knowledge/infinite time.

This is somewhat speculative and mystical, and I don't have a proof (yet), but I believe it to be true.

The closest humans can get to perfect non-violence is to starve
themselves peacefully to death while meditating. Even this is not
perfect because it involves violence to your own body. Orthodox Jains,
among other meditative traditions, actually recommend this path of
meditative suicide. There are those who define meditation as the art of
conscious (i.e. mindful) dying.

You can also consider creative-destruction in the Nietzsche/Schumpeter sense as a necessarily violent dialectic that gets (hopefully) more knowledgeable and less violent over time.

If that's too philosophical for you, let's do the more "human" answer. Unfortunately the human answer is necessarily rather long.

The Human answer

The problem of cruelty in nature

To me, lions killing gazelles is sufficient proof that violence is neither good nor bad. If you force fit notions of good or bad on violence, you will logically get to absurd extremes such as "lions shall lie down with the lambs" and uncomfortable conclusions like "we should build robotic tofu-based deer for lions to chase."

If you want an even more extreme case, a volcano can kill you in horribly violent and painful ways, but obviously a volcano is not a sentient entity that can be either good or bad (but there's some subtlety even in the volcano case that I won't get into...).

The restricted human case

Okay, so perhaps only human violence is under consideration, since we possess enough of a mix of sentience, empathy, omnivorous dietary systems and intelligence to even consider the question? This is a common position. Traditional (not modern) law in fact adopts this position. Manslaughter and murder are both violence. The difference is whether or not there is intent.

The question of intent

John Steinback's Of Mice and Men, is a startling exploration of the relationship between intent and violence.

http://en.wikipedia.org/wiki/Of_...

In it the mentally disabled but immensely strong Lennie accidentally kills a puppy and later, in fright, a woman. Lennie's cynical friend George later kills him as painlessly as possible, to save him from a lynch mob.

The 3 acts of violence in the novel are born out of mental retardation, fear and compassion respectively.

So can you conclude that intent is key, and that an intention to be violent makes violence wrong? In that case, we can absolve both Lennie and George. And we can tentatively call the actions of an intentionally cruel person "bad."

From Intent to Utilitarianism

But intent falls apart with a little poking. What if the only way to serve some greater good in a utilitarian sense is to employ violence, and the person employed to administer that violence happens to enjoy cruelty, but uses minimal cruelty under the direction of the utilitarian judge directing the proceedings? Is it bad merely because the instrument of (minimum necessary) violence happens to enjoy his job?

Agatha Christie's And Then There Were None gets into this question, but I won't say more, since that would involve spoilers.

This conundrum forces you to back away from the subjective experience of being cruel/violent as a reliable guide to anything. You can only work with the processes and outcomes of violence. People's experience of administering violence is just too tricky.

At small and large scales, intended violence can (and often has) been justified on utilitarian grounds (for instance torture one enemy spy to save an entire "good" country from an "evil" one).

Once you allow for this sort of moral relativism as a premise for utilitarian analysis, good and bad become useless as a frame of reference. Violence cannot be absolutely good or bad if there is no absolute good and bad to reference.

So having abandoned intent, we arrive at the classic post-utilitarian philosophical question about violence: do the ends justify the means? It's an age-old question, but only recently have codified institutional practices emerged around one answer: No.

Post-Utilitarianism: From Bentham to Rawls

Note that this is not an application of utilitarianism. This is about questioning the fundamentals of utilitarian philosophy, upon which intent-based judgments of violence are based.

In law, this was the transition from the classical utilitarianism of Bentham to what you might call the modified post-utilitarianism of Rawls.

http://en.wikipedia.org/wiki/Joh...

The general conclusion of modern law (not traditional) is that the ends do not justify the means. (Here I am playing fast and loose and ascribing certain general views to a "Rawlsian style of reasoning" rather than to Rawls himself).

So modern nations voluntarily attempt to make capital punishment humane instead of painful, and forgo the benefits of making a violent example of one criminal as a deterrent to others. Modern nations also, at least on paper, unilaterally limit their use of torture in war or criminal interrogation and make up stuff like the Geneva convention, and express appropriate horror at things like the Hutu/Tutsi conflict.

Note though that this is not a reversion to an absolute notion of "good/bad" morality. The decision to avoid violence rests more on the practical notions of empathy and pain than on abstract and absolute notions of good and bad.

But ultimately, I do not find this an interesting debate, because you can only resolve it in cases like capital punishment, where humane capital punishment is an option. You're still left with higher levels of irreducible violence in situations like war and cops chasing criminals in a live fire situation etc. The means-vs.-ends debate ends up being about arbitrary thresholds of justifiable violence and tradeoffs between effectiveness and violence over the long term. It is only applicable when you actually have a choice about how much violence to use.

More importantly, grand moral positions about whether it is okay to torture one person to save all of humanity have never really been tested in extreme ways (Sophie's choice is an interesting read here).

Most violence is not chosen under such conditions of extreme choice. If you loosen conditions, Rawlsian views work less well.

The Gandhi Sidebar

So Rawlsian "ends do not justify the means" is a practical solution within
limits, that breaks down into absurdity if you push it too far, as
Gandhi did. As an iconified symbol of practical non-violence, Gandhi is a good
meme. As a real model to follow, he is not. His life was full of
hypocrisies and disingenuities, which is why he is not quite as revered
among Indians as he is among non-Indians. I am among those who do not
admire Gandhi much. This recent Atlantic article gives you an idea why:
http://www.theatlantic.com/magaz...
So to clarify, I think the ends do justify the means. An Indian
text that is much less known globally, the Bhagavad Gita, actually
adopts this position (and is controversial as a result).

So intent and utilitarian analysis and means-ends justification debates are ultimately useless. You cannot resolve the question of violence that easily within pragmatic social-morality frameworks. You have to dig deeper into the true nature of violence.

So you have to ask, when and why is violence a necessary condition for action?

The Necessary Conditions for Violence

The Steinback example provides a hint: violence is usually required when there is ignorance. With perfect knowledge and omniscient foresight and planning, in principle, you might not require violence. This is the Sun Tzu ideal of winning all your battles without a fight.

This even makes us rethink our opinion of people who enjoy administering violence. Perhaps they are only ignorant? Perhaps they are sick in some way, lacking empathy, and therefore more like volcanoes than humans? Perhaps we can only justify putting them away for the safety of others, just as you might evacuate the region around a volcano (effectively "jailing" it within an uninhabited zone)?

Is Sun Tzu right? Can all battles be won without a fight if you're smart enough?

It gets even more complex: there are Zen ideas about how the perfect warrior is really one with his enemy, and the mindful sword fight that ends in the death of one is not violence at all, but a sort of enlightened destruction of an aspect of oneself.

Is this sort of Zen philosophizing correct? Can you always avoid violence or reframe it into meaninglessness if you are enlightened enough?
Here I part ways with the wise monks. I don't believe so. My reasons have to do with my philosophical answer based on ideas like entropy. Perhaps the universe as a whole can get more enlightened and less violent, but not individual sentient decision-makers within it. They do not have infinite time or wisdom. They are not thermodynamically closed systems. They are plugged into larger systems of ongoing natural violence. They will necessarily be violent. If it goes away at all, violence will only disappear long after human beings are destroyed.
The Universe is an irreducibly violent place. We are part of the universe. There is irreducible violence within us. At a deep level, that's why we must die.