Your assertion that we may never know whether anyone was murdered was a surprise to me, which is why I responded. It seems relevant because whether the good outweighs the bad depends on how much bad was actually done. (Technically what's relevant is how much bad one would expect to have happened in advance, but what actually ended up happening last time will affect our guesstimates of what is likely to happen next time.)
Don't know we also need to know how much good was actually done? Can that be quantified? (Can we name somebody who wasn't murdered because of the leaks?)
Asking how to quantify the good and bad that resulted from a thing is something we can ask about literally anything. It's especially hard in this case, because the potential harm was immediate and personal (people being murdered, etc) while the good was broad long-term (a public with greater insight into what its government is doing, etc). But there are plenty of areas with similar tradeoffs; if raising a speed limit lowers commutes for millions of people but raises fatalities by some amount, then that's equally hard to compare.
