I had one of my occasional moments of getting annoyed and deleting things the other day, though this time it was focussed on my first couple of attempts at novels, which were frankly rather awful, and were never going to see the light of day. Not least because I had already cannibalised one of them for parts.
Now, I imagine that there are some people who would be horrified by the very thought of that. Rather like those people who horde every piece of paper until their house looks like the inside of a recycling centre (minus the conveyor belts, large vats, etc), they just can't bring themselves to get rid of old pieces of work, in case they should turn out to be useful later on. It's like little old ladies and pieces of string, or posties and rubber bands (which do, admittedly, let you create the bounciest rubber balls in the known universe; great for driving cats insane).
But clearly, there is a case for getting rid of work that is simply unsalvageable, even if your computer hard-drive could conceivably take the strain. It clutters the place up, your occasional wanderings back to it distract attention from what you're really doing, and more importantly it is never going to contribute meaningfully to a worthwhile piece of work.
That strikes me as the important point when deciding to get rid of old work, though it's something I have to be careful with. In some phases, I can decide that nothing I've written is ever going to be any good. That is not a good point to start hitting the delete key, as I have found out once or twice to my cost. As I have probably mentioned before, 28000 words seems to be a barrier of sorts for that with me, presumably because it's the point where things inevitably require some extra work around the middle.
The trick seems to be spotting the point where a piece has nothing more to offer. That's easier said than done. Take a post a few places below this, about changing styles in writing, which I wouldn't generally have thought of as a great source of ideas. It happens to contain the line, "Now unless we're saying that all you need to do to vastly improve a story is introduce an eight foot, hairy bloke in a brown overcoat, I'm not sure that this is what made the difference." It has suddenly occurred to me, skimming over it, that there might be a short story (or five) in that somewhere.
This is where I think it becomes important to understand what it's useful to save from failed work. It isn't generally the body of the work. If you can cut and paste pieces wholesale, then what you're probably dealing with is a very different draft of the same piece rather than its death, as I'm doing at the moment with my alternate versions of the sequel (I think I prefer this one from the point of view of series continuity, but I've got to make it work first). What you can take from old pieces are story ideas, characters (who will no doubt appreciate the change of scenery), scenery... or rather settings, especially wonderful lines or jokes, that sort of thing. Individual pieces, in short.
But the best way of doing this doesn't seem to me to be keeping the original story. Doing that means that any piece you try to use clings on to other bits, octopus like (which might be useful if you need some jars opening, because apparently they can be taught to do that, but not otherwise). What you need, more than a single piece tying up all these ideas, is to move them back into little files of characters, premises, settings, etc, where they can sit happily twiddling their thumbs (inasmuch as settings have thumbs) until you can find somewhere else to use them.
In short, don't hang on to things that don't work for the sake of it. It might be the first thing you wrote, but I promise you, you'll write better. The very fact you can't make it work proves that. Instead, take it, squeeze it dry of anything useful, and then let your finger drift gently unto the delete key. Just not quite as often as I do.