Composer Derek Bermel has a great post about making revisions:
About ten years ago, in Den Haag, I was chatting about the thorny process of revision with my friend Peter Adriaansz, a fellow composer whom I hold in high regard. "I am a chronic reviser" he said. "It's my curse; I'm never satisfied with a piece. I rethink and rewrite until I'm absolutely satisfied. It can take years. And some of my pieces I just won't release again until I make all the necessary revisions."
I thought Peter overly dramatic. "Why don't you just write a new piece, with these insights in mind?" I asked. I showed him an orchestral score I had written recently; the work had already been performed twice, and I still wasn't entirely happy with the last section. However I had decided to leave it unrevised, as a document of my compositional mindset at the time; I explained to him my feeling that returning to that piece and reinterpreting it within my current aesthetic would be anachronistic and untrue to the original conception.
Peter smiled. "I suppose you and I are just different kinds of composers", he murmured wistfully.
Around the time of my encounter with Peter, I was writing a piano piece, which I called Turning. One reason I gave the piece that title was because I could sense my compositional process beginning to shift; I had determined that I was most satisfied as a chronic reviser. These days, to the chagrin of my publisher, I tend to revise after virtually every performance. I recall my encounter with Peter and I find it hard to identify with the composer I was then.
We live in an era of marketing makeovers, in which politicians deny their mistakes, change original rationales to suit the polls of the moment, and take credit for events and trends that have nothing to do with their own policies. If politics is rooted in appearances, perhaps art (with a small ‘a’, just to be safe), is the other side of the coin: truth-telling. Such truth-telling must of necessity start with oneself, and painful questions follow: Why write this? Is this interesting? Does it go on too long? Not long enough? Is it clear? Is it muddled? Is it pretentious? Simplistic? Someone else can – and probably will – answer those questions for us, but we only become good composers when we answer them ourselves, and then make appropriate changes.
I'll cop to being a chronic reviser as well. In the age of Finale, it's almost impossible to resist the temptation to tweak, and to be honest, I find the act of revising a lot more enjoyable than the act of composition. It's always more grueling and anxiety-provoking to trudge out into blank staff paper (or, um, blank screens) then it is to tighten and sharpen a piece that I'm confident basically works already. If I had the time, I would probably revise every chart in the Secret Society book before every performance. Of course, some of the changes would be one-time adjustments to accommodate the strengths and weaknesses of the specific players involved, or just to mix things up according my personal whims at the moment. But by necessity, the more substantial revisions are focused around newly written material.
If you compare the first three versions of "Phobos" -- from November, January, and April -- you will hear all kinds of variations, from the absence of cajón on the first version (which now, with the benefit of hindsight, feels so very, very wrong), the length of the introduction, the orchestration of the melody, the changes and duration of the solo section, and the rhythm of the concerted ensemble passages following the tenor solo. Next time you hear it, there will be still more changes -- all of the for the better, I hope. I flatter myself that all this revising isn't the result of chronic indecision or directionless fiddling -- I'm just trying to get all of my pieces to the point where I feel comfortable abandoning them.