Dear old Clive James - never blocked for long - has revealed to us something about his way of writing: "I do the first draft in longhand because I want to see what I've crossed out - easily; because sometimes the bit you've crossed out is the bit you should put back."

This applies to the act of composition too does it not? Is it then an argument against the use of Sibelius or Finale and in favour of music-paper? Or does a really first-rate composer do it - or have it - all complete in his head anyway before putting pen to paper or finger to keyboard?

I wonder whether Mr Grew's analogy between writing words and writing music is actually quite as obvious as he suggests: from my own (admittedly limited) experience of writing music, it's a far longer process than writing prose or poetry, and what is being marshalled is far more complex, both in terms of recording accurately exactly how the material is intended to be played, and (even more complex) how its structure fits together: even allowing for idiosyncratic transliterations, the words one places on a page are by and large an exact representation of what one wishes to say, whereas finding as accurate an interpretation in notes of what is intended musically is often far more difficult to arrive at.

There are further complications, too, concerning whether a composer works at an instrument or directly onto manuscript (or varying permutations thereof), or indeed whether score-writing software is used throughout the whole process or only in the later stages for producing a fair copy and parts, and indeed even which method of inputting the information is chosen. Whether manuscript or software be used, there are so many more combinations which may be used to arrive at the final representation of the composer's conception, and the journey there itself is generally so much more complicated, that it may be misleading to try and link the process directly to writing words. In any case, with Sibelius at least, previous drafts are saved automatically for a while, so that even if one doesn't start each session working on a new saved copy of the previous work - strongly advisable in case of computer malfunction - it is usually possible to make comparisons between earlier and later versions of the same passage: I am not aware whether Finale works in the same way, but I would still expect that most users would retain previous drafts whilst working on a score.   

The difficulty is that when writing words, the text is linear, straightforward. A line of words can be crossed out without confusing the text too much (and one of my typing exams many years ago required transcription of a very confused piece of writing, sometimes with several crossings out in the same area and new words written in a smaller and smaller hand using every bit of available space). But if a composer starts crossing out individual notes, or worse still, chords, to write in others the score can very rapidly become desperately muddled in the instance of anything more complex than a very simple melody.

A bit like reading my thoughts, I suppose.

eee! I remember when it were "xxx.fred" "xxx.jim" and "xxx.sheila"

Being myself not a composer but a performer and very occasional musicologist I can't see any argument in favour of doing the actual composition (as opposed to copying) in a software program unless what you want to do corresponds closely to what the program allows for...

I often find myself wondering where a note came from, so to speak. Often with archived sketch materials of various formats (and I don't at all mean previous drafts saved in a program, but sketched marginal fragments, reminder notes to self to do or undo something, and yes, crossed-out versions) that sort of thing leaves a trace - and every aspect of it is capable of leaving some kind of information when for example the pen or ink are different or it's clearly a second thought in some other way; the nature of the handwriting can show if something was dashed off or considered, for example. This for me is important musical and human information.

I can't see that sort of information being left for posterity in a computer version. I know that many composers don't see that as a bad thing but would humbly suggest it's not only for the composers to say - this is for me the sort of information that sheds light on the process that gave rise to the music as opposed to the product which results. That's very much how I like to approach music.

As far as sketching goes, regardless of what's most useful to the composer, pen(cil) and paper are infinitely more useful for the rest of the world, I'd say...

I wasn't speaking solely as a very occasional musicologist in fact but just as much if not more so as a performer who likes to know what a note is doing there. Or more to the point: if a note has a documentable history then I would very much like to have the possibility of knowing about it because as far as I'm concerned that has the potential to lead to a deeper performance. Quaint or not...

To take a concrete example: when I perform the Ferneyhough Time and Motion Study I, what comes out is very much influenced by what I've seen of the sketch materials, not just the kind of thing that might have ended up in a Finale or Sibelius draft version but the snippets on the back of envelopes, the hints to self on the back of envelopes and such... there's a kind of focussing of my attention on the salient stuff that that leads to.

Hmmm. For me the key issue is (as Richard mentioned) the transparency of the software. No matter what software you use to typeset, it locks you into a certain way of operating it. Given the extra level of abstraction necessitated by the interface, software tends to lend itself very readily to some typographical tasks, and very poorly to others. Also as a result of this abstraction (composer uses mouse and keyboard to tell software to employ complicated MIDI/graphic algorithms x and y, so that the result will look like z) I don't think that software is particularly useful to actually 'compose' music in, unless your music is either very simple, or the composer knows exactly what the music is before entering it. Pen/pencil and paper, on the other hand, makes it very easy to alter things which, for most composers, is pretty necessary.

That said, there are a large number of Australian composers of the younger generation who write (and, I think, audibly so) directly into Finale. They seem to be doing rather well for themselves in a local sense.

I don't think that there's any real evidence that this is the case, although perhaps the only way to be sure would be to lock each of them in independent studios with only manuscript paper and pens, pencils and other such antediluvian writing instruments (did I mention litre bottles of correction fluid? - no, of course I didn't!) and see whether any palpable changes in the "styles" of any of them might occur after a time (and it occurs tio me that such an experiment might even be conducted formally under an arrangement that could be known as a draughtsman's contract, but let's not get into that one)...

Best,

Alistair