Why I am writing this before the program is finished

For twenty years I wrote and arranged on other people's software, and a great deal of that time went into learning the software rather than writing the music — manuals, tutorials, forum threads, a private list of workarounds I had to keep in my head. I got fluent in the programs. That was never what I wanted to be good at.

The hours are not really the cost. The cost is stopping in the middle of writing to go and find out how to input a note, or how to make the program do something I could have done by hand in ten seconds, and coming back four minutes later to an idea that has gone. Nobody ever gets those back, and they are the part that mattered.

Arietto is the program I wanted during those twenty years. This is the record of it being built, and it goes up on Tuesdays.

There will be one of these every week from today until the program ships. Some weeks it will be a long one about how a page gets broken into systems. Some weeks it will be four hundred words saying that I spent five days on a bug in multi-bar rests and have nothing to show you. A thin Tuesday will always go up. A skipped Tuesday will not. That is the only promise in this post that costs me anything, which is why it is the one worth making.

Why write before it is finished

Because the alternative is worse. When it ships, a stranger is going to land on a page belonging to a notation program they have never heard of, made by one person they have never heard of, and be asked to build a working year on it. Nothing I can put on that page then will answer the question they are actually asking, which is: is this person still going to be here in 2031, and can I check?

A run of dated Tuesdays is not an answer either. But it is evidence, and evidence is the only thing that works on this question. A run of dates is checkable. A sentence about commitment is not.

There is a second reason, less noble. Writing this every week forces me to look at the week honestly. It is very easy, alone, to spend four days polishing something nobody asked for. It is harder to spend four days on it knowing you have to write down on Tuesday that you did.

What will be in here

• Tuesday — this. What I built, what broke, what I reversed. Weekly.

Other kinds of post will come when there is something worth showing, not on a timetable.

What I am promising about the program

Three things, and they are all structural rather than aspirational.

One version. Every stave, every part, every engraving control, and the sounds to hear it back. There is no reduced edition with a stave limit. A Blasorchester score is forty-five staves, and a program that falls over at forty-five staves is not finished.

It runs on your machine. No account and no log-in to open your own file. When it is online, Arietto fetches a small signed licence list in the background and never waits for it, so it works just as well offline. Every score exports to MusicXML, MIDI and PDF as part of the program. If you ever want to leave, you leave with everything.

Everything it cannot do gets written down. The list is dated, it is longer than you want it to be, and everybody who joins the beta gets it before they say yes. If what you need is on it, I would rather you knew now.

What happens next

The private beta opens this autumn. It is small on purpose. Use it for the music you actually write, and when something breaks or feels wrong, say so; one sentence is enough. What comes out of that is the only proof of a notation program that has ever meant anything: music that got played from it.

If that is you, the beta page says what it involves in both directions.

If it is not, leave an address at the foot of the page. You will hear when the beta opens and when the program is ready, and a short note every few weeks in between about what has been built. Nothing else.

Next Tuesday, then.