[linux-audio-dev] Developing a music editor/sequencer

Chris Cannam cannam at all-day-breakfast.com
Wed Feb 2 19:49:42 UTC 2005

On Wednesday 02 Feb 2005 19:02, Paul Davis wrote:
> I continue to think that this crazy.

So do I.

The design process at work here reminds me a lot of the way I approached 
Rosegarden: look at how other applications work and what they do, and 
then add in a few interesting generalisations to make it more 
potentially flexible in ways that happen to meet my own preconceptions 
about how people might use the system, thus resulting in something that 
looks innovative and interesting to me, and just looks like another 
sequencer or score editor to everyone else.

For example, Rosegarden contains structure intended to support things 
like arbitrary layout engines for editing; multiple different layouts 
on the same music data; event-based systems that are not MIDI, and so 
on.  Yet because it has taken so much development work just to do the 
basic MIDI and audio support that people expect from a sequencer, and 
because we are so few developers, most of this is still unused.  It 
would have been far better, for my own personal aims, to have worked on 
something that was not so obviously a sequencer application and that 
instead focused on the one or two experimental features I was really 
interested in and ignored everything else.

Don't get me wrong: I think Rosegarden is a successful piece of work and 
I'm very proud of it.  But if I was starting out now, I wouldn't be 
working on anything like it.  You should see some of our planning 
emails from five years ago, excitedly talking about how we just had to 
do one or two generic bits and bobs and the rest would simply fall into 
place.  It's a very hard delusion to avoid, but it's always a delusion 


More information about the Linux-audio-dev mailing list