I *like* this idea!
On Thu, Apr 21, 2011 at 01:04:06PM +0200, Arnold Krille wrote:
One would basically split the interface in two parts. A daemon in
contact with the main engine (using OSC or any other means), which would
then export all parameters to a file system building a suitable
hierarchy. Each parameter would have a handle node e.g "freq" and a
description node "freq.desc". The description node would contain
relevant information, such as a label for this parameter, a type, a unit
value when relevant (such as hz, bpm, %, etc.) and a range of values (or
choices if it's a list).
Then the actual UI would simply have to crawl this filesystem and
populate its tree based on the values and their description. Likewise,
it could modify the values simply by writing to the appropriate node.
There could even be a separate FS subdirectory containing "actions"
(such as save, load, reset, etc.) which could once again simply be
triggered by writing to the appropriate file.
This opens up a tonne of possibilities (scripting, control from remote
applications, multiple UIs).
I've no idea how feasible it really is, but it sounds great!
Linux-audio-user mailing list