Re: [Rosegarden-devel] Matrix Editor Progress and Proposal

[D. M. M. <ros...@gm...>]

Now, moving on, your proposals...

First, I definitely agree that having multiple segments open in multiple 
separate matrix windows is inefficient and undesirable.  I'm fairly certain 
everyone would agree.

I think we've never quite worked out how to solve any number of problems for a 
multi-view matrix.  I don't know if there's some old discussion document 
laying around that might have already covered this ground, but thinking about 
it from a fresh start, I see:

1. Selections on the segment canvas can and should definitely continue to be 
able to span multiple tracks.  If we continue to display an IPB for the 
matrix, this makes deciding which IPB to display problematic unless the 
individual segment views are self-contained and independent of each other.  
This is an argument in favor of a tab-based sort of paradigm over some kind 
of all-on-the-same-grid overlay as has been discussed more recently.  
(Currently there is no IPB in the matrix at all, and I was planning to leave 
it that way.  I don't find the controls useful, and as they only affect 
controllers and whatnot at the very start of playback, I don't think it's 
terribly useful to be able to manipulate those controls inside a segment 
halfway through the composition.)

2. Regardless of where they reside in the track structure, segments can 
overlap in time.  This is necessary for dealing with multiple voices playing 
different lines on the same staff, and it's just generally useful.  Any 
all-on-the-same-grid display would have to take this into account and expand 
the grid along the time axis to cover the whole span of time covered by all 
open segments.  This in of itself would be more desireable than a tabbed 
collection of individual segments, as it would make the time relationship 
more clear, but it's probably really hard to get right.

So after thinking about those points, I'm finding that I'd probably favor a 
tabbed approach as more practical than some kind of all-on-the-same-grid, 
even though the latter would look better and potentially be more intuitive to 
operate (if executed well.)

Now, as to having track controls inside the matrix, that's an idea I just 
don't like, but I'm having trouble articulating why not.  I'd be interested 
in hearing other people weigh in on that one.  It just doesn't seem like a 
properly Rosegarden way of doing things, and it misses some point, but maybe 
I'm just stuck in my traditional view, and it's really a good idea.  I simply  
don't know yet.

Beyond all of this, I guess I just find it mildly disappointing that all this 
effort is going into a feature I rarely use anyway.  If anything is going to 
get revamped here, I'd far and away prefer to see it be the notation editor.  
If I ever wanted to insert program changes in a controlled and convenient 
way, I'd definitely want to do it from the notation editor.  All of this is 
kind of wasted on me, really.

You talk of how we could be something great if we just had solid direction, 
but in this case it's really a question that the top people left standing at 
this project all pretty much agree on a different direction from what you 
have in mind.  We're notation people.  I'm definitely a notation person.  I 
see Rosegarden's other capabilities as taking a back seat to notation, and 
I'm content to be a "sort of" sequencer.  I've looked at the matrix more in 
the last few months than I have in the last five years, and that's only 
because it's currently the only editor we have that's any kind of functional.

If you want to get me really excited, let's talk about pumping up the notation 
facility to the same kind of level you're proposing for the matrix.  I know 
it could be done, but I don't have the talent to do it myself, and Chris 
doesn't have the time.  Some of what you've already done sounds like it could 
be carried over, and if that is your eventual intention, you'll see far more 
enthusiasm coming from my corner.
-- 
D. Michael McIntyre