This post is intended as a courtesy notification to any who may be
interested. If not or nobody, I apologize for the waste of time and
bandwidth.
I have released an independent fork of Rosegarden at
https://github.com/thanks4opensource/rosegarden-fork/. This is compliant
with Rosegarden's license ("COPYING").
Note that forking the project was not my first (or Nth) preference. I
believe that forks are in general bad for open source software as they
increase the FUD factor put forth by proprietary/closed-source vendors.
("You use Rosegarden? Which version? See, that's why you don't want to
get locked into open source software!")
The fork is also bad for me personally. When I first considered working
on the Rosegarden sources I was elated to find that the project was
actively being maintained and developed. (Also surprised, given the
prevalence of open source "abandonware", and then even more so when I
learned how far back Rosegarden's history goes.) My hope, and frankly
expectation, was that my contributions -- initially small, but which
grew in size and scope -- would be incorporated into the codebase,
hopefully with collaborative back-and-forth improvements. And that
eventually, as the release schedule progressed and distributions picked
up the latest version, I could delete my own development branches and
simply use distro binaries.
But it has become increasingly clear that my merge requests aren't going
to be accepted into the Rosegarden mainstream, at least not in any
timely fashion and probably not ever. I have invested far too much time
and effort (currently 20x my original estimate) (and Rosegarden is very
much a sideline to my main open source work) to relegate it solely to my
own private use. I believe the fixes and new features I've added
represent worthwhile improvements that others could benefit from. Time
may or may not tell.
On an even more personal note, in retrospect I regret having chosen this
development path. I had thought I could "hack in" two minor changes: The
one bug and one feature I originally posted about, i.e. MIDI input
playing the current editor's active segment's instrument, and key-aware
matrix editor highlighting. But at each step in a long chain of
development I ran into further missing features that I needed for my own
Rosegarden use, and even more so internal architectural failings and
omissions that made implementing anything far more difficult than it
should have been.
I also realized in retrospect that I could have implemented a
matrix-editor-only application from scratch in far less time. As I once
posted to a bug report, I had previously written a primitive non-GUI
application with the basic algorithms and an ALSA MIDI back end. I
wouldn't have ended up with the myriad of other capabilities Rosegarden
provides, including the much more difficult to implement notation
editor, but as I personally only use notation for communicating with
tradition-bound friends, and could have always exported MIDI to
Rosegarden (or Musescore) for producing offline output, that wouldn't
have presented an insurmountable problem. "Live and learn", as the
saying goes. Or maybe more appropriately given Rosegarden's original
developers, "In for a penny, in for a pound".
I hope this explains my motivations for forking Rosegarden (again, if
anyone is interested in them). As of now the fork still isolates its new
code in the separate thanks4opensrc_devel branch (with the exception of
a modified README.md file in master that clearly indicates the repo is a
non-official fork) although that may change in the future. I have one
more major feature planned, plus a raft of smaller ones, but currently
hope to slow my development on the project in order to return to the
other work mentioned above. I'm describing the branch structure on the
off-chance that there's any future interest in looking at parts of the
fork for potential inclusion in the official repository, either as code
or merely as ideas for independent/separate implementation. But note
that the branches have already diverged significantly (my last merge
with master [833ea5] was very difficult) and will likely continue to do
so. Of course my recommendation is that the forked branch be merged
wholly, as-is (git-merge "theirs), but I don't anticipate that happening. ;)
Finally, and as publicly stated in the fork's README.md, please accept
my appreciation for Rosegarden and acknowledgement of the man-decades of
work that have gone into it. As much as I think there are significant
architectural issues (not surprising considering the long development
history) and that a deep refactoring/re-implementation (far beyond the
recent "lint"/const-correctness merges, as valid as those are) should be
undertaken (not something I'm likely to undertake on my own) I still
believe the basic code design is sound and that the program's features
and workflows are exceptional. My best wishes for the future of
Rosegarden, in any and all of its forms.
--
MARK aka "thanks4opensrc"
_______________________________________________
Rosegarden-devel mailing list
Rosegarden-devel@lists.sourceforge.net - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-devel
