@TW It used to not be so if that is the case. But it appears you are correct in that KDE libs are completely removed from build requirements now. So that could be a significant improvement.
That being said, my suspicion is that the 27 MB number may be a little off, not saying you don’t see it but my first question is, does that include the shared libs used by Rosegarden or not?
Now keeping my above comments about memory in mind, in as far as desktop environments, here is the first comparison of memory usage article I found, and is almost recent at least being about 2 years old, so it could be out of date…
You will notice that by default razorqt was a bit heavier than other DEs mentioned here, but the second part of that article does mention you can switch it to run on top of openbox instead of kwin which apparently gives a significant boost in using less memory.
Now don’t get me wrong, I have programmed in Qt before, and I don’t think it is bad, especially not compared to Qt3 etc. Qt4 is a significant step forward, and I haven’t played with Qt5 to be able to comment there, but with their focus a bit more on mobile types I would expect good improvements in memory usage.
It is also worth noting that even on Rosegarden’s site they acknowledge it can be fairly resource hungry…
http://www.rosegardenmusic.com/getting/requirements/
Any modern desktop or laptop and most netbooks should have sufficient processing power and RAM to meet Rosegarden's minimum requirements. However, Rosegarden is fairly memory-hungry, and can demand a lot of processing power, so the more you can throw at it, the more enjoyable your experience will be.
... and pretty well echo what has been said in this thread...
Editing, playing and recording MIDI and editing score are less resource-hungry than recording or playing audio. The use of multiple audio plugins or soft-synths will especially impact system performance.
Now I am not saying Ardour can’t be resource hungry as well, just that by default I find it a bit less so than Rosegarden personally, but I haven’t gone through the entire process of measuring both Rosegarden and Ardour in a non-qt and non-gtk environment, though I suppose since I run e17 it wouldn’t be to difficult to do, in a similar fashion to the original article I posted on memory usage, actually measuring the free memory after loading each. Obviously that doesn’t take into account swap if any is paged out, but I should be able to disable swap for testing purposes most likely if needed.
But frankly I consider that to be a lot of work for not much benefit honestly.
Seablade