It’s both and. The fundamental concept is both, but the piano roll actually didn’t come along until [relatively] recently. You are correct that a rest is the absence of a note, which was one of dorico’s stellar features from the outset (that it could take care of those for you). But correct notation and thinking in terms of actual note values was also a fundamental concept. This works both in the piano roll sense, but it was very much also a notation thing from the get go (another early banner feature).
Even in version 1.0 the Play mode was there (although I did not start using Dorico other than simple testing in version 1.0). My point was that the way it represents the data underneath is more similar to the way DAWs store the data instead of the way notation programs do, and it always struck me that this was likely on purpose as a way to ease the path to Cubase integration at some point, and I think that design is quite brilliant in that sense. My explanation is very likely an oversimplification of a much more complicated process. I don’t have visibility into actual code to see what really happens underneath. But the Play mode and the way things are stored underneath is basically the same as it was in 1.0.
And certainly, I agree that having it do correct notation out of the box by default is one of the goals, so that if you import MIDI or something, or just enter things on the piano roll, they are already notated correctly. But what a lot of people think is that Dorico is changing what they wrote because it thinks it was not correct, and changing it to what it thinks is correct, which is not the case. I’ve come across a lot of composers who don’t want to use Dorico because they see in demos that people enter a half note on the second sixteenth of a 4/4 bar and the fact that it changes to a series of ties is viewed as a negative, because they think it is deciding by itself to change what they wrote. I would say Dorico has correct defaults, but it doesn’t force you to use correct notation and doesn’t change what you wrote. It’s a case of people misunderstanding the abstraction, in that making a new “half note” doesn’t actually make a literal half note as a single note, instead it makes a new note that is a half note in length in the data store, like a piano roll rectangle, and then decides how to render that on-the-fly, which allows you to do things like change time signatures, rhythmically shift it, and do other DAW-like operations.
That is a great explanation and helpful to understand the current behavior of Dorico, but, it doesn’t really matter to a user how the engine is working inside; the engine should simply work properly. I want to step on the gas and have the engine go, and step off the gas and have it stop going, regardless of whether the vehicle is a diesel or gas engine and regardless of whether it is manual or automatic transmission. So while it is informative, it is not an excuse for buggy behavior or improper functionality. Excusing playback timing errors or engraving errors by saying, “That’s because it is a DAW underneath the hood,” doesn’t really apply to a composer who is playing back an engraving, which is what Dorico is designed to do. Dorico is sold as engraving software for notational composers (with DAW features, and I also hope Dorico replaces all DAWs permanently).
For comparison, Guitar Pro (only about $60 compared to Dorico at $600) handles the incomplete bar situation very simply, by coloring the measure red, and repeatedly returning the edit cursor to the position of the missing/extra beats when the measure is next edited. Yes, Guitar Pro has a very different method of internally handling time and the grid; but, it counts beats and beats in the measure; and functionally as an engraving/playback app (while not having as deep an engraving or VST toolkit as Dorico), it does not have the problems that Dorico has in this area. If Playback is performed on improper measures I believe the playback cursor will stick on the missing beats to “add time” where it is missing, and it skips over beats to “ignore time” where there are extra beats, to always ensure that Playback is valid according to a real metronome. I note this only to say that these problems are able to be worked out in an engraving app to the satisfaction of composers. Guitar Pro can also have any audio track (i.e. a live recording, or an audio track exported from a video) imported along-side the score as an additional instrument with a visible waveform, and allows the user to align the barlines in the score to markers “stretched” on the audio waveform, which is a behavior similar to many DAWs, and this does not harm the notational correctness of Guitar Pro’s engraving. This functionality allows Guitar Pro to have both notational correctness with a time signature, and true swing during score playback, by allowing marker alignment to a live audio recording, etc. They handled “time”; so Dorico can handle “time” too.
Considering these various perspectives on how Playback should work with missing beats, it could equally be said that Dorico should extend the length of the prior beat, to fill the measure entirely, until the measure completes. In a 9/8 measure with only 6 beats, why doesn’t Dorico extend the last note to fill-in the remaining time? This is hypothetically what a DAW would do, if a DAW were keeping meter to a time signature properly: not send the MIDI Off, because there is no next note; the same as a fermata. Obviously this is not a good solution. But, it is the flip side to the argument. At the end of the day, I simply want to trust Dorico to engrave and playback music correctly as notated. Currently, I can’t do that. I write music with changing time signatures. When I change time signature, it introduces engraving mistakes and playback mistakes.
I have wanted something similar to this:
What I want is, the ability to select (highlight) one or multiple bars, use popup to enter a new time signature, and have Dorico add both the new time signature at the start of the selection, and add the restoring time signature after the end of the time selection (so after the selection region, the time signature has not changed). Currently, if highlighting bars and inserting a new time signature, the time signature is added at the start, only. Adding this functionality would ensure to the user, theoretically, that “only the highlighted region has had a time signature change” thus avoiding the ambiguity of where to restore the time signature since all the notes may have moved position.
Importantly, I will note that I have previously sent full projects and long feature requests to the Dorico team for prior versions, and the Dorico team has implemented features as I described, or implemented functionality to fill the holes as I described, and solved bugs as I described. This is far beyond what other engraving software teams will ever do. Unfortunately, the specific couple of users (the same as always) who lambast those with feature requests or discourage those who note implementation errors, are behaving incorrectly and are acting improperly on this forum. I have confidence in the Dorico team to implement a good solution when broadly informed of the problem users face when using Dorico. In contrast, I do not have confidence in a specific couple of users on this forum (who I have called out before in prior threads) and they are absolutely not helpful in moving Dorico forward.
As I indicated earlier in this thread, the reason why both DAWs and Dorico work well for film scoring is how you can change time signatures at any point without disrupting the synchronization with the picture - everything still synchronizes, because it is the same number of beats in length (and therefore time in seconds) to each hit point. All notes arrive in correct synchronization with video frames even if you change time signatures. If Dorico were to change the main mode of operation so that it added or removed time whenever you changed time signatures, thus making your 9/8 bar complete, this would throw off the synchronization of anything in the following bar that was already correctly synchronized to picture. That’s why it was discussed as an additional optional mode rather than replacing the current way of doing things, as having Dorico add or remove time in these cases would basically make it useless for film scoring, in my view.
DAW’s won’t enlarge a bar (add space) to make it complete. Cubase is my example here. Although DAWs do preserve complete measures, they do this by shifting the notes by a certain number of beats so that the notes fall in the same places on the timeline, and they maintain sync with picture. In Dorico, using this same behavior could be incredibly destructive if there was a lot of custom engraving done. (Numerous times I’ve accidentally added time in Dorico through insert mode, and after shifting it back, it didn’t look quite right, and I had to go back to a previous save, or undo.) So Dorico handles this a bit differently, by having an incomplete bar just before the time signature changes, instead of shifting the notes over by X beats like a DAW would, probably to avoid potential custom engraving from getting wrecked if it were to follow the normal DAW behavior of shifting all future notes over by a beat or two so that they all landed in the same place on the timeline to maintain sync with picture. But the end result of maintaining sync with picture in these types of edits is common to both Dorico and DAWs.
Other notation programs like Finale and Sibelius and MuseScore are not usable for film scoring in my view because of how changing time signatures in those programs will cause loss of synchronization with picture by adding or removing time without prompting. Your Guitar Pro software likewise would not be usable for film scoring with the behavior you described.
And after all this circular discussion, since it sounds like you really just want all the bars to be complete, I think we’re back to the consensus proposed by Michel here, which most of us I think were in agreement with: