Eliminate selection duplicity

[q]

  
This discussion began here in a Hairpin topic.

On the Mac, if I drag select the last note of a line, and shift drag select, Encore will tie notes, but it will NOT slur them.

Yes, same here on Encore/Windows 5.0.2. Another unfortunate example of the inconsistency in implications of selections attained with different gestures.

Thanks for confirming this on windows. Beautiful description of the general problem and the UI issues at hand.

I mistakenly thought the only way of slurring slur across lines is to shift+double-click the last note of line x, and shift-double-click the first note on line x + 1 ... but tonight's test shows that one CANNOT tie or slur across a end of line with this type of selection.

I'm not familiar with selection by Shift+double-click - we do not have that in Encore/Windows 5.0.2. Thus I cannot test here that scenario.
In Encore/Windows 5.0.2, If I individually select the last note in a system with Shift+Click, then individually select the first note in the same staff in the subsequent system system with Shift+Click, I can apply a slur (with Ctrl-L) or a tie (with Ctrl+T), if the notes are eligible for a tie, which extends across the end of the first system.

Curious difference between the two platform versions.

Thanks for your scrutinizing eye, Doug.

Indeed, Shift-clicking a note (and then the next) will indeed allow slurring—and if the notes are of the same pitch, tying is allowed, beaming is not allowed ... and when the user tries to beam the selection is lost. Actually the beam command works but ONLY to unbeam. And there is an option (at least on the Mac) to shift+double-click to select a note, and subsequent note(s) ... but slurring and tying fails with this type of selection.

This is truly messed up.

Encore is simple and intuitive in so many ways, but when we find peculiarities, they often run quite deep.

For a gander at some really odd stuff (possibly mac only) slowly shift-click a single note successively (slowly enough to fall below the double-click threshold.) In most applications the second shift-click would deselect a selected object; in Encore the object stays selected. Interestingly the third click does deselect—its behavior is what we'd expect from the second shift click. And there's more

Here's a summary of the behavior of successive shift-clicks:

shift-click 1: selects note
shift-click 2: imparts no observable effect
shift-click 3: deselects the note
shift-click 4: imparts no observable effect
shift-click 5: appears to restart the cycle ( i.e. same result as shift-click 1, suggesting that this is a four-step cycle.

q

[q]

 
For a gander at some really odd stuff (possibly mac only) slowly shift-click a single note successively (slowly enough to fall below the double-click threshold.) In most applications the second shift-click would deselect a selected object; in Encore the object stays selected. Interestingly the third click does deselect—its behavior is what we'd expect from the second shift click. And there's more

Here's a summary of the behavior of successive shift-clicks:

shift-click 1: selects note
shift-click 2: imparts no observable effect
shift-click 3: deselects the note
shift-click 4: imparts no observable effect
shift-click 5: appears to restart the cycle ( i.e. same result as shift-click 1, suggesting that this is a four-step cycle.

q

Even stranger:

Shift double-click a note
Shift-click another note
Result, as expected: Both notes are selected

Press Command [l] to slur
Unexpected results:

a) Encore does not slur
b) Encore deselects the Shift double-clicked note ... but leaves the shift-clicked note selected

q

[q]

This seems out of parallel: Guess Durations requires a double-clicked measure selection whereas Revert to Raw does not. But to me that's a moot point, as I'd like to see Revert to Raw abolished.

The keyboard shortcut (command/control g) is problematic. It operates only with full measure selections (for instance, when a measure is double clicked.) A drag selection does not suffice, because that approach never enables the Guess Durations option. When command/control g meets drag selection Encore should produce an alert sound, but it does not. The ONLY indication that the keystroke has failed is that the Notes menu does not highlight/unhighlight.

The disabled status of Guess Durations is obvious to a user viewing the Notes menu, but not to the user invoking the keystroke. It just looks at though Encore is unresponsive.

Until the selection duplicity is resolved we'd benefit from some "can't comply with command" alert.

q

[polarbreeze]

It seems to be inescapable that Guess Durations requires a full-measure selection (or an integral multiple of). The reason is that to do its thing it needs to decide exactly where in time the first note starts and also exactly where the last note (of the selection) ends. It then quantizes between those two points in time. That works fine if the selection is a full measure because the start time and end time of the measure is known and fixed. So far so good...

But suppose the user selects a group of notes that does not start and/or end exactly on a barline (ie at a known point in time). What is Encore to assume about the exact timing of the period that has been selected? Is it to take the notes before and after the selection as precisely timed in order to determine that? Well, maybe, but when you think about it, that's more or less the same as quantizing the whole measure anyhow.

I'm not sure if I'm expressing this very well but I'm convinced that "Guess Durations" has no practical use/meaning unless it operates on integral numbers of complete measures.

Note: I'm not saying that the raw note data always starts and ends on barlines. Clearly that's not the case. But the point is that the only safe place for Encore to "draw the line" when it does its quantization is at the barlines - because the barlines are the only available reference to fixed, known points in time.

[q]

It seems to be inescapable that Guess Durations requires a full-measure selection (or an integral multiple of).

Of course. I hope nothing I've written implies that I think Guess Durations should/can be brought to bear on partial measures (except those which are partially full, which indeed it handles.)

The reason is that to do its thing it needs to decide exactly where in time the first note starts and also exactly where the last note (of the selection) ends. It then quantizes between those two points in time. That works fine if the selection is a full measure because the start time and end time of the measure is known and fixed. So far so good...

But suppose the user selects a group of notes that does not start and/or end exactly on a barline (ie at a known point in time). What is Encore to assume about the exact timing of the period that has been selected? Is it to take the notes before and after the selection as precisely timed in order to determine that? Well, maybe, but when you think about it, that's more or less the same as quantizing the whole measure anyhow.

I don't know the ins and outs of any behavior behind the veil—and I wonder if development does—but I think its dangerous to make assumptions about Encore's internal operation.

I'm not sure if I'm expressing this very well but I'm convinced that "Guess Durations" has no practical use/meaning unless it operates on integral numbers of complete measures.

Again ... agreed. Perhaps I should review my messages to see why you believe I support an opposing view.

My point is about selection duplicity:

Why can't Encore equate the "drag selection of ALL notes in a measure" with a double-clicked measure?

Drag selecting all visible objects should mean that the entire measure is selected, at least for functions that operate on fully selected measures.

Update: The two statements above do not accurately describe what I meant. I meant that "drag-selecting ALL note/rest objects in a measure" should qualify to enable Guess Durations for that measure.

Do we really expect users to discern the difference. I don't think the User Manual differentiates, nor should it. The behavior should change to conform to the very likely preconceptions of the user.

The indented points above are independent from what Guess Durations can or should do. And there's a lot to worked out about that.

Thanks to our input (yours, Doug's, Eveline's and mine) on Guess Durations and Swing Playback, hopefully a spec and a reasonable User Manual explanation will manifest. As you know, these functions are full of anomalies, and not explained properly in the User Manual, nor in their current state could they be.

q

[polarbreeze]

I'm not sure if I'm expressing this very well but I'm convinced that "Guess Durations" has no practical use/meaning unless it operates on integral numbers of complete measures.

Again ... agreed. Perhaps I should review my messages to see why you believe I support an opposing view.

I wasn't directing my comments at you - just musing in general. Others have suggested that Guess Durations should apply to any selection - I'm just making the point that this isn't really possible.

My point is about selection duplicity

I know, sorry for the OT.

[polarbreeze]

[list]Why can't Encore equate the "drag selection of ALL notes in a measure" with a double-clicked measure?

It seems to me that it's not necessarily the same thing. Selecting a measure includes all objects in that measure. Drag-selecting all the notes doesn't do that and isn't intended to do that.

As you say, a case could be made for making drag-select of all visible objects in a measure to be equivalent to selecting the measure but I think there could be troublesome boundary cases. For example (of a limit case), consider a full measure followed by a second measure which is empty except for a hairpin which starts in the first measure and ends just into the second measure. The proposed rule would cause both measures to be selected but I don't think that's what the user would expect. I'm not saying it can't be worked out but I think it's non-trivial.

[q]

Why can't Encore equate the "drag selection of ALL notes in a measure" with a double-clicked measure?

It seems to me that it's not necessarily the same thing. Selecting a measure includes all objects in that measure. Drag-selecting all the notes doesn't do that and isn't intended to do that.

I know nothing of Encore's intentions, not without a spec. I'm only reporting what I believe is potentially confusing to users. I know about the issue and still it trips me up.

As you say, a case could be made for making drag-select of all visible objects in a measure to be equivalent to selecting the measure

Agreed (obviouslyn ... that's my position!)

...but I think there could be troublesome boundary cases.

I can't think of any, off hand.

For example (of a limit case) ...

Sorry, not familiar with the term "limit case"

... consider a full measure followed by a second measure which is empty except for a hairpin which starts in the first measure and ends just into the second measure. The proposed rule would cause both measures to be selected but I don't think that's what the user would expect. I'm not saying it can't be worked out but I think it's non-trivial.

I think each individual function, like Guess Durations, would have to decide how to deal with something like a hairpin that extends into a neighboring measure ... but since Guess Durations operates only on notes and rests, I don't see an issue with your example. Guess Durations should simply ignor the hairpin. Nor do I understand why you say that the hairpin straddle would cause both measures to qualify as selected. Was there something in what I wrote that suggests that?

q

[polarbreeze]

...Nor understand why you say that the hairpin straddle would cause both measures to qualify as selected. Was there something in what I wrote that suggests that?

I think you said that if (in the proposed mode of operation) all visible objects in a measure are selected, then that measure is deemed to be selected. So if your selection rectangle included the whole hairpin, and if there were no other objects in the second measure, then both measures would be selected.

[q]

...Nor do I understand why you say that the hairpin straddle would cause both measures to qualify as selected. Was there something in what I wrote that suggests that?

... if your selection rectangle included the whole hairpin, and if there were no other objects in the second measure, then both measures would be selected.

Ah, yes. The utter Booleanality of it all.

I think you said that if (in the proposed mode of operation) all visible objects in a measure are selected, then that measure is deemed to be selected.

Indeed I did. I wrote:

Why can't Encore equate the "drag selection of ALL notes in a measure" with a double-clicked measure?

Drag selecting all visible objects should mean that the entire measure is selected, at least for functions that operate on fully selected measures.

This was inaccurate, so I updated the original post to express what I meant, adding the following:

Update: I meant that "drag-selecting ALL notes in a measure" should qualify to enable Guess Durations for that measure.

I know I've mentioned this point previously, but have you noticed? Even with a single note selected Revert to Raw (for whatever it's worth) is enabled and affects an entire measure. I do not endorse that.

If a justification is ever found for continuing to include Revert to Raw in Encore's interface, perhaps Revert to Raw and Guess Durations should be moved to the "Measures menu” as they presently operate only on whole measures.

q

[q]

 
Here's a recent example of a selection issue:

http://www.gvox.com/forum/viewtopic.php?f=7&t=3602

q