Name 'Divided' Objects

[Image Surgery]

I have been using the divide operand in Designer and notice that all object names are discarded when using this operation. The other boolean operands keep a reasonable name intact, in my testing.

Steps to reproduce:

1. Create two overlapping shapes and give them non-default names
2. Use the 'Divide' operand (`Layer > Geometry > Divide` or toolbar button)
3. Note the resulting objects have no names, naming is lost

Tested on Designer 2.1.0 on MacOS 13.4, MBA15" Model `Mac14,15`

(edit 2023-06-29 11:30)

Before dividing, objects have properties, such as colour, many of which are maintained after the divide. If the object's name is considered to be a property in the same way colour is, then apply it to the resulting objects in the same way. That's the whole request, parity with the rest of the tool(s). Other enhancements suggested below should be considered on their merits for sure but they're not my feature request (bug report?).

Edited June 29, 2023 by Image Surgery
added clarity to request after discussion below

[GarryP]

Which name(s) would you suggest that the software gives each of these new layers?

For instance, if I have three layers named “red”, “blue” and “green”, and “red” and “blue” overlap, and “red”, “blue” and “green” overlap – as per my screenshot – what name should each of the seven resultant layers, after the division, be given? Any what is the rationale for those names?

[fde101]

Just tack a "1", "2", etc. onto the end of the name of the original?

[GarryP]

The name of which "original"?
If two shapes overlap then the resultant shapes, after the divide, will come from one or more, maybe many more, layers.
So which layer name does the software “tack” a number onto?
And is this good enough to give the user the information they need?

Which leads me to the question: “Why would it be useful to name the divided layers in the first place?”
If that can be ascertained then we might be able to progress better on this.

[fde101]

32 minutes ago, GarryP said:

The name of which "original"?
If two shapes overlap then the resultant shapes, after the divide, will come from one or more, maybe many more, layers.

From what I can tell, that is not the case; rather the overlapping part only contains the shape of the part that was on top.  The layer that was on the bottom guides the place where the layers are cut, but the resulting color comes from the layer that was on top, so that is the one that remains, and the parts that were tucked underneath (from the layers on the bottom) are disposed of.

If I have two rectangles that overlap one corner, then I get three parts: the part of the rectangle that was underneath, which did not overlap with the upper rectangle; the part of the upper rectangle which overlapped with the other; the part of the upper rectangle which did not overlap with the other.  As a result, I would have the remaining part of the underneath rectangle, which could simply keep the old name of the underneath rectangle (as there is only one remaining part), plus two parts taken from the upper rectangle, which could be named with "1" and "2" added to the names of the originals.

[Alfred]

1 hour ago, fde101 said:

As a result, I would have the remaining part of the underneath rectangle, which could simply keep the old name of the underneath rectangle (as there is only one remaining part), plus two parts taken from the upper rectangle, which could be named with "1" and "2" added to the names of the originals.

But you don’t have one part from the lower rectangle and two parts from the upper one! You have one part from the lower rectangle, one part from the upper rectangle, and one part from both.

[HCl]

.

[fde101]

5 minutes ago, Alfred said:

one part from both

Then why are the colors not mixed together?  Try it with two shapes where the opacity is reduced.  The color of the part that you claim comes from both shapes, is only the color of the upper shape, so the color of the one underneath is lost and ignored.

What seems to be happening is that both shapes are simply cut at the intersections and the parts that are underneath are then deleted, leaving only the parts that were on top.  In the end, the effect of the operation is actually the same... thus the way I am looking at it is that there are two parts left from the upper shape, and one from the lower one.  None of the individual parts represent more than one of the original shapes.

[GarryP]

I think there may be some confusing over what I meant by “will come from”.
What I meant is that the geometry of the resultant shapes may be partially, or wholly, obtained from that of one or more of the original shapes.

By way of explanation, here’s another screenshot (document is also attached).

On the left are the original shapes (in the “Original” group).
On the right are the shapes after division (in the “Divided” group).
The layer I have renamed as “X”, is selected, and takes its shape from all three original rectangles.

As such, what name would be given to layer “X” during the division process, and why that specific name?

It could be argued that “X” was 'part of' the green rectangle and could be named “green<something>” but it could also be argued that it was 'part of' the blue rectangle and could be named “blue<something>” since layer “X” is wholly ‘inside’ the area of the original “blue” layer.

My point in asking this is that: “If the names of the resultant layers are not meaningful to the user in some way that is useful then there’s probably no reason to name them.”

divided-shapes.afdesign

[HCl]

.

[GarryP]

If there is always a way to name the divided layers that gives them some useful meaning to the user then I’d be happy with that.

I’m not against this request in any way (unless, as I said above, the names given are meaningless); I’m just trying to find out what the details of the requirements are.

[walt.farrell]

22 minutes ago, Return said:

In your example X would be green-1 as this is green and was part of the green rectangle.

In his example, where that object is named X, it looks like it is green and red and blue, to me.

[HCl]

.

[walt.farrell]

35 minutes ago, Return said:

The selected part is obviously green and was never part of the red or blue shape before the dividing.

Then why does the bounding box have handles on lines that are green, red, and blue?

[HCl]

.

[GarryP]

41 minutes ago, Return said:

The selected part is obviously green and was never part of the red or blue shape before the dividing.

True, but its geometry of the resultant shape is dependant upon the red and blue rectangles also; its ‘sides’ (not all of which actually exist) are green, red and blue.

It’s coloured green because that’s the way the software has divided the shapes but it could be argued that the shape should be blue because its geometry is wholly within the geometry of the blue rectangle.

It’s also wholly within the green rectangle, so is it ‘part’ of the green rectangle or ‘part’ of the blue rectangle?

It could be either and the user has no control over it (as far as I know, maybe it’s a layer-ordering thing, I haven’t checked) so they may be expecting something else.

[fde101]

5 hours ago, GarryP said:

It’s also wholly within the green rectangle, so is it ‘part’ of the green rectangle or ‘part’ of the blue rectangle?

There were originally two shapes, one green and one blue, that overlapped completely the same area.  Both were cut up by the other (acting like cookie cutters to split each other up) and the blue one was deleted because it was underneath the green one.

That is how I read it anyway.

[HCl]

.

[GarryP]

I think it might, hopefully, help to get my concern across if I simplified the question.

In my new screenshot (and attached document) we have:
On the left, in the “Original” group, two identical rectangles, called “Left” and “Right”, overlapping by 50%;
On the right, in the “Divided” group, the result of duplicating the layers on the left and performing the division.

What name should the selected layer “X”, on the right, be given?
Or, to put it another way, which layer was layer “X” taken from? Left or Right?

divided-shapes-2.afdesign

[bures]

The application knows from which object new objects were created. I wouldn't complicate it with the edge case where the objects look the same to a human. Let the application take the names of the original objects and add a sensible label to them.

Of course, the application must resolve conflicts with existing names.

[HCl]

.

[Oufti]

5 hours ago, GarryP said:

What name should the selected layer “X”, on the right, be given?
Or, to put it another way, which layer was layer “X” taken from? Left or Right?

It should take its name from the layer which was above on the time you divide.
In this case, Left. 

[Image Surgery]

Hi folks, OP here. I thought this was simple but a lot of people are quite confused, and I'm seriously surprised at how many well meaning but quite intense responses there's been. A touch overwhelming, honestly. I'd suggest not looking at it visually, but instead to think about the logic behind the operation, and how it already works. It works perfectly if you consider that a colour is a property of an object and is maintained following a boolean operation. In this case, the name is also a property of an object, and that should be maintained. This would sound so much stranger a request if I was saying 'my blue square becomes a default black square after a boolean operation…' but instead just read it as 'an object' property is removed after a boolean operation', in which case simply copy the existing behaviour for _other_ properties which are maintained.

The part about the name, I'd be inclined to keep it the same. There is no problem having multiple objects share a name, that's the job of the user to give good names to their work. And in the case of splitting an object, we maintain the colour and other object properties where possible (not green-ish, but identically green), so I would posit that the best, most general purpose naming would be no change at all. Not `Object-1` or `Object-1-2-1` which when split sooner or later turns into your phone number; just like the colour, leave that to the user to choose. (Also a shout out to the Affinity team for showing default names differently, it's nice to know that 'Rectangle' if it's grey, was not named by me. It's just super clear and nice, thank you!)

As for my use case, it could be anything, and that's my point in requesting this feature. But if you. must know specifically, I was using Designer to split some lumber that I have. Each piece is unique, and I wanted to preview how much yield I could get from my stock. However I quickly noticed that with every division, I needed to re-name my pieces, which made the task unpleasant and harder. I was using names as the identifier, it could be said that a workaround is to apply the unique identifier to the colour instead of the name. However that's clearly not as good a solution so I request the change.

I'd also love to encourage the rest of the discussion, but I'd put it under the 'feature enhancement' category, should it exist, rather than what seems more akin to 'bug report' (or in this case, possibly technically a feature enhancement however the enhancement itself is actually feature parity).