Is it possible to change the rule direction

[R C-R]

8 minutes ago, Old Bruce said:

I was speaking generally of the coordinate system Windows and Macs use. Cannot say anything about Linux because I didn't spend too much time with it back in the 90's.

This topic is not about the coordinate system the OS uses; it is about the coordinate systems that different applications use.

They do not all use the same one.

[Old Bruce]

10 minutes ago, R C-R said:

This topic is not about the coordinate system the OS uses; it is about the coordinate systems that different applications use.

They do not all use the same one.

I am well aware of that.

[R C-R]

Just now, Old Bruce said:

I am well aware of that.

Then I do not understand why you were talking about OS coordinate systems. That seems irrelevant here.

[Old Bruce]

1 minute ago, R C-R said:

Then I do not understand why you were talking about OS coordinate systems. That seems irrelevant here.

It is easier to program using the same coordinate system. Add in the arithmetic for each and every screen drawing operation and you have a significant number of cycles used up. Back in the day it would seriously slow down an application, not so much of a problem now but the convention stuck. Again I am not against using the Cartesian plane, I won't worry about it if it never comes to pass though.

If it does happen I'll just get my head around the 'new' system and continue on. 

I have no dog in this fight, I do support the request even though I don't care very much about it. I just felt that a little context about the OS would be helpful.

[GarryP]

If the user were allowed to change the direction of the Y axis, wouldn’t that mean that some (or all) layers, which were previously pixel-aligned vertically, would now be non-pixel-aligned?

For instance, creating a new A4 document (300 DPI, for example) in mm and then changing the UOM to pixels gives us a document which is 3507.874 pixels in height (without bleeds). If the origin was then moved to the bottom-left of the page/artboard/whatever, and the Y axis reversed, wouldn’t that make every vertical distance ‘off’ by 0.874 pixels (or maybe 0.126 pixels)?

If so, that could cause a lot of problems later on without the user necessarily being aware of them until it all ‘goes wrong’ at export.

Or have I got this wrong?

[lacerto]

7 minutes ago, GarryP said:

If the user were allowed to change the direction of the Y axis, wouldn’t that mean that some (or all) layers, which were previously pixel-aligned vertically, would now be non-pixel-aligned?

The code could perhaps check whether vertical pixel alignments are used in the document, and ask if the current vertical position is wished to be retained (at the cost of losing the vertical pixel alignment), or if the vertical pixel alignment should be retained (or the whole operation be cancelled). I checked whether Xara gives any warning (when using snapping to pixel grid), and the change causes losing vertical pixel alignment, but no. I am not sure -- maybe this should just be left to the user, since similar questions should then also be asked when the user e.g. is about to resize the canvas using a positioning option that causes losing of pixel aligment.

[GarryP]

1 hour ago, Lagarto said:

The code could perhaps check whether vertical pixel alignments are used in the document

How do you imagine this happening?
I’m not saying that it shouldn’t be done, I’m just wondering how the software would ‘know’ if pixel-alignment was ‘meant’.

1 hour ago, Lagarto said:

or if the vertical pixel alignment should be retained

How would the software retain the alignment?
Again, I’m not saying that it shouldn’t be done (if that’s what the users wants) but what would the software actually do?

It would be interesting to know what other software does under these conditions but I’m in agreement that the user must be in control over whatever happens with necessary warnings when something might happen ‘under the hood’.

[Old Bruce]

2 hours ago, GarryP said:

If the user were allowed to change the direction of the Y axis, wouldn’t that mean that some (or all) layers, which were previously pixel-aligned vertically, would now be non-pixel-aligned?

I would think that this would be a change that would be set in stone, not a user definable option. It could be user definable but as you point out there would be so many problems created by changing the 0, 0 origin on the fly.

Probably why the old 'wrong' way has persisted. Then think of applications with the ability to draw or make selections using scripts, macros or 'actions' that are using x, y coordinates, they would all be broken. 

[GarryP]

11 minutes ago, Old Bruce said:

Then think of applications with the ability to draw or make selections using scripts, macros or 'actions' that are using x, y coordinates, they would all be broken.

I didn’t even think about macros but that’s a really good point.

If users could change the direction of the Y axis – either when creating the document, or on-the-fly – would that mean that all macros, which used a different Y direction to that document, would not work properly? I imagine that would make macro creation and distribution a bit messy.

[lacerto]

12 minutes ago, GarryP said:

I’m not saying that it shouldn’t be done, I’m just wondering how the software would ‘know’ if pixel-alignment was ‘meant’.

I do not think it can. But assuming e.g. that all objects are vertically pixel aligned then it is a safe guess...

I guess that simplest method (seeing that the top bounding box of all objects is originally pixel aligned), is making sure that the same bounding box edge is aligned to the nearest pixel also after the vertical ruler has been flipped. But it is often much more complex, so therefore probably best to just leave it to the user.

[firstdefence]

It would be nice if the OP @duncang replied and explained what the reason for the request was.

[GarryP]

I think this area is probably a lot more complicated than we know about.
It sounds like it should be easy at first but it just gets more complicated as the problem is examined, even at this high level.

(Also, wouldn’t layer Y placement have to be measured from the lower-left, rather than the upper-left, after a Y-direction flip?)

firstdefence: That information would be nice to have.

All-in-all, if this Y-direction flip were to happen then an on-screen warning would definitely be required as it could cause some (most?) people a lot of difficult-to see-at-first problems.

[Old Bruce]

1 minute ago, firstdefence said:

It would be nice if the OP @duncang replied and explained what the reason for the request was.

I think that duncang was just asking if changing it was possible, or he may have made a request in the questions forum. We all have hijacked this. Mea Culpa.

[firstdefence]

2 minutes ago, Old Bruce said:

I think that duncang was just asking if changing it was possible, or he may have made a request in the questions forum. We all have hijacked this. Mea Culpa.

I know what Duncang wanted but just wanted to know why. sometimes using an app that fits the need is much less work than wishful thinking, the alternative in Affinity is to work upside-down.

[lacerto]

9 minutes ago, firstdefence said:

I know what Duncang wanted but just wanted to know why.

We can only guess, but one possibility would be having used CorelDRAW for years, which uses Cartesian coordinates. Or being accustomed to Cartesian coordinates because having worked a lot with 3D apps.

As already mentioned, apps need to translate coordinates all the time, reading in from one format, rendering or writing out in another. I do not think that there is much point in wondering complexities involved in flipping the coordinates and maintaining things like pixel alignment since if e.g. pixel alignment were crucial, I'd assume that the user would avoid flipping the coordinates just for fun, in the middle of designing something that requires exact positioning. The coordinate system can be changed any time at all in Xara Designer without any visual change in the document; display Y coordinates naturally change, but otherwise this seems to be a "trivial" operation.

As for Affinity apps, packaging functionality of three apps in one might make things more complex, but I would not expect this feature to be added in this app trio, anyway. 

[R C-R]

4 hours ago, Lagarto said:

As already mentioned, apps need to translate coordinates all the time, reading in from one format, rendering or writing out in another. I do not think that there is much point in wondering complexities involved in flipping the coordinates and maintaining things like pixel alignment since if e.g. pixel alignment were crucial, I'd assume that the user would avoid flipping the coordinates just for fun, in the middle of designing something that requires exact positioning.

I am not so sure that at least some users would be aware that flipping might mess up pixel alignment. Some do not even seem to be aware that pixel misalignment is possible, that objects can be placed or sized at non integer pixel values.