This reproduces the unfortunate behavior that I encountered while constructing something more complicated over time.
The steps here save to a different file each time .
That is not necessary to reproduce the behavior.
Only the Save/Close/Load cycles are needed.
Start a fresh Cinderella session
Create a circle C0 about point A
Create three lines thru A
Use point to create one intersection point with the circle for each of the lines (a,b,c)
Save as LinkedAppearance-a?.cdy
Close associated windows
Load LinkedAppearance-a?.cdy
Create the other intersection point opposite point B
Inspector shows a white zero size point; change it to size to 5 & color lime & set labeled & identifier to E
Save as LinkedAppearance-b?.cdy
Close associated windows
Load LinkedAppearance-b?.cdy
Observe the other intersection point opposite point C is displayed in lime (it was not visible before the save)
Inspector shows color is lime, size is 5, labeled is set; change identifier to F
Save as LinkedAppearance-c?.cdy
Close associated windows
Load LinkedAppearance-c?.cdy
Observe the other intersection point opposite point D is displayed in lime (it was not visible before the save)
Inspector shows color is lime, size is 5, labeled is set; change identifier to G
Save as LinkedAppearance-d?.cdy
Now change the color of any of the lime points (E,F,G) to yellow.
Now redraw the view (the easiest way to do this is to zoom by using the mouse wheel or doing any translate, zoom-in, zoom-out or fit of the view) .
Observe how all of the lime points are now yellow
Select any of the linked points (E,F,G) and in the inspector, clear "Labeled".
Observe how all of the (E,F,G) labels disappear only to return when "Labeled" set again.
The "Draw Border", "Point Size" as well as all font settings (Font,Size, Bold & Italics) affect all of the points when the view is redrawn.
Observe that if LinkedAppearance-a?.cdy is loaded now, the other intersection point opposite point B already appears as a yellow colored size 5 unlabeled point.
This is why starting a fresh Cinderella session was necessary as the first step to get predictable results.
Apparently something is being shared between constructions so much so that when the color of the now yellow unlabeled point in LinkedAppearance-a?.cdy is changed to cyan,
redrawing LinkedAppearance-d?.cdy shows that its yellow linked points are now also cyan.
Observe that if either one of LinkedAppearance-b?.cdy or LinkedAppearance-c?.cdy is now loaded, all of the linked points (unnamed, E,F,G) in all of the constructions are now colored lime immediately.
The same behavior can be observed in "Angular Bisector" lines:
Start a fresh Cinderella session
Create three lines thru A
Use "Angular Bisector" to create one bisector line for each combination of the lines (a,b,c)
Save as LinkedAppearance2-a?.cdy
Close associated windows
Load LinkedAppearance2-a?.cdy
Open a "Construction Text" view and click on the unnamed bisector in that view
Inspector shows a white zero size line; change it to size to 1 & color lime & set labeled & identifier to h
Close the "Construction Text" view
Save as LinkedAppearance2-b?.cdy
Close associated windows
Load LinkedAppearance2-b?.cdy
Observe that another unlabeled bisector line is displayed in lime (it was not visible before the save)
Inspector shows color is lime, size is 1, labeled is set; change the identifier to k
Save as LinkedAppearance2-c?.cdy
Close associated windows
Load LinkedAppearance2-c?.cdy
Observe that another unlabeled bisector line is displayed in lime (it was not visible before the save)
Inspector shows color is lime, size is 1, labeled is set; change the identifier to l
Save as LinkedAppearance2-d?.cdy
Now change the color of any of the lime lines (h,k,l) to yellow.
Now redraw the view.
Observe how all of the lime lines are now yellow
Other linked setting also change as happens with the points example above
Observe that if LinkedAppearance-c?.cdy is now loaded, all of the linked lines (h,k,l) in the LinkedAppearance2-d?.cdy construction are now colored lime immediately.
Finally, just what elements are impacted by this behavior?
The presumption is that any element multiple possible output elements would be impacted.
The list would then be following:
Line and circle/conic intersections, two points.
Angular bisector, two lines.
Circle/conic intersections, four points or only two (no conics involved).
Ellipse by foci and point (the other solution of course is the hyperbola)
That these are involved can be shown using:
Two lines thru a point
The angular bisector of the two lines
A circle about the point
An intersection point of the circle and one of the lines
An ellipse by foci and point positioned to intersect the circle in four places
An intersection point of the circle and the ellipse
Then cycling Save/Close/Load and giving the identifier a non-blank name, and changing the setting to make the element visible.
Sometimes selecting the unnamed element requires using a "Construction Text" view.
The result, attached (could not be attached) here along with the other two construction sequences, is LinkedAppearance3-a?.cdy thru LinkedAppearance3-g?.cdy.
If the original bisector in LinkedAppearance3-g?.cdy is deleted and the construction is saved as LinkedAppearance3-h?.cdy, closed and reloaded it reappears unlabeled and lime in color.
Now the appearance of both angular bisector lines are linked with the other linked elements.
By deleting the original output elements and and using a Save/Close/Load cycle and setting their identifiers, it would be possible to link the appearance of all of them.
Beware that the window shrinks by 14 pixels in height (different unfortunate behavior :-) with each Save/Close/Load cycle.
Just open a new Euclidean view and close the original when the view becomes too small.
Observed using Version: 2.9 build 1783 (2015/08/03 18:17)
Best regards,
Alexander Elkins
PS: Attach File is not working here!
|
