[edit: the change would improve edge scrolling in monster cities, too. It's the potential complexity to build the perimeter that might take 0.01 seconds or longer in larger cities.]
When edge scrolling or following vehicles the hard line enforcement of the city outline is restrictive. Often city outlines are non-convex (thinkhaving just three squares at a four-square corner) yet benefit from and use roads and tracks outside the city perimeter (think highways crossing the not-yet-purchased tile). When following a truck it seems odd to hard stop at the city line when the route will resume within the city.
As a starter, add 1/8 (oe 1/2) diagonal;ly of the unbought tile to the viewership perimeter at the three-tile corners. This will solve about 75-90% of the routes and significantly improve edge scrolling around the city.
In the extreme, the viewable perimeter could be extended to the convex hull of the city tile perimeter. (which would add a diagonal half of the unbought tile) That might also be simpler to code. But the closest diagonal corner half-tile by half tile unbought tile is the most important to add. If your system prefers squares to diagonals, then add 1/16th (the NW/4 NW/4 "northwest corner of the northwest corner" or 1/4th (the NW/4 "northwest corner") of the unbought corner tile.
Linking up the at-most 8-10 extreme points on a 9-tile city could be the simplest way to go. If this slows down "perimeter rebuild and recheck" in 25-tile or 81-tile monster cities, then they can chenge their mod to add the capability to turn off beneficial edge scrolling and route following.
QED
vs vg,
scattered
PS, if you already have separate ownership and viewership outlines, then this suggestion is to add more unbought land within the viewership profile that is outside the owned-profile. And since tiles are bought by appointment, the viewership profile will be static compared to all the other data in the game.
When edge scrolling or following vehicles the hard line enforcement of the city outline is restrictive. Often city outlines are non-convex (thinkhaving just three squares at a four-square corner) yet benefit from and use roads and tracks outside the city perimeter (think highways crossing the not-yet-purchased tile). When following a truck it seems odd to hard stop at the city line when the route will resume within the city.
As a starter, add 1/8 (oe 1/2) diagonal;ly of the unbought tile to the viewership perimeter at the three-tile corners. This will solve about 75-90% of the routes and significantly improve edge scrolling around the city.
In the extreme, the viewable perimeter could be extended to the convex hull of the city tile perimeter. (which would add a diagonal half of the unbought tile) That might also be simpler to code. But the closest diagonal corner half-tile by half tile unbought tile is the most important to add. If your system prefers squares to diagonals, then add 1/16th (the NW/4 NW/4 "northwest corner of the northwest corner" or 1/4th (the NW/4 "northwest corner") of the unbought corner tile.
Linking up the at-most 8-10 extreme points on a 9-tile city could be the simplest way to go. If this slows down "perimeter rebuild and recheck" in 25-tile or 81-tile monster cities, then they can chenge their mod to add the capability to turn off beneficial edge scrolling and route following.
QED
vs vg,
scattered
PS, if you already have separate ownership and viewership outlines, then this suggestion is to add more unbought land within the viewership profile that is outside the owned-profile. And since tiles are bought by appointment, the viewership profile will be static compared to all the other data in the game.
Last edited:
Upvote
0