What is the reasoning for the asymmetrical boundaries of Miller oblated stereographic?

[Piotr]

Those four maps are the same, but they have different boundaries, even though they are unconstrained.

[daan]

The projection is defined for a particular aspect and is not meaningful with a different aspect. The boundaries have been defined against that aspect.

— daan

[Piotr]

The projection is defined for a particular aspect and is not meaningful with a different aspect. The boundaries have been defined against that aspect.

— daan

So why is then there an option to change aspect?

[daan]

Probably because the implementation never got updated in the transition from Geocart 2 to Geocart 3, and nobody noticed. It’s a bug.

— daan

[Atarimaster]

At least it’s a bug that lets you create funny if useless projections…

[Piotr]

At least it’s a bug that lets you create funny if useless projections…

Or it could be used to map oval-shaped regions, even though this projection was meant to map one specific region, just like how GS50 could be used to map US-shaped regions, or Mercator could be used to map great circle shaped regions.