Here is the link to the *.svg version of the images:
== less important side-notes ==
The calling context-hierarchy is potentially disjoint and thus forms a half order. Connecting stuff via disconnected simultaneously open window-views should be supported. This is related to navigating the name-space-context - another topic - maybe browse like a wiki??
(Cloning window-views should be supported.)
In case one plans to extend super-contexts beyond project borders (possibly unbounded – well I guess that is more a goal of the Paul Chiusano’s “unison” project than “luna”) then the number of super-contexts might become to big for the tab solution and a general search becomes more appropriate.
Nonlinear squashing, greyout, fadeout at context-borders may be useful. In case of the pagoda-tower view instead of discrete magnification steps one could do something like “extreme Mercator”: https://mrgris.com/projects/merc-extreme/
or log polar / complex logarithmic:
Joachim Böttger: “Complex-Logarithmic Views and Map Warping.” University of Konstanz 2011, ISBN 978-3-8439-0180-2, pp. 1-106
** Multiple sub-contexts to the current-context would only appear if one goes crazy and provides language features for rewiring to extensional equivalent alternative implementations of the same function (equivalent not just by type but by logical behavior too) but IMO thats should most definitely be handled by standard conditionals with hard-coded redirection input instead.
Related topics on Wikipedia: Zooming_user_interface & Interface_metaphor
(end of part five - last part)
PS: Editing in the third link to the _.svg afterwards (exceeding the two link limit) did work (just once) but editing in a second image did not so I can’t fuse the posts by after-editing.