Ah, I remember that I linked to the package in my original
message, but indeed I do not find any link there (let me see if
this works, otherwise the package can be found at:
https://github.com/dalanicolai/sketch-mode).
Anyway, although I do not immediately see any big advantage of
having the image overlayed on a 'semantically corresponding
text', it sounds like an interesting idea. Maybe you can explain the
advantages in a little more detail.
I was thinking of creating a togglable side-window that can show various
'levels` of definitions (layers, objects etc.). It is not very useful to get the
full definition of the svg image, as it is 'polluted' with the definitions of
the grid and the object's labels. So a sketch-mode image is build up modularly
using svg-groups, where the grid and labels and the different layers are added
as different modules/groups. So when I overlay the image on semantically
corresponding text, I would not immediately know which part of the definition
best to use for the text (the svg definition, the full elisp image definition,
only the layers?)
Indeed, the package is only an isolated editor emulated in Emacs, but it has
certain advantages over more specialized software especially for simple svg
images; i.e. it is in Emacs, sketches can be created really quick and finally it
is an Emacs package and therefore ultimately/easily hackable. So users can
implement whatever they like, e.g. snippets, and use elisp functions for
drawing certain patterns etc.
I have written a little motivation and vision in the
sketch-mode wiki
Daniel