Hi,
I’d like to bring this subject on the table as it seems important and not decided yet.
The basic reason we need image dimensions flags is that a standard cannot rely on definition only but has to include resolution too, and both at the same time if possible.
The world is going high dpi, (or retina for fruit people), and the quality of a document should not rely on a design based on one only (old) type of screen device.
The same way you are pleased to have vector quality for your fonts, why do you want to force the image item type to be stuck at a basic level of quality ?
Example: I want to include a screenshot in my markdown to show how to use a GUI, I don’t want to have to reduce the quality of my GUI screenshot to 640x480 where it becomes unreadable just so it can fit into my markdown page, a simple example like this should be enough to prove the current limitation of the image markdown support.
Ps: The dimension feature should not rely on percentage because the layout needs to be source quality/definition agnostic.
here my humble syntax proposal:
 for setting width / height at the same time
 for setting the width and auto-ratio for height
 for setting the height and auto-ratio for width
Thanks for reading me, I hope it can help and make things moving the right way.