iframe@srcdoc

Hi everybody!
I tried an experiment with initial content for iframes. But @srcdoc seemed
really difficult to implement and yet so incomplete. Why has it been chosen
to rely upon an attribute to integrate markup content inside an element?
Isn't it confusing? it needs that a user agent/data mining tool/fetching
script is aware of its peculiar syntax (quote escaping, double escape for
ampersand and reserved characters in prose, resolving entities to obtain
markup) and IMO it also has other problems. Has anyone else ever had
similar doubts?
Maybe it has already been discussed, otherwise other solutions could be
found.
Thank you!