Hi daimip
Your proposal is admirably clear and well explained. I would love to be able to implement it, but some of those capabilities remain out of reach with current browser capabilities:
* When dragging a file into the browser, JS code only sees the content of the file, and not it's path. Thus there is no opportunity to link to files in their original locations
* When saving changes, the default HTML5 fallback saver only permits a single file to be saved at a time. There is no convenient way for the user to control saving the subsidiary, attached files
Tantalisingly, a lot of your proposal does already work out of the box. Using the `_canonical_uri` feature makes it possible to reference images, PDFs etc as "external tiddlers" which reside in separate files and are only loaded when required for display. We can easily support operations such as downloading attachments. The obstacle is around the input of attachments.
We can do a lot better when using something like TiddlyFox or TiddlyDesktop; both of them support unrestricted file access.
Best wishes
Jeremy.