I know this is off Topic, please forgive me
Mark,
You say " they have to be served up via a web-like address" and that is exactly where I am coming from. Personally I can dump images on the internet (my host) and reference them as URL's and load them as needed. However for local only wikis, say my photo library, the next version of tiddlywiki server apparently lets you also serve files, never the less, I am about to test a simple nodeJS or similar file server that can publish files/images on localhost addresses, I could have a configurable prefix say
http://localhost:8088/images/ then if I do publish this on the internet I can change that configurable prefix to
https://domain.com/images/
What I have being thinking is why fight the file:// limitations, just use http addressing.
Then it will import the image to tiddly wiki. Try it.
You could then embed the image in tiddlywiki multiple times with {{export_import.jpg}} or effectivly lazy load using
which is a form of lazzy loading, because it is retrieved and displayed only when the tiddler is displayed
This is actually bad practice with other people images and websites, but if you have your own image repository online, even password protected, it would be fine.
Given all the above I am now looking at a way to import such resource addresses from a directory list so that the tiddlywiki contains the names and folders of selected resources (not the resource them-self) to allow me to browse these resources AS IF I had access to the server file system. The "File" Directory is effectively maintained in TiddlyWiki and gives me the full address to the resource as I need it. Perhaps I can make a fake file and folder directory appear in TiddlyWiki.
TiddlyServer already can do this, see my attachment
Even if "only" displayed, not imported I can transclude a tiddler containing the <img src html, and the image will display.
You can also import files and URL's from using the <$browse> widget.
Too cool if you ask me.