Firstly, it’s a shame to hear you are thinking of abandoning TW.
(Note: I use WikiText instead of Markdown so you may have to convert to your syntax.)
For my part I reference most images using either <img>
or [img[]]
. I will usually do it the HTML way if using in combination with figure
and figcaption
, and the WikiText link way on most other occasions.
If you chose to move away from TW, the chances are that the next tool would not be truly automated, but I would also say that the format is not that complicated either, and it can be programmatically converted without much effort.
Standard HTML should be supported within most Markdown compilers, theoretically (not that it’s a real standard, but still).
The same cannot be said for Markdown-based tools, so you should probably avoid using HTML directly.
As to option #2, images-as-tiddlers would probably be less ‘future-proof’. I only create tiddlers for content that have meaningful content, such as magazine scans. If it’s just pretty headings or illustrations, they don’t need unique tiddlers. I cannot think of many other tools that follow TiddlyWiki’s model for storing information, so you might struggle with this style. Use it when it is meaningful in your data structure, but really consider if it’s necessary for them to be tiddlers.
My key question is if I could ‘consume’ the tiddler standalone to some extent. If not, then it should be an embedded img
.
When I do have images as tiddlers, I always do them as _canonical_uri
and follow a loose folder hierarchy (I don’t sweat this since hierarchies don’t represent the TW way).
Above all, I’d also say not to let the idea of possibly having to migrate put you off from using TW to its full potential.