Iconography question

[eric-schuh]

Discussion in the VCALM threat model:

I am not a fan of the shapeography / iconography used here. The folder concept is from the 90s, having objects in the folders makes little sense. The external actors and objects are hard to differentiate. The objects are marked with a "D" instead of an "O". It's not clear which object is used in which flow. The right-most line in the objects is missing for some reason. The storage mechanisms are not defined in the DFD. The nested container boundaries are difficult to see. Accessible text for the diagram is going to be a nightmare to understand for someone that does not have the ability to see. The diagram probably does not work well in dark mode. It's not scalable nor does it have an SVG alternative. The trust boundaries difficult to make out.

None of these are blockers to merging the diagram, but I do think the Threat Modelling group needs to rethink it's visual language.

Originally posted by @msporny in #655

I strongly recommend we stick with the DFD iconography defined by Adam Shostack https://github.com/adamshostack/DFD3

Defining our own icons will just make it harder to have functional interoperability across the W3C.

Originally posted by @jandrieu in #655

I don't see anything here that resembles a folder.

That said, I have nothing against using the files-and-folders (including nested folders) analogy, and think it can be put to good use, even today, even if nothing in the UI/UX of any eventually implemented systems will look anything like files-and-folders.

The age of a system does not automatically define it as useful nor useless. I believe Monsieur Jacquard would have things to say about the value of punch card systems even today. Likewise Otlet and La Fontaine about interconnected index cards.

Originally posted by @TallTed in #655