I know in this community the overloaded value for vanilla in this context means just javascript without other people's javascript. But "vanilla web" should really mean HTML and CSS.
Sure, so which subset of HTML are we talking about?
Because unless I'm misremembering things, HTML 2.0 (literally the first version intended to be a standard going forwards) came out exactly the same year that JS dropped in a commercial browser (1995), and superseded the previous HTML and HTML+ markups.
So surely you don't mean HTML with things like file uploads, or GIF support. Because that's not "vanilla" in your world.
In the software sense, vanilla means "in it's original form, without addons, modifications, abstractions, third party libraries, ...".
Like a "vanilla kernel" is a Linux kernel without (distributor) patches. Or a "vanilla debian" is a Debian system without third party repositories or software.
Regarding parent, I could say... you cannot make a permanent To-Do web app, with HTML+CSS... drag and drop? data persistence? you need a backend (and JS for event handling) or as in this case local storage via JS.
OK, you can make a simple To-Do with static forms being sent to a backend... but then you have HTML+CSS+something (java, python, php, ruby, perl, golang, whatever), not only HTML+CSS.
