(I think I've already posted a similar post/topic a few years ago, but this has some changes.)
Just out of curiosity: what's the big idea behind the language you'd want to create? Is it going to be a hobby project or a serious attempt at creating a solution for a wider audience?
Either way, in order for you not making it too difficult for yourself you may want to start building a prototype within Snap!
I guess Snap! is powerful enough to support a very large range of potential primitives of your future programming language. With Snap! as an extra layer, execution of programs in your new language is probably not going to be very fast - but for a prototype that shouldn't be much of an issue.
Perhaps share some ideas and ask others to hone them. Some forum contributors contrived languages of their own; they, and others, may be interested to reflect and think along with you.
Do you have, like, a programming paradigm in mind, and / or an application area? A niche other visual languages don’t (seriously) support?
Not exactly. I want to add blocks (like those in Scratch, Snap!, and Blockly), but I would like to introduce a mechanic that allows you to edit the data of the blocks instead of creating your own or adding extremely complex alternatives. I think it would be fun because the basic commands and functions exist, but you can learn ways to alter them.
Yes, scratch did, create the idea. The first version of scratch was made in squeak, which, I'm pretty sure, is a gui system that is very old and outdated. I don't know the specifics. When they made scratch 2, they made it work in the browser by using flash. When they made scratch 3, I think they collaborated with google to make blockly, which they then used to create scratch 3. Don't quote me on thi, I'm just saying what I think I've heard, I'm not fact checking anything.