Great piece! Reminds me of my old project https://github.com/rnd0101/eternal , where I am trying to design a system, which can preserve old programs over hundreds of years, using easily bootstrappable programming language, which can then be enriched, and given the code, old data can be revived. Now it seems to me that anything too artificial will not survive. Only ideas, which were discovered mathematically can give raise to such language.

As for the need for more languages and such, I think the movement is not purely to preserve the best from the old languages, but also to add usability to programming languages ( for starters: https://en.wikipedia.org/wiki/Cognitive_dimensions_of_notations ). Modern PL design is not just about it's technical characteristics, but humans are taken into account too.

There probably are some limitations. Majority of software developers work with some ready frameworks, and do not bother (or lack abilities) to boost their productivity with meta-tools like sport their own DSL.

I've recently wrote an article on frameworks ( https://medium.com/star-gazers/framework-as-a-way-to-painless-software-development-c3382f35a334 ), so answering your question regarding DSL vs framework: They are NOT opposites, they are complementary. Some problems require DSL, some frameworks can benefit from DSLs. I think, the angle should be: As a developer, can I notice when I need to make my own mini-framework AND/OR my own DSL to solve some specific domain problems elegantly and more efficiently?