I don’t get it. Why limit yourself like that? My coworkers and I have never have any problems with using .ai to generate deliverables in the process of building an app.
Do you mean to draw the logo in codes? Because if something can be drawn in illustrator, it almost always can be drawn natively in an app too, though it might not worth time doing so in some cases.
The example logos in the article are mostly from the time before people use computer to design, so it makes sense that the logo have to be optimized for reproduction in different materials that way. Today’s laser cutting machines and super large-format printers allow for much more flexible designs.
The vector file of a logo is essentially a map for its creation, just like those old grids photo you show. It really is just a list of points sitting relatively to eachother in 2d space with no solid measurement.
Maybe I missed something or need that second cup of coffee (in which case please point out to me so we can both laugh at my foolishness), but why going backward? Modernism in graphic design is celebrated for a lot of reasons, but I’m not convinced by the one you mentioned…