Interesting!
I think the hurdles George Hotz encountered at his first contact with Metamath are very common, especially with people with programming background.
Nowadays there are less limits in naming, e.g. the 8+3 DOS file naming is no more, same for symbols in programming languages.
Actually, some of George’s complaints could be easily worked around:
- Metamath already supports file inclusion, and we could very easily split
set.mm into several files, in several directories, with a hierarchy.
- Nothing prevents us either to use longer, more expressive names, for symbols and for theorem names, except that doubling theorem names would double the database length.