Good link. But for those who won't bother reading[1], that paper doesn't support the Fortran assertion. Fortran is only mentioned twice in the HOPL conference paper, once regarding syntax familiarity (lack of semicolons), and again in the context of FFI. In neither case is Fortran mentioned in a privileged manner, but rather alongside multiple other languages to highlight non-C, non-C++-specific considerations and motivations.
I know there are Fortran shops that make profitable use of Lua as both an embedded extension language and as a glue language. But Fortran users do so in the same way Lua is used with C, C++, and other languages. That merely reflects that the Lua creators had a good understanding of real-world software ecosystems, and successful in how they shaped the design and implementation to minimize impedance mismatches between Lua and other environments, across syntax, semantics, API, and various implementation details that are invariably ignored or overlooked in comparable languages.
I know there are Fortran shops that make profitable use of Lua as both an embedded extension language and as a glue language. But Fortran users do so in the same way Lua is used with C, C++, and other languages. That merely reflects that the Lua creators had a good understanding of real-world software ecosystems, and successful in how they shaped the design and implementation to minimize impedance mismatches between Lua and other environments, across syntax, semantics, API, and various implementation details that are invariably ignored or overlooked in comparable languages.
[1] For those who like reading, see https://www.lua.org/docs.html#papers