I would flip everything around: At scale, doesn't a private enterprise run out of money to support large codebases?
This isn't a facetious suggestion. Every employer I've had in the industry has been drowning under their heavy codebases. They have massive amounts of vendored dependencies, and long tails of single-author modules. The bigger employers have to create special teams just for pruning codebases and cleaning up after old projects.
And they're losing the battles. Code sprawls between multiple cloud vendors. Services exist in ecosystems so large that they have echo chambers, network effects, and politics. New languages are slowly introduced and even more slowly removed, with most code being written in mid-level glue modules. Code ontology and navigation is primitive and limited to indexed search.
What makes FLOSS different? Political discussions have to bring technical justifications to the table, mostly; this is Linus' famous "talk is cheap, so show me the code" slogan. The openness of collaboration ensures that CI systems, code repositories, and other shared services are available to folks, although this is still a cultural situation that isn't quite uniform yet. (There's too many folks who only publish tarballs and don't share the rest of their infrastructure, which is fine but not very collaborative.)
The resulting ecosystem can directly and openly compare codebases, and deprecate code which is shown to be inferior or unsafe or otherwise undesirable. We have shown an ability to not just change code artifacts, but the protocols and languages which we use to communicate, at a very rapid rate.
Rust, Zig, IPv6 in Linux, Wayland, EGL, PulseAudio, PipeWire, and of course BusyBox. Compare and contrast with the fact that the banking industry is still searching for COBOL maintainers.
This isn't a facetious suggestion. Every employer I've had in the industry has been drowning under their heavy codebases. They have massive amounts of vendored dependencies, and long tails of single-author modules. The bigger employers have to create special teams just for pruning codebases and cleaning up after old projects.
And they're losing the battles. Code sprawls between multiple cloud vendors. Services exist in ecosystems so large that they have echo chambers, network effects, and politics. New languages are slowly introduced and even more slowly removed, with most code being written in mid-level glue modules. Code ontology and navigation is primitive and limited to indexed search.
What makes FLOSS different? Political discussions have to bring technical justifications to the table, mostly; this is Linus' famous "talk is cheap, so show me the code" slogan. The openness of collaboration ensures that CI systems, code repositories, and other shared services are available to folks, although this is still a cultural situation that isn't quite uniform yet. (There's too many folks who only publish tarballs and don't share the rest of their infrastructure, which is fine but not very collaborative.)
The resulting ecosystem can directly and openly compare codebases, and deprecate code which is shown to be inferior or unsafe or otherwise undesirable. We have shown an ability to not just change code artifacts, but the protocols and languages which we use to communicate, at a very rapid rate.