A way out of this crisis could be an Enterprise Linux Standard, a document
describing in detail a standard installation, together with a vast set of conformance
tests.
It would be a supremely boring project, but if interested orgs such as suse, oracle, ibm etc suport it, it might be feasible.
It could start from a description of rhel, but ideally could contain some extension
of the file system standard to allow compatibility with debian.
Then, you would have a reference platform that multiple distributions could provide and there would be competition, not just one major provider of a system `as is' and some more or less incompatible others.
Yes, it would continue POSIX. But POSIX is about a small part of a linux distribution.
There was LSB that was more extended, a bit. Then there was the file system standard,
and perhaps others that I don't know about. But a standard covering what one should expect
from a modern 'base system', in full detail, updated every 10 or 5 years, does not exist. There is a de facto `standard', RHEL, but it is by fiat. A proper standard requires a documentation (the `standard') that defines what is essential to be found, _and_ an implementation--or, ideally, more than one, _and_ a means to check that a given implementation matches that documentation.
If you have that documentation, it works both in space (alternative implementations) and time. In the future, you can be sure that a new implementation, perhaps of a future version of the standard, is still exactly compatible with the previous version.
This has been done for languages (C, Ada, Lisp, Fortran etc), libraries and kernels (that is POSIX), processor ISAs, but not for linux distributions AFAIK.
It would be a supremely boring project, but if interested orgs such as suse, oracle, ibm etc suport it, it might be feasible.
It could start from a description of rhel, but ideally could contain some extension of the file system standard to allow compatibility with debian.
Then, you would have a reference platform that multiple distributions could provide and there would be competition, not just one major provider of a system `as is' and some more or less incompatible others.