Ah, I see. I misunderstood your comment. I thought that by
links to something deleted lead to previous versions of the documentation
you meant
links to something deleted lead to previous versions of the documentation (of that thing)
but you meant
links to something deleted lead to previous versions of the documentation (of python)
Although in this case it links to the current version of the python documentation, not the previous one.
With that ambiguity resolved, I still think that this is a suboptimal way to organize documentation.
I think that this is a weak argument. Python might not be backwards compatible, however the Python documentation does not need to follow those rules exactly. I don’t see, why the docs can’t include a “stub” cgi.rst
file that simply informs the user that the cgi
module no longer exists and then links to the 3.12
version of the cgi
documentation.
The current documentation already includes historical notes for individual functions/features that are no longer present in the current version of Python. For example, http.server — HTTP servers — Python 3.13.0 documentation currently contains the following note:
And venv — Creation of virtual environments — Python 3.13.0 documentation has this note:
My understanding is that both of them are caused by deprecated-removed::
directives in the rst
and the difference in wording (“will be removed” vs “removed”) is automatically adjusted based on the python version.