Quoting Clarke Sideroad via Dng (dng@???):

> I think the purpose was to avoid tainting the initial installation
> with proprietary firmware, but not have the lack of it it ruin the
> party.

Proprietary firmware files divide broadly into two categories:
1. Copyright owner granted sufficient permission for distros to
redistribute them.
2. Copyright owner didn't.

You might think it would be really incredibly stupid for say, chip
manufacturer Broadcom, Inc. to just totally forget to write down 'Oh, by
the way, anyone may freely redistribute this firmware file without
modification provided that no right to reverse-engineer it is granted
and no warranty of any kind' -- and you would be correct. Yet, that is
exactly the mistake Broadcom and a number of other such manufacturers
repeatedly make. I've been known to mock Broadcom, Inc. on this matter,
observing that they were too busy putting profits up their nose to find
first-class postage. (Web-search 'broadcom cocaine' for details.)

This is why quite a number of firmware files simply cannot (lawfully) be
included in Linux distributions, for lack of someone lavishing the cost
of a first-class postcard on the matter. Typically, instead, someone
crafts an open-source 'cutter' utility for *ix that, if installed and
run, downloads the MS-Windows driver .exe or .cab files from the
hardware manufacturer's Web site, extracts the firmware BLOB image,
parks that with the desired filename under /lib/firmware, and discards
the rest.

See, for example, https://packages.debian.org/stretch/b43-fwcutter
(speaking of Broadcom).

Anyway, don't assume that a distro's omission of desired firmware is
necessarily on grounds of ideology. It might be because of
nose-injuring incompetence from firms like Broadcom, Inc.