Back in the early days, the printed manuals for Research Unix and BSD consisted of two volumes. Volume 1 was the reference and consisted of all the man pages for every command (section 1), system call (section 2), library function (section 3), etc. Volume 2 contained longer documents - what this article calls tutorials, how-tos, and explanations.
The man command let you read all the pages in volume 1. Volume 2 only existed in print, with the troff source in /usr/doc but no obvious way to find it if you didn't know where to look. So naturally volume 2 fell by the wayside. When I was learning Unix in the late '90s and early '00s I had no idea there was supposed to be "official" documentation besides man pages, and filled in the gaps with random web tutorials and borrowed O'Reilly books and other "unofficial" sources.
Nowadays some "Unix purists" are insisting that man pages are all the documentation you could ever possibly need, and if the man page is too long that means the software is too bloated. I find that attitude to be ahistorical. Like anyone's going to learn to effectively use troff and eqn from a cut-and-dried syntax description.
(I could ramble a bit about the other documentation formats that have sprung up to replace troff and how, nice as they can be, they don't replace the convenience of manpages, but this comment is long enough.)
> The man command let you read all the pages in volume 1. Volume 2 only existed in print, with the troff source in /usr/doc but no obvious way to find it if you didn't know where to look. So naturally volume 2 fell by the wayside.
I guess my feeling that man pages were insufficient is not without basis.
Notably, Linux had a whole lot of howtos and faqs—not sure about authoritative sources but I guess The Linux Documentation Project is/was the largest. That was how people learned to do stuff without tearing their hair out, in the early–mid-2000s. I probably still have some of them lying around thanks to stockpiling like the apocalypse is nigh.
Of course, sparse or arcane documentation also leads to proliferation of educating books, by people ready to help for a reasonable sum. The existence of which market should say something about the truthfulness of ‘manpages are enough.’
I've had the cynical thought that Eric Allman made the documentation that came with Sendmail intentionally shoddy to increase sales of the Bat Book. Likewise Larry Wall with older versions of Perl and the Camel Book.
The man command let you read all the pages in volume 1. Volume 2 only existed in print, with the troff source in /usr/doc but no obvious way to find it if you didn't know where to look. So naturally volume 2 fell by the wayside. When I was learning Unix in the late '90s and early '00s I had no idea there was supposed to be "official" documentation besides man pages, and filled in the gaps with random web tutorials and borrowed O'Reilly books and other "unofficial" sources.
Nowadays some "Unix purists" are insisting that man pages are all the documentation you could ever possibly need, and if the man page is too long that means the software is too bloated. I find that attitude to be ahistorical. Like anyone's going to learn to effectively use troff and eqn from a cut-and-dried syntax description.
(I could ramble a bit about the other documentation formats that have sprung up to replace troff and how, nice as they can be, they don't replace the convenience of manpages, but this comment is long enough.)