I actually always supposed that Melbourne's publishing of the Spectrum ROM was by consent of Sinclair Research; after all, it greatly helped software developers to make new software titles and devise even greater hacks for the machine, thus making it more attractive for buyers. The ROM did display a copyright message, so it should have been clear that it was Sinclair's property.
I recall I wrote my own loader for an adventure game that I designed; it was so nice to have a slightly different loading sound and different colours flashing at the sides of the screen. The loader of course was based on the ROM routine, but with different timings (to increase the loading speed from 1500 bit/s to about 2200 bit/s). The game was probably not very imaginative, and sold less than a dozen copies, but it later got me a job with the local Spectrum/Amstrad distributor. Alas, I've lost all copies.
Yes, it was a very different environment. Here's one for the BBC Micro:
http://www.computinghistory.org.uk/det/6637/BASIC%20ROM%20Us...
From other details on the page, it looks like the publisher, Adder Publishing, had the tacit approval of the manufacturer Acorn.
Disclaimer: the author is a personal friend.
Said manufacturer Acorn later designed their own CPUs, and spun off a subsidiary called "Acorn Risc Machines" to market those chips, later renamed ARM. You people may have heard of them....
The fact that they could publish a book like that is a nice reminder of how much the environment surrounding copyright/IP has changed since then... try to do that with the BIOS/firmware of a modern system today and, the issue of it being many times bigger aside, you'd probably just get DMCA'd.
I suppose companies back then were just far more open about and approving of users knowing how their systems worked; there's a similar book for the C64 (http://www.amazon.com/Complete-Commodore-Disassembly-Peter-G... ), and IBM even published the BIOS source code for its PC.