Every published book - RPG or otherwise - has errors. You can't fix that. Best you can do it mitigate against errors with more time and resources allocated to editing and proofing.
Honestly I think that's BS. You can totally have an error-free book. It's going to cost you more, but you can.
Is there a compiled list of typos, errors, and omissions anywhere? I don't really want to browse through that whole thread to find them. At a glance, most of them seem to be mis-named spell references and the like.
I'd like to have such list too, it should be easy to build it collectively and post it on ENWorld as a reference for everyone.
One would think, but I'll gladly eat the next 1st print RPG book set in front of me with zero errors.
It's a game book. There are medical and industrial manuals whose accuracy have lives (and lawsuits) depending on them. Errors still get through.
RPG seems to be more prone to errors. I own tons of scientific books, and despite their sheer size the number of mistakes is much lower, and in many cases it's
zero.
But clearly, a serious books has
factual information, while a game book has
fictional information, so sometimes it might even be impossible to say if something in the RPG book is right or wrong, and only the writers know that.
And because it's made-up stuff, it gets re-designed and re-written many times before publications, and there are multiple people working on it (even if splitting up different chapters, they might get related things differently in those 2 places). This highly increases the risks of mistake. Just think about it: they wrote some playtest document 6 months ago, the rules for e.g. mounted combats or hiding are spread to 3-4 different locations, then they playtest and adjust the rules a couple of times, and someone may not remember to update one of those locations...
Another reason is, they are doing verifications on their own written products, which is the easiest way to overlook mistakes. They might have a proofreader or two, but is probably someone focusing on grammar & language stuff, not someone especially knowledged in the rules of the game.
In theory, they had almost 200.000 proofreaders in the last 2 years. However, all the changes in the last 9 months were seen only by a few hundreds/thousands of closed-group testers, which likely focused on testing how the game worked, not the written docs.
Had they shown the rules to those 200000 playtesters before publications, we would have got rid of all the mistakes, but obviously this isn't the best thing to do business-wise. Also it would have needed to be
much earlier, to account to all the delays in re-editing and sending to the printers. And it's very likely that if they had 1 more month delay, they would have been tempted to sneak in some more changes, with their relative errors.
Unfortunately this always means that the real proofreading starts when the customers buy the books. If the mistakes are too many, they will be fixed in a future re-print, if they aren't many there will be no updates to future re-prints.
But all this is just to explain why games always have much worse editing than scientific books, not to justify the traditional bad quality of a RPG book. It
can be error-free, but between people wanting to have the game asap, designers needing to have it ready to sell it, publishers wanting to minimize cost, and everybody wanting a completely new edition built from scratch every few years, eventually it's damn hard to lower the error rate.
If it's really only 8 mistakes in the whole 5e PHB, it's bordering the miracle...