> Why should I put any value in what W3C recommends?
For the same reason you put any value in, e.g., ISO or ANSI. Because others recognize them as the standard organization and put value in what they recommend.
The W3C citations on MDN are a historical artifact by the documentation folks, not influenced by the browser folks. There's a slow movement to consolidate around what's implemented and stop referencing W3C forks, and I think today's news will help hurry that along. See e.g. the discussion in https://github.com/mdn/kumascript/issues/1019 or my previous work in https://github.com/mdn/kumascript/pull/220.
If I complain to Mozilla that they deviate from the W3C spec in some respect, will they treat that as a bug, or will they more likely say "no-one implements this, so no website uses this, so we have no reason to be the first to implement this because we don't think this is part of the spec is really important"?
Depends on the deviance. If it's 'this was changed in the whatwg spec' the answer will probably be 'the next w3c spec will document this change, for now here's <an explanation and/or an update to mdn>'.
If the deviance breaks production apps sometimes Mozilla will roll it back, it's happened before.
In other cases they'll treat it as a bug and fix it.
Although I still worry about the case where fixing the deviance would break production apps. Potentially break them, since people won't complain until you release the breaking browser version. That's what's usually kept browsers from hewing to standards.
I think it's more misleading to claim that any written standard has authority. In practice these written standards are ultimately subservient to the actual implementations—particularly where the competing implementations happen to agree. And especially in disputes that affect real websites.
If there is an inconsistency between what the standards say and what the browsers have done, the browsers will more than likely trump the standards. In all practical senses the standards are descriptive, not prescriptive.
For HTML and DOM, this is true. There are many more things at the W3C, the most obvious of which is probably CSS. Browsers totally pay attention to what these spec says, and are the dominant participants in producing these specs.
For the same reason you put any value in, e.g., ISO or ANSI. Because others recognize them as the standard organization and put value in what they recommend.