Change 257624 had a related patch set uploaded (by Ricordisamoa):
[POC] Implement NOANCESTORLINK magic word

https://gerrit.wikimedia.org/r/257624

In T41395#2272004, @Jdforrester-WMF wrote:

Adding new behavioural switches adds complexity for users (new things to remember), for software (new checkboxes/etc. with which to prompt users), and the servers (more complex parsing), so it needs to pass a pretty high bar in terms of value when it comes to their features. Though I see the two use cases given (in the description and in the merged task), I'm not sure this is sufficiently valuable to do, and am minded to Decline. Thoughts?

IIRC there was some request for Tabs extension, so the first usecase would be better handled by the extension itself, which would take care of disabling the autonavigation. (Speaking about extension, how does actually Breadcrumbs extension deal with that?) Beside that, if all wikiprojects do use the same style of navigation, there can be easy CSS written for that. CSS can be written even for certain list (but that's not scalable though). Javascript gadget could also work...

The latter example could be solvable by JS gadget. Or with CSS like above, though in this case it will be the second alternative - un-pretty-scalable growing list.

From the accessibility point of view, the navigation should be consistent across all pages on site, so introducing various different ways is rather disruptive.

Also, all usecases basically dealt only with first level of subpages, but the automatic subpage navigation automatically links all the way down.

Given the reasons above, I am also in favour of declining of this.

This item has been referenced on the English Wikipedia's "Village Pump (technical)" page: https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical)#Way_to_disable_link_to_base_page_and/or_root_page_on_a_subpage?

That brings up perhaps the most important use case: the talk page of a mainspace article with a slash in it, such as [[Talk:AC/DC]], [[Talk:n/a]], [[Talk:R/K selection theory]], [[Talk:9/11 Commission Report]], [[Talk:S/PDIF]], etc.

A more general fix would be to automatically disable it for pages in a talk namespace when (a) the corresponding content namespace has subpages disabled and (b) a page with a matching title in the corresponding content namespace exists.

In T41395#6308500, @Ahecht wrote:

That brings up perhaps the most important use case: the talk page of a mainspace article with a slash in it, such as [[Talk:AC/DC]], [[Talk:n/a]], [[Talk:R/K selection theory]], [[Talk:9/11 Commission Report]], [[Talk:S/PDIF]], etc.

A more general fix would be to automatically disable it for pages in a talk namespace when (a) the corresponding content namespace has subpages disabled and (b) a page with a matching title in the corresponding content namespace exists.

If I'm reading this correctly, the issue I see with approaching the problem in that manner, specifically via resolution "(b)", is that it would erroneously disable pages when viewing talk page archive subpages. For example, when viewing [[Talk:AC/DC/Archive 1]], the link to [[Talk:AC/DC]] would be disabled ... when viewing [[Talk:9/11 Commission Report/Archive 1]], the link to [[Talk:9/11 Commission Report]] would be disabled ... etc. In other words, such a resolution would create more false positives than fixed pages since it would even disable parent page links such as ... when viewing [[Talk:NASA/Archive 1]], the link to [[Talk:NASA]] would be disabled ...

In T41395#6308522, @Steel1943 wrote:

If I'm reading this correctly, the issue I see with approaching the problem in that manner, specifically via resolution "(b)", is that it would erroneously disable pages when viewing talk page archive subpages.

I think you misunderstood. The intention of (b) would be to disable the link if an article exists that matches the current, not the target, talk page. In that case, the link to from [[Talk:AC/DC/Archive 1]] to [[Talk:AC/DC]] would not be disabled unless there is also an article called [[AC/DC/Archive 1]], and the link from [[Talk:NASA/Archive 1]] to [[Talk:NASA]] would not be disabled unless there is an article called [[NASA/Archive 1]]. However, the link from [[Talk:AC/DC]] to [[Talk:AC]] would be disabled since the article [[AC/DC]] exists.

In T41395#6308658, @Ahecht wrote:

In T41395#6308522, @Steel1943 wrote:

If I'm reading this correctly, the issue I see with approaching the problem in that manner, specifically via resolution "(b)", is that it would erroneously disable pages when viewing talk page archive subpages.

I think you misunderstood. The intention of (b) would be to disable the link if an article exists that matches the current, not the target, talk page. In that case, the link to from [[Talk:AC/DC/Archive 1]] to [[Talk:AC/DC]] would not be disabled unless there is also an article called [[AC/DC/Archive 1]], and the link from [[Talk:NASA/Archive 1]] to [[Talk:NASA]] would not be disabled unless there is an article called [[NASA/Archive 1]]. However, the link from [[Talk:AC/DC]] to [[Talk:AC]] would be disabled since the article [[AC/DC]] exists.

Yep, I definitely misunderstood, and you clarified what I didn't understand. Thank you. What you stated makes sense, and I agree with this suggestion being a possible, and highly probably, approach.