I think you're confusing the "reliability" of a source (in the peer-reviewed article sense) with its reliability as an indicator of the notability of some subject X.
The peer-review system basically establishes (a) the reviewers believe the result is non-trivial and (b) the reviewers believe the result is correct.
However, I'd say that the very existence of documents such as third-party tutorials/introductions/discussions, etc., can be adduced as evidence of notability. Think of determining notability more as "sociology" than some academic judgement about the intrinsic worth or originality of the topic.
For instance, "article citations as notability" is patently absurd in the case of, say, celebrity or TV show articles (the existence of both of which seems accepted on Wikipedia) or even news topics, and only somewhat less so in the case of PLs, especially those developed outside of academia. (It's arguably a bad criteria anywhere where results aren't very expensive to reproduce, such as population surveys and lab results -- surely many important results in math and physics living only on the arXiv are more notable than some minor topic which generates higher citation counts through constant re-citing by the same group of devotees).
In contrast, the three languages mentioned on the front page were all ones I'd previously heard of.
I see there is a paragraph in the Wikipedia guidelines for notability regarding refereed paper citations, but in line with my above comments I'd suggest that this not be read too literally. Perhaps it's personal preference, but I find such "long tail" articles useful, even if that means WP contains millions of articles on topics I personally find irrelevant.
Perhaps the reason you've incited such anger is that people feel you're imposing what can be seen as elitist and subjective views about how notability is defined (certainly it's not true that everyone's pet/undergrad PL project should have a page, but the three languages mentioned certainly have received wide attention). From this perspective, the question here is not about any particular property of these PLs but what the WP criteria are or should be, so it would certainly be best to err on the side of non-deletion in all cases.
I hope you reconsider your views on this topic (disclaimer: I am not associated with any of the projects mentioned).