Status quo
We currently have seventy-four questions tagged palindrome, but most of them contain a twist. As I see it, there are only two challenges about the basic palindrome string decision-problem.
While the first one is a basic decision problem challenge, it excludes punctuation, whitespace, control characters and letter case such that palindromic sentences are considered palindromes. To me, those are very human language specific exclusion and I do not feel like this challenge captures the core property of palindromic strings.
Furthermore, it only asks for a "method bod[y]" -- not conforming to our current standard method of determining byte count.
While the second challenge linked to above actually defines palindromes as reading byte-for-byte the same from both directions, it is also a restricted-source challenge -- which, to my eyes -- seems to be the far greater task.
There is one other challenge with a palindromic theme I found -- which is not tagged palindrome --,
which also is a restricted source challenge. Furthermore, this challenge focuses more on two-dimensional matrices rather than string palindromes (even though the palindrome tag description could be interpreted as also covering this challenge).
A new challenge
Based upon all of the above, I think it would be appropriate to have a new palindrome decision-problem code-golf challenge which simply asks to decide whether or not a given ASCII string is a true palindrome, as in being byte-for-byte invariant under horizontal reflection.
It would be about nothing more, nothing less -- no restrictions on the source or other interpretations of what a palindrome is. It would also comply with our current byte count and I/O standards.
While this challenge will be trivially answerable in a lot of languages, there are definitely (especially esoteric) languages in which palindrome decision is no simple task.
A pure palindrome decision program in such languages, however, is -- as I see it -- currently not a valid answer to any question.