tl;dr: I don't necessarily think this is going to work well, but support anything that might engage people around here, ultimately, if it works, great, if it doesn't, theres no lasting damage and it can easily be removed.
I thought, given that I'd tried to make 'synthesis golf' a thing, I should comment on 'salt golf'. In general, one of the big concerns with synthesis golf was that it (perhaps unfairly) necessarily excludes a certain proportion of the users active here on chem.SE (though looking at the tag usage, and general questions, I think its fair to say that organic chemists are over-represented generally amongst active users).
At the time of writing the FAQ for synthesis golf, I talked about possibly expanding the organic nature of it to inorganic/polymer synthesis, but also commented that other things (computational/spectroscopy/nmr-golf) could also work well.
The big question then becomes whether 'salt-golf' is one such thing that would work well.
Scope:
My main concern with this 'salt identification challenge' is the scope.
Synthesis, by definition, is quite fluid, and as such any given problem will have an infinite number of solutions (none of which can be identified without some knowledge of reactions/previous synthetic work).
Salt identification, on the other hand, is rather formulaic (and arguably doesn't really need knowledge so much as a book on qualitative analysis). Any given salt identification question will necessarily have only one answer, which can be easily arrived at by using the same flowchart that we all used as undergrads doing inorganic labs. Vogel's Qualitative Inorganic Analysis has all the information that could ever be needed (as does SciFinder, but arguably its not quite so straightforward to apply the junk results that SciFinder yields), and as such I worry that essentially one person will answer, and nobody else would bother because 1) the right answer has already been found and 2) necessarily anyone else trying will just arrive at the same point (albeit by a slightly different order of reactions).
Purpose:
A second, and perhaps unfair concern is that nobody really does qualitative analysis now, modern techniques such as elemental analysis and mass spec have vastly superseded it, and as such it becomes quite a trivial issue.
Level/audience:
The flip side to my arguments above is that actually, salt identification could in theory be answered by anyone who's willing to invest some time into doing it, as theres no fundamental skill/knowledge involved other than being able to think analytically. For this reason, it could be a way of getting certain users engaged who otherwise wouldn't be able to answer many questions here.
Conclusion:
Overall, just post a question and see what happens. I'd perhaps wait to see how it goes before committing to making 'salt-golf' a thing (i.e. just post a question and see how its received. I dont actually think the question you linked to as an example of this is a particularly great question, it very much falls into the 'amirite' category, but thats subjective on my part, especially not knowing much about the topic other than nightmares from inorganic labs in my first year of undergrad with the 'what is this white powder' challenge). My feeling about these things is that they're easy to try, and easily removed if they flop, but they could just do something great and help get people engaged and answering questions.
Good luck have fun!
