TITLE: When to publish minor results?
QUESTION [35 upvotes]: One of the main things I enjoy about MO and MSE is the chance to solve somewhat difficult problems.  On a couple of occasions, I've found proofs of results (posed as questions mostly on MSE) which seem original and not entirely trivial.  The examples I'm most fond of are short but tricky, not in any way revolutionary, but have a satisfying "aha!" at a crucial point.  
Disappointingly, these questions seem to always disappear into the mists of time.  While my answers to "soft questions" often get enormously upvoted and see lots of views, those answers to serious questions which actually have some original intellectual content often go largely unnoticed.  Once they fall off the top of the question list, they essentially are never seen again.  
Perhaps this fate is inevitable, but it seems a shame to let a cool trick die.  This has got me wondering whether to try to publish any of these.  I'm not a professional mathematician, so this likely would not materially benefit me in any way.  The only goal would be to put the results out in a place where they might be of use.  The options I know of are:

Drop it on arxiv.
Publish in the usual way, in a topical journal.
Publish a note in an expository journal like American Mathematical Monthly.
Submit as a problem to a journal that publishes problems. 

I've published in the Monthly before, and it seems a good place for small results of broad interest.  However, it's not without cost (submitting takes work), it's not totally clear that results published in the Monthly "do any more good" than those rotting in the bowels of MSE, and moreover not all minor results are of sufficiently broad interest for an expository journal to be a viable option anyway.
So my questions are:

What's the best way(s) (in the sense of "being maximally useful to those who may find them interesting") to disseminate minor mathematical results? 
Is it worth the effort?  Or are all minor discoveries bound to languish in obscurity until such time that they are rediscovered and applied to a major result?
Are there any good rules of thumb to determine what minor results are worthy of dissemination (if any)?


A couple afterthoughts:

I've noticed stackoverflow doesn't have the same problem with questions not having any later use, presumably because programming questions are more likely to be shared by many people.
I can give examples of the minor results that inspired this question if requested.
I think this question should be community wiki, but apparently I'm not allowed to designate it as such myself.

REPLY [14 votes]: If you're not interested in self-promotion but are purely thinking about how best to disseminate your results, then I think that we're at a point in history where the question you should be asking isn't "Where should I publish this?" but "What can I do to maximize the probability that when people in the future are in need of this idea, their search engines will find it for them?"
One of the motivations for the Tricki was to serve as a repository for "useful tricks."  The Tricki has had limited success IMO but I believe it is a step in the right direction.
I looked at the two examples you provided and it seemed to me that in both cases, even if you went to the trouble of formally publishing the results, the chances of someone discovering your paper with a search engine are not as high as one would like, because there's no obvious keyword or search term that they would use.  Thus, the work required to publish the result doesn't seem quite worth it, unless you also find some way to solve the problem of making your work discoverable by search engines.
Of course I'm making an assumption here—that for example you would want your answer to fedja's question regarding a proof of Enflo's theorem to be discoverable by more people than just the small number of people who are interested in different proofs of Enflo's theorem.  Under that assumption, I think that it might be fruitful to ask yourself, "Are there other results similar to mine that are scattered across the literature and the Internet, and if so, what is a good organizing principle that would make it easier for people to navigate this corner of the mathematical universe?"  Obviously, this takes way more work than simply publishing your result, but the benefits will be substantial if you succeed.