Written: April 1(!), 2015
I have always hated snobbish journals, who think that only high-brow subjects, like the Langlands Program, are worthy of publication, in other words, papers that in order to understand them, would take ten years of (very boring) study, and that at most ten "experts" in the world can follow.
But even more despicable than so-called "prestigious" journals, are "prestigious-journals-wannabees", who want to increase their "social standing", and start rejecting papers that they belive use "low-prestige" methods, like heavy computations. I regert to say that the Electronic Journal of Cominatorics became such a journal.
I used to love this once innovative journal, that was co-founded (with Neil Calkin), by my good friend Herb Wilf, and in fact, I coauthored (with Dominique Foata) the very first paper, and many other ones, and was a member of its Editorial Board for many years.
I recently resigned in disgust. This happened after a little-known "amateur" mathematician, who requested to stay anonymous for now, (his day job is a software engineer) submitted a paper proving (correctly!) the notorious 3x+1 conjecture. This was rejected based on an obnoxious referee report.
After the author forwarded me the above rejection slip, along with the paper, asking my advice where to submit it, I told him that so-called "peer-reviewed" journals are a thing of the past, and since he is not interested in academic jobs, the best thing would be to just post it in arxiv.org, and he would gain instant fame. He should also make the source-code, and the huge planar map (that he proved would need five colors if 3x+1 were false) in his website, and anyone with modest computational powers can verify the correctness of the claim. Of course coming up with that planar map took months of computations (and ingenious software engineering!).
So, if you are in academia, and still need journal publications in your CV, you would have to put-up with such obnoxious journals, and their henchmen, the "anonymous" referees. But be careful, if the paper is too good, there would be a good chance that it would be rejected. The average referee can only appreciate mediocre stuff that resembles his own research. But if you are a free-lance "amateur" mathematician, or fully promoted, then say NO to "peer-reviewed" journals, that foster mediocrity and the status-quo.