I did not in fact attempt to contact the original author as I did not feel that this was the appropriate way to comment on his paper (see my earlier posts!). However, discussion with the original author could indeed a viable option within a somewhat looser commenting structure, but does require that he/she is contactable at the same address after maybe over a year, is willing to engage in discussion, and is amenable to publishing a correction should this become necessary. Assuming all this is possible, direct contact is perhaps on a par with discussion in the Forum, lacking the wider view of other members but (hopefully) gaining from the author’s insights. A combination of the two is perhaps the best solution if actual publishing is not to be used but does, as you say, risk a protracted discussion.
Indeed, direct contact is what I felt the inadequate responses I had received from the Editor forced me to do when I wished to comment on a later JBAA paper by Jean Meeus. My remarks were met with Jean’s usual courtesy, and he graciously agreed that some phrases he used could have been worded slightly differently to make his points more accurately. However, it is most unlikely that he would be keen to publish these minor corrections himself and of course if I were to do so my contribution would have to surmount the hurdle of being “of interest to the Editor”. While this is of no great import in this particular case, if a significant deficiency was identified in a paper but was unable to be fed back into print then the consequences could be more serious.
I therefore feel (and it seems you do too) that, whatever the means used to facilitate the discussion, there is a need for a mechanism by which comments on papers (whether by the original author or not) can be freely published in the Journal. If not in the Letters section, then where? I was certainly not suggesting that ArXiv be used, as that would throw up all sorts of problems of its own, but maybe the Journal should permit of a contribution which is part-way between a Paper and a Letter? i.e. something which is only lightly refereed, maybe just by the Papers Editor (allowing it to be quickly published), and does not have to fulfil any particular “interest” criteria – its interest being inherent in the fact that it is a comment or correction on an already-published paper.
Are we heading in the right direction here, do you think?