I've heard from John Vig (http://johnvig.org/ [1]) that the IEEE is considering adding a feature to their online paper archive which would be "like a blog at the end of each article, i.e., to allow people to comment on possible errors in the paper, references that should, or should not have been included, etc."

I think it might be worthwhile to add a comment system like this for ISCA papers.

Dr. Vig mentioned some valuable potential uses, such as posting corrections. Another potential use is publication of disappointing follow-up results. Some senior people in the speech technology community have expressed concern about under-publication of disappointing results. I have personally seen several cases where positive initial results for a new technique were submitted for publication, but disappointing results discovered in follow-up work (by the original authors or by people who read their paper) were not. I believe that two key aspects of this bias are that it can be hard to make a competitive full-length paper out of nothing more than a disappointing result, and that people feel less motivated to put in the effort and time required to create a full-length paper when the results are disappointing rather than pleasing. It seems plausible that an online comment system could greatly help with these two aspects of the bias problem, because comments could be of any length and would not have to out-compete other work in a competitive peer review process. I think that the ability to make comments of any length is an important difference from the Journal of Negative Results in Speech and Audio Sciences which was launched by CMU in 2004 and did not receive many submissions.

Apart from disappointing results in particular, a comment system could be useful for other sorts of follow-up results which are not substantial enough for a new paper. I've noticed that I often learn new information after I publish a paper which is of potential interest to readers of my paper, but not substantial enough to be published in a new paper.

To prevent spam, commenting could be restricted to ISCA members. Membership could be checked with the same password system used to restrict access to the archive. Even better, a unique user name and password could be computer-generated and emailed to each ISCA member. This would allow the commenter's name to be automatically associated with their comments, preventing any abuse of anonymity.

I'm very curious what people think of these ideas.

Regards,
David