Hi Paul,
I'm a researcher who uses arxiv daily and I love it. I love it to the point that I sometimes wonder if it could become more than it is. Did anyone involved with arxiv ever toy with the idea of attaching message boards to papers in which the papers could be discussed? i.e. One of the links from the abstract page could be to a message board specifically for discussing that paper.
One of the worst things about the peer review process at journals is the lag and one-way nature of communication between authors and referees. It is tremendously frustrating to have one referee torpedo the whole process because he/she has interpreted the paper incorrectly. I'd love to see somebody try to replace the traditional referee process with something like an invitation-only forum so that authors and referees can interact in a timely basis with anonymity preserved. Arxiv could do a variation of this by creating a forum open to users who have published papers on the arxiv (just to weed out spammers). This could provide rapid feedback to strengthen papers as well as a place to collect answers to questions from readers.
(1) arXiv.org is amazingly successfully because people use it and love it. However, the low funding of arXiv.org also means that arXiv.org has missed many opportunities. Had we gotten 4x the funding we got, and spent it well, we might have come to dominate several more academic fields and we could have launched some awesome features.
(2) I think we could have solved the problem of "non-scientists" commenting, but there's also the problem that the physics community is very snarky. What if, say, a grad student writing his first paper gets a really nasty comment by one of the luminaries in the field? We didn't want to get involved in all the suffering that would have entailed.
You're totally right about the ineffectiveness of peer review and that was a subject that we talked about a lot. Some service for giving feedback to papers and evaluating their significance could have been a great help. Yet, a serious investigation of this could reveal that many of the assumptions about how science works may be wrong.
For instance, we found that by far, the papers on arXiv.org that are of most enduring interest are review papers. Like the U.S. patent system, scientific publishing fetishizes being the first to do something, not the first to do it right. Similarly, the idea of "reproduction" is core to what people say the scientific method is, but almost no scientific results are reproduced and many or most would not be reproducable if somebody tried.
Remember, despite peer review, the median scientific paper is wrong.
I totally agree with you and congratulations on creating such a widely used and transformative product, I hope that in time biology will also adopt the arXiv model.
One of the worst things about the peer review process at journals is the lag and one-way nature of communication between authors and referees. It is tremendously frustrating to have one referee torpedo the whole process because he/she has interpreted the paper incorrectly. I'd love to see somebody try to replace the traditional referee process with something like an invitation-only forum so that authors and referees can interact in a timely basis with anonymity preserved. Arxiv could do a variation of this by creating a forum open to users who have published papers on the arxiv (just to weed out spammers). This could provide rapid feedback to strengthen papers as well as a place to collect answers to questions from readers.
Cheers!