In calling for release of algorithmic implementation, the host of the star blog Nuit Blanche (dedicated to compressive sensing) has put up a post with the title “Nobody cares about you and your algorithm”. Interesting, this kind of parallels my recent thought on authors’ role in advertising their research papers that potentially make scientific progress, which is actually lesser than release of implementation .
I would often browse through all the titles in recent conferences of interest, like CVPR, ICCV, ECCV, ICML, NIPS, and etc. For CV papers, there’s a nice website cvpapers.com archiving all recent major CV conferences and providing links to the electronic version of the papers and appended materials if available (there are more and more such social-professional websites available, e.g., sciweavers.org that holds papers of a wider range of fields besides CV ). This is a pretty great way that helps save me a lot of time to collect such papers and their access, as otherwise I’d have to go to IEEE Explorer and download the papers one by one, or crap them from the CD proceedings which are normally not at my disposal. Going to individual researchers’ website and checking their updates in publication is only by rare chance.
The titles not linked from archiving sites will most likely be simply ignored by me and never knocked in again. After all, given the large number of research papers published in each conference, there’s little chance I missed a very important paper with this strategy. And assuming rational people as researchers, I just don’t believe anyone who’s very proud of their work would wait and hide their work from publicity. Supporting my sentiment is my observation that mostly the early released papers are truly interesting ones!
Copyright could be another concern by many for personal release. That really depends on the publisher. Some publishers will hold the copyright jointly with the authors, so there’s no problem with authors’ release of the material. For some others like Springer for archiving journals, they have adopted radical publication policy that allows the authors to choose grant open access in which case the copyright is held solely by the authors. For publishers that kind of deprive of the authors copyright, we could vote down them . But even this worst case, I guess the authors should still have the right to release their manuscript, do we?
With rapid growth in the number of accepted papers each conference (and in fact also the number of conferences, and explosion of workshops), we could anticipate the growing probability that our papers being ignored and the diminishing citation for each paper. To our intelligence, we create our papers; to our responsibility, we should better advertise our papers in the Internet age.