On his blog Lawrence Lessig notes that he's commited himself to open access for future works to be published in academic journals.
Quote: "I will not agree to publish in any academic journal that does not permit me the freedoms of at least a Creative Commons Attribution-Noncommercial license."
This is worth an applause. I hope more will follow. On the computer science side we already have a reasonable level of open access, and services like CiteSeer provide an invaluable resource, but it only works when people stand up for their rights and refuse to give everything away to academic journals that far to often try to lock things away.
I don't agree at all that CS has a reasonable level of open access. The main publishers, IEEE and ACM, both wall off their archives. They may be available to you under a site license, but they are not available to me. The same goes for Springer, who publish most of cryptography.
I agree with you about IEEE, ACM etc., and I didn't mean to imply that the academic journals are open, but that we have a reasonable degree of open access because the academic journals are less relevant.
I've been in the field for more than ten years professionally, and I've yet to work with anyone who were a member of the IEEE, the ACM or similar organisations, though many of the people I've worked with have had an avide interest in the academic aspects of computer science.
To me these organisations are increasingly becoming irellevant when it comes to research, since more and more CS research is bypassing the journals and end up getting posted online.
Their position is further eroded because CS is one of the fields where hobbyists are still responsible for a large part of the worthwhile research, though the quality of the resulting papers/articles vary greatly.
That's the things I was referring to with "reasonable access".