It's been argued that the ACM's policies aren't all that bad, since they allow authors to host preprints of articles on their personal websites, or on the archives of their employers.
But CS deserves better than "not all that bad" -- why can't we have a professional society that's actively good?
The ACM constitution lists the unimpeachable purpose of serving "both professional and public interests by fostering the open interchange of information and by promoting the highest professional and ethical standards". It's hard to imagine how the public interest is served by limiting access to research -- research that the public typically funded in the first place. Or how paywalls serve the open interchange of information.
Furthermore, the ACM publishing board is aware of the calls for open access. And yet, while they've written that "the age of open access is upon us", so far they have only considered models that will leave the paywall intact.
So let's push the issue. We should be proud of the ACM, but for that, we need an ACM that we can be proud of.
The "Tear Down This Paywall" website was designed and built primarily by Alex Rudnick, a graduate student in computer science at Indiana University. He works mostly in natural language processing, and will go on at length about how the Association for Computational Linguistics is great and their online archive of proceedings is basically the right thing.