I agree with your general perspective. (I have publicly vowed to never act as a reviewer for Elsevier, for example.) However, ignoring the traditional role of

I agree with your general perspective. (I have publicly vowed to never act as a reviewer for Elsevier, for example.) However, ignoring the traditional role of publishers does not help our case. Regarding your individual points: - Peer review for CS conferences is usually organized by volunteers, yes. However, this is not true for other, journal-focused disciplines, where there are paid assistants - One might argue that spelling/layout checks are not really necessary. I personally appreciate them. In any case, these _are_ services that publishers usually offer. - I was not trying to argue that the typical publisher does a good or cheap job. What I wanted to say is that publishers can and do add value. - Universities usually try to handle scientific fraud discretely (for PR, HR, and legal reasons). They are usually not interested in disseminating bad news widely. - Sure, if we could retroactively bring all previously published works into the public domain, things would be much easier. However, in the copyright system we live in right now, publishers provide a useful service by taking care of author's rights and facilitating reuse.