I just wanted to make matters more precise. I don't want to beat your comment (I'm an academician too), but let me comment on another example you give:
Open Access is an option for most of the journal I know of, whereby the author pays the publisher up front to have content accessed by anyone free of charge. If an author does not want Open Access, he/she won't have to pay, but the article is no longer free. [Edit: put quotes in the right place]
Even worse, you must sign a copyright form that transfers ownership to the publisher.
True, and that is quite ridiculous, but many publishers are OK with you keeping a copy of the article (e.g. a final draft) on your webpage or on some pre-print or re-print websites.
In the debate over which makes the most impact -- journal vs. conference paper -- most academics will say the same thing: journal, journal, journal.
In Computer Science, this is not true. Publishing on conferences such as STOCS or FOCS is sometimes much more desirable than on any CS journal.
I understand that there is a staff behind the scenes that puts these publications together -- editors, admin. staff, etc. -- and they need to be paid. But some journals have acquired such prestige over the years that they can name their price, and the manuscript submissions will keep coming. So perhaps these prices are not so surprising.
I agree. I dream of a totally open, web-based publishing system where papers are submitted and reviewed by volunteers, and accepted (i.e. put officially on a publication list) only when above a certain threshold. Perhaps the reviewers could also receive comments by both the authors and the public. If both the referees and the authors have a chance of receiving feedback, it could benefit the paper.
Admins and staff would be replaced by the web-based system, while editors (when suggested that the paper has poor wording or language) could be outsourced to the author. [edit: typo]
This is not always the case. In Applied Maths and in many fields of Engineering, there are no publishing fees.