When you agree to publish with Elsevier, you sign a contract where you agree not to publish the final version of the article on-line. Many (most?) Elsevier journals DO permit you to post the so-called "author's manuscript" if you also include the appropriate disclaimer. (Whether this is permitted is also written in the contract you sign.)
In that light, I don't think Elsevier is the worst among bad guys. For example, you're also permitted to upload the unreviewed version to arxiv before submitting it to journal. Some other journals will outright reject the article because of its "prior publication".
For example: http://zvrba.net/writings/fusion.pdf
In that light, I don't think Elsevier is the worst among bad guys. For example, you're also permitted to upload the unreviewed version to arxiv before submitting it to journal. Some other journals will outright reject the article because of its "prior publication".