Is all of that necessary? Would a model like the following work:
Have a web site where scientists can submit papers. Other scientists can then review the papers. The reviews are publicly (though optionally anonymously) published along with the papers, and the authors of the paper can submit revised versions of the paper.