Evaluating Scientific Work by Means of Diffusion

There are two approaches for evaluating scientific papers. The classic way is to choose well established representatives of the specific scientific community and have them evaluate their colleague’s work. The other method of evaluation, the so called peer-evaluation method, is where peers (famous or otherwise) of the author evaluate the paper. Peer-evaluation resembles the diffusion process in which a new substance spreads out to the whole solution. Similarly the new author and article are diffused among the scientific community, smoothing the level for accepting scientific papers. Using the classic-evaluation system of accepting new papers, the average starting scientists writes their first number of articles as collaborators with a renowned scientist, thus gradually building up their image. Only afterwards do these authors dare to independently publish. What are the pros and cons of both these types of scientific article evaluations?