Is this enough or do we need to go even further?

in my oppinion we need to take much more steps. Journals and publications where good institutions in times 50 years ago. Where it took mankind much efford to spread valuable information. But heaving the web and network technologies a scientific journal seems to be pointless in a certain way. The entire reviewing process of course gives some trend of what papers have good quality and what papers don’t. Afterall science is envolving at a faster and faster speed after all. 
But as we can see from open projects like wikipedia, linux, wordpress and many more those procedures also yield amazingly good results. They take less effort and are much faster in their deciession process. I have observed that open repositories like http://arxiv.org/ helped mathematicians a lot. But as I know from my good friend Dr. Heinrich Hartmann who is a postdoc at Oxford is that many advanced scientific research discussions among the younger generation already proceed on http://mathoverflow.net/ an “unscientific”, not citeable, not driving your citation count webservice. There people can ask questions and post answers. The best thing is they can rate the answers and receive “carma points” for providing good answers and receiving much feedback. 
We see this outside of academia envolving even faster. When I have a question about anything I type it into google search. Chances are high that someone asked the question on yahoo answers or some similar service where best answers are voted and crowed sourced! If the entire society has figured out the strength of this system, why not using it in academia as well?
Obviously these kind of communications and collective intelligence efforts are possible in todays world. So why do we still stick to our old fashoined “good” working methods despite the fact that everyone I talk too is complaining about the reviewing process? Never heard someone complaining about mathoverflows feedback mechanisms!

It is also much cheaper!

It is unbelievable that Elsevier is making a revenue of $3 bn / year. This money comes from our education system! This money could be invested to researches. Maintaining a reliabale citeable academic website compared to mathoverflow would cost us a fraction of this money. Making it possible to afford more researchers actually working on problems! 
But actually the best part is, we don’t even need to maintain such a site. Servieces exist. Maybe not designed specifically for academic research but we could still use them. Why do we need a conference or e journal. Every idea every solution can be made publicly and discussed with a broad audience. My first research paper about graphity is still in the reviewing process where some comitee decides weather the idea and results are good enough to be published. This is happening while the corresponding blogpost already received 1465 views on my blog and about twice as much on dzone and made quite some buzz already setting me in the discussion with some co’workers from linked in, yahoo and microsoft… proving that my research results are actually of interest to people. All this happens with me being a new fish in science having no precompiled trust or authority on any topic at all.
So everyone it is your descission how you act. Of course it is easy and probably efficient as well as time saving to rely on some authority to select high quality information sources for you. And I am not saying that the quality in top conferences and journals is not high. But – besides the money – you might pay a very high price in the sense that these authorities filter a lot of also good information for you. Why not using todays modern technologies and have the crowd decide which resources and ideas are worthwile spreading among people and which ones are not?