Anyone who has tried to replicate an experiment based on the description published in a paper knows that this can be difficult, frustrating, and often close to impossible. The protocols in the Methods section can be incomplete, even inaccurate, and sometimes lead the hopeful reader down a trail of never-ending references to previous papers, eventually arriving at a protocol only marginally related to what the reader actually set out to find.
One answer to this problem, in a few cases at least, might be a new video journal spearheaded by Moshe Pritsker, a postdoc at Harvard Medical School/Massachusetts General Hospital. The Journal of Visualized Experiments (JoVE) (which has already been mentioned at A Blog Around the Clock and The Frontal Cortex) allows researchers to submit their own detailed video protocols (accompanied by a written description) in hopes of more effectively transmitting scientific knowledge and allowing others to actually be able to replicate their experiments in practice, not just in theory. The site already hosts a few protocols in genetics, animal research, stem cell biology, and cell biology, and it will make its first full edition available at 11:00 pm EST tonight.
Unfortunately, the protocols are not currently peer-reviewed (so I would hesitate to call it a "journal"), although the site claims that this is an eventual goal. On the other hand, the site is built on an open access model, which should help maximize its potential to enhance the sharing of scientific information. Regardless, this will certainly be an interesting site to watch out for, and hopefully we'll see it eventually transition into a full-fledged peer-reviewed journal.
It is a journal. Our Editorial board includes leading scientists from best universities (check Editorial on our website). Videos submitted reviewed editorially with help of experts in specific fields.
Totally open access. Submission is free, and publications are freely available online.
The problem you point out is what allows so many scientists to get away with fraud.
The myth that this fraud is always discovered and that science is self correcting is long overdue for exposure.