actually thats not a good idea. A paper may already take 1+ year to go from experiment to publication, it should be open for everyone to read immediately to be most useful - science moves fast.