As I always say: If you have one data point you publish in Nature, two data points Phys. Rev. Letter, anymore than that publish in Rev. of Sci. Inst. :-) I agree with Schekman as far the field of physics goes. How many Nobel prize works are published in Science or Nature? Not too many it turns out. The good work is published in so-called archival journals. Peter Higgs apparently agrees. http://www.independent.co.uk/news/science/new-nobel-laureate-randy-schekman-hits-out-at-academic-publishers-over-their-publication-of-only-the-flashiest-research-8996376.html
Add to this that most "Top Level Science" in medicine can't be reproduced, a main criteria to be called science...http://online.wsj.com/news/articles/SB10001424052970203764804577059841672541590
With these two things (unreproducibility and published wrong results) one of the most basic tenets of science, that it is "self-correcting" is absolutely not true, as wrong papers are rarely retracted and unreproduced or null results are rarely reported (as they are not flashy science). We need a Journal of Null Results, or a Journal of Don't Try this Path! :-) But who would read it?
I agree with you, you have summarized the story in a cool way:
"If you have one data point you publish in Nature, two data points Phys. Rev. Letter, anymore than that publish in Rev. of Sci. Inst. ",which I believe is applies to other fields as well.
I like the term 'luxury journals', because those journals cost a lot money for all of us and in the end tend to limit our funding opportunities. After all, our institutions have to pay their exorbitant subscription fees (or individual article costs). What's more, I like Open Access. It's stabilizing the budgets of many countrywide organizations engaging in science (pay once, use often).
I wish we all the fields have a journal for negative results. Thanks Islam for pointing this out.
Michael I totally agree with you OA publishing is merging fast and I think it will play a role in changing the publishing market (if the idea is adopted carefully).
Don't forget the tens of thousands of any currency they charge Libraries for subscriptions, and force Libraries on limited budgets to cut subscriptions "lesser" journals. Three cheers I say ...