Not in astronomy as most (all) papers are submitted to astro-ph and so are freely available.

1. I think that there is questionable value for US astronomers to publish in Nature vs. ApJ, but the same is not true in other countries. In many smaller counties, research funding is directly tied to the “impact factor” of the journals that scientists publish in. ApJ’s impact factor is the highest among astronomy journals, but Nature beats it by a factor of 5 or so, although I think that Science is usually slightly ahead of Nature.

2. Nature’s enforcement of the embargo policy is somewhat flexible depending on how much they want to publish your paper. If they want your paper more than you want to publish in Nature, then they’ll be quite willing to forgive “accidental” violations of their policy.

3. Embargo rules are basically designed to get equal access to the press. If one press outlet gets an early article out on a discovery, then all the other outlets are “scooped”. It becomes old news, and many news outlets don’t do the story. Embargos are quite common with papers that are accompanied by press releases, and they allow the press releases to be distributed to the journalists ahead of time, so they can try to understand the discovery and write there stories ahead of time. These are done on the honor system, but journalists who violate the embargos risk being excluded from future press release distributions.

4. Obviously, this implied threat of denying future early press releases does not apply to arXiv. And since scientists would generally much rather see the paper than the press release, there doesn’t seem to be any rationale to trying to apply this embargo to CV or Bad Astronomy bloggers.

5. Nature and Science generally do press releases every week, and generally insist that any other press release follow their schedule. ApJ allows the authors to decide if they want an embargo.

6. Nature actually has a much more offensive policy than this embargo issue. Nature, unlike Science, insists that the editors and not the authors have the right to decide on the title of a paper. Typically, they don’t tell the authors in advance that they are planning to make a change (although they do send you a boilerplate warning), and the authors see the change with the proofs of the paper. And typically, they will change the title to hype the result in a way that most scientists would be too cautious to do. Since Nature insists on publishing only cutting-edge results, these changes often change the whole meaning of the paper, from the report of a potentially exciting result to a bold claim that later turns out to be wrong. Or more often, they change the title to one that claims more than the paper actually shows.
In contrast, Science allows the authors to control the title of their papers, so I would recommend Science over Nature for future astronomy papers.

Several years ago, I had an extended email correspondence with the maintainers of
ArXiv, trying to get them to introduce a REQUIRED field containing the publication status,
with mandatory updates. (If a paper moves from submitted to accepted, most will probably
update voluntarily, but if a paper is reject, it might linger on in the submitted status,
perhaps indefinitely.) The impression I had was that this wouldn’t be implemented because
it would essentially use external criteria (publication status) to judge the quality of ArXiv
submissions, whereas in some sense ArXiv wanted to become an alternative to the
established system, though of course it accepts papers published elsewhere.

Why do people use ArXiv? Because it is a one-stop shop. Journals nowadays have online
access, and it would be easy for them to make accepted or even submitted papers visible.
Most people would still browse just the daily ArXiv listings. Why? Because that is ALL they
have to browse: it covers all the journals, both popular and obscure, as well as conference
proceedings etc. In other words, it’s more efficient.

This highlights the difference between serial reading and browsing. Internet pioneer,
comedian and troublemaker has some interesting thoughts on this at:

http://ideas.4brad.com/archives/000112.html

and near the end of:

http://www.templetons.com/brad/clarinet-history.html

I’ve taken to reading and commenting on a few blogs recently. However, in almost all
cases the same functionality could be implemented better and more efficiently if the
blog were not a blog but a newsgroup (though to make the most of this, it would have to
be read with a real newsreader and not just accessed by a web browser via an http
interface, e.g. Google Groups).

In other words, the daily ArXiv listings are like a new newsgroup post. Checking up on
all journals and other sources for new content would be like catching up on blogs. Which
is more efficient? ArXiv, like a newsgroup as opposed to a blog, provides a standard
interface to all articles.

This gets back to the reason for journals to go online at all, rather than just relying on the
stuff being posted on ArXiv: vanity. In other words, journals want to remain players and
they do what they can to try to achieve this (embargoes etc) for the same reason that people
have blogs rather than creating an equivalent newsgroup.

Sure, some serial content benefits from links to URLs elsewhere, and some web content can
be read sequentially. But almost always, one wants either web-like content or serial
content. If I book a holiday, the website might have thousands and thousands of pages,
but in booking I navigate through the ones I need. I don’t care about the others, and I visit
the site only when I need to. In contrast, some things are better read sequentially, like
news and most blogs. OK, the blogger himself can present his posts sequentially (though,
unlike with a newsgroup in which a subscriber automatically notices new content, the
reader has to check for new content), but the comments are no longer sequential (especially
when more than one blog is concerned).

In other words, the world would be a better place if there were a uniform interface to
blogs and comments which automatically alerts the reader to new content and doesn’t
bother him with stuff he has already seen or doesn’t want to see. But why re-invent the
wheel? This has been around for years with the NNTP protocol. In other words, why not
have alt.talk.cosmic-variance as a newsgroup? What, apart from lack of vanity (the
equivalent of the hair-metal poser), would be lost? Much would be gained.

If most bloggers continue with the inefficient blog format out of vanity, of course it is too
much to expect journals to behave rationally; they prefer vanity to efficiency as well.

The only context in which an embargo has any meaning is as a contract between two consenting parties. Authors: “I will let you see the paper, if you agree to X, Y, & Z.” Recipient, “I agree to X,Y, & Z – now please let me see it.” Authors: “OK, here it is.” We do this all the time with colleagues, etc.., who might look over something before you submit it. Referees are always under an embargo, given the obvious conditions under which they agree to see a paper. But in all cases, the conditions AND agreement to the conditions are clear before an exchange of the paper itself takes aplce.

One can under no circumstance, however, declare an embargo in the public square and expect it to mean anything or have any power.

I was at a conference in 1996 when a speaker declared an embargo before he spoke to an audience of 300 or so, which included science reporters. The fact that he then immediately launched into his talk without verifying consent of all 300+ present, or even asking those unhappy with this to leave, rendered it instantly and ludicrously irrelevant. As a member of the audience I certainly had no responsibility to keep secret what he openly stated.

With regards to the science press, they are professionals and not under our control or management. They are welcome to get their information where ever they care to, and do what ever they want with it under their own journalistic system of ethics. If they want to read an astro-ph paper and make a front page story out of it, or completely mangle it, that is their call. Again, an embargo here requires two consenting parties who define a verbal contract before exchanging information.