Why isn’t anyone publishing open-access articles in Elsevier journals?

“A couple of days ago, we noted that PLoS ONE publishes more open-access articles in a month that all of Elsevier’s 2637 journals put together publish in a year.  This time I would like to consider why that is... I can take four guesses... First, there is an ideological reason.  Authors who care about open access probably don’t just want their article to be open; they want it to be in an open journal...  Second, there is a legal reason.  Articles published by PLoS are fully open access, using the Creative Commons Attribution License (CC-BY).  This means that data from PLoS articles can be freely mined, interpreted and republished provided only that the author is acknowledged, which means that PLoS article meet the original definition of the term “open access” as stated by the Budapest Open Access Initiative... it turns out to be pretty much impossible to discover exactly what the terms of Elsevier’s “open access” option are... Third, there are technological reasons to prefer PLoS.  They impose no limits on manuscript length, figure count, image size, etc... Articles are made available not only as PDF and HTML, but also as semantically marked-up XML that is very amenable to automatic processing.  Article-level metrics let you see how you’re doing... And finally, there is an economic argument, too.  If it costs $3000 to publish as “open access” (whatever that may mean) in an Elsevier journal, you could save more than half that money by going to PLoS ONE instead...”