Not shockingly, there's quite a bit of confusion in the research enterprise about what exactly "open" means. Open access is bad enough, with its green and its gold and its gratis and its libre and its cha-cha-cha (okay, I made that last one up). Open data is worse, partly because it started happening on a noticeable scale before the Panton Principles could frame it properly.
It's pretty safe to say, though, that the Cacao Genome Project ain't open data. Glen Newton has all the details, but the basic upshot is that to get to this supposedly (and trumpeted-ly) "open" data, one has to register (pseudonyms need not apply) and agree to an extraordinarily restrictive license that precludes data mashups and publications, among other things.
Now, I don't know what happened here. It may not have been the researchers' fault. Maybe somebody's lawyer wasn't clear on the concept—funny how often this seems to happen in an "open" context, as open-source developers in academia and industry will tell you at great length. Maybe somebody's website developer was asleep at the switch. I won't poke fun at the researchers themselves, nor assume malice or cluelessness, until more is known.
Just for a moment, though, let's think about what this false claim tells us about the brand value of "open." With regard to data, particularly genome data, it seems to be higher than I would have guessed at this early date. The CGP didn't just quietly put their dataset out there; they made a big deal of its supposed openness. That's fascinating, and it's hopeful. Science is prestige-mad. If open data is a prestige brand, that's a good thing for those of us who want to see more of it.
Curiously, I can't come up with analogous cases of "fauxpen" in publishing. There are the lovely hybrid publishers who can't tweak their website designs enough to get rid of demands for money on articles whose authors have purchased the open-access option, I suppose. In my head that's not quite the same thing, though; it's not really trying to leverage the "open" brand falsely, more trying to ignore "open" in order to grab at more money. I might suggest that publishers have done enough of a smear job on open-access publishing that the "open access" brand is worth less than "open data." I hope that sad situation can be reversed.
The comments are open for LOLresearchers, LOLpublishers, or any other (PG-13, please) illustration of the post title. If I get some good ones, I'll post them here (with credit, naturally).