Contrast that approach with Mandelay’s roach motel attitude to user data (one-way import of CiteULike and now 2collab). Coupled with their TOS’s express prohibition of tinkering around, I see a troubling lack of concern about future access to their users’ data.

“RT @bengoldacre: Zotero file storage has suddenly secretly appeared, $20 for 1gb, all my PDFs synced across any computer i use, with my refmanager, w00h00″

which sounds like more of the same issue…!

Individuals don’t buy articles, just like they don’t buy reference management software. The bulk of these markets is institutional subscriptions (often together).

Finally, it’s difficult to believe that raising the group cap from 10 to 12 is going to justify paying. Maybe you’ve been misled or you’re leaving something out?

So far, publishers have indeed taken a wait and see attitude, given the lack of lawsuits. But time is not on Mendeley’s side. They’re the ones looking for traction and successful business models in a very crowded market. They’re the ones looking to partner with, and sell to publishers. They’re unlikely to see much success at either as long as they insist on including what you note is not an important part of their functionality. The journals my company runs have direct links to upload paper information to Connotea. Mendeley will never get such placement or links as long as they enable copyright infringement, no matter how limited. Nor are we likely to purchase tracking statistics from them when many analytics packages are readily available from companies that pose no such threat.

I honestly think you’re making a much bigger deal out of this than it is, but since you don’t appear inclined to agree that Mendeley would be a really poor vector for widespread filesharing (by design), I’ll just have to step back and let time make the argument for me.

I wish I knew a way to ease your mind about this, but I think you’ll gradually realize over time Mendeley is doing no harm.

Also, since Mendeley is actively trying to partner with publishers, and as you note above, wants them as customers for its statistical reporting, it would seem to me that the onus is on them as far as proving they’re a reliable group to work with. There are lots and lots of reference managers out on the market, so why encourage one that has such potential to do damage?

Perhaps we could table this discussion until there’s some credible evidence that it can and will be used in an infringing manner by a non-insignificant number of people? Removing a feature because of some future possibility that might not occur seems a little extreme.