The linked article is directly contradicted by the very same email thread it quotes from. In fact, John Lenton's reply to similar CouchDB bashing in that same thread pretty much says it all:
That's not a fair thing to say. We (the Ubuntu One team, working with the people from CouchIO/CouchBase/MemBase) couldn't get CouchDB to scale the way we wanted it to work; millions of users, many oauth tokens and several databases for each. It's not reasonable to say that it doesn't scale at all, and that's not the case; there are plenty of people using CouchDB to store billions of documents, and we don't want people saying that Couch is "fundamentally flawed", because it isn't; it just wasn't right for the Ubuntu One use-case. Please don't denigrate people and projects unjustifiably; that helps no-one.
That's not a fair thing to say. We (the Ubuntu One team, working with the people from CouchIO/CouchBase/MemBase) couldn't get CouchDB to scale the way we wanted it to work; millions of users, many oauth tokens and several databases for each. It's not reasonable to say that it doesn't scale at all, and that's not the case; there are plenty of people using CouchDB to store billions of documents, and we don't want people saying that Couch is "fundamentally flawed", because it isn't; it just wasn't right for the Ubuntu One use-case. Please don't denigrate people and projects unjustifiably; that helps no-one.