The VAR Guy dons a hairshirt for getting five things wrong in 2008 on open source, but other than his optimism on Mandriva, there's little in his predictions to criticize. Indeed, it may simply be too soon to call the game one way or the other on his other predictions related to Novell Suse, Google Android, Ubuntu Server Edition, and Sun/MySQL.
My own biggest failed prediction actually predates 2008, but it took me all of this year to understand just how wrong I had been (or, at least, to admit it). I talked about "burning the boats" back in 2006, by which I meant that the only way forward with open-source models is to cut one's ties to past security (license lock-in, proprietary software, etc.). But I was wrong. Or, at least, I wasn't astute enough to recognize that the lock-in battle would simply shift to hosted applications, data, etc.
Tim O'Reilly got it right. In my quixotic quest to keep everything free, I got it wrong.
Having said that, I do think that lock-in changes in open source. Proprietary software gates use, meaning you can't use the software until you pay, but this is clearly not what any vendor should want. Adoption and ubiquity are the name of the game with Web-enabled businesses. Open source makes both easier.
The "right" model seems to hearken back to an experiment Trolltech made at the end of last millennium. Trolltech used the GPL license for its code, but added a clause requiring commercial users to pay. This, of course, wasn't open source according to the OSI definition, and Trolltech took heat for its position. But eight years later this is essentially the model followed by SugarCRM, Zimbra, MySQL, and others (including Red Hat, if you look closely at its business), though each company uses different means to get to this result.
Whatever the means (proprietary extensions, commercial add-ons that boost production value, etc.), the effect is the same: many users try the software for free but will upgrade to a commercial version when moving to serious production. It's a fair trade-off, and one that I should have acknowledged sooner. Mea culpa.