UPDATE: Please also see the revision to this post that harmonizes the terms “personal data store” and “personal data service”.
I’ve been blogging about “personal data stores” for two years now, but as of last Thursday I’m done with it. The term, that is. Effective immediately I’m moving to “personal data service” and it’s companion term “personal data server” and not looking back.
Here’s why. Last Thursday was the second World Economic Forum workshop on “Rethinking Personal Information” in NYC. The first one was in June, and there the concept of a “personal data store” was relatively new to some of the attendees (though very well received). I’m happy to say that by last week’s meeting, the concept was no longer new to the majority of attendees. Rather discussion was much more focused on how these “control panels for personal data sharing” would actually work in practice — both technically, socially, and legally.
But what I saw happen at least a half-dozen times during the workshop was attendees tripping up over the word “store” because it clearly implies that a “personal data store” actually stores all your personal data. Each time those of us working in the space would have to explain, “No no — it’s actually a ‘virtual’ store — it doesn’t need to centrally store all your data at all, just provide you with a locus of control — a dashboard — for managing it”.
As soon as that misconception was cleared up, discussion and analysis of the possibilities for this new personal data ecosystem moved swiftly forward — and in fact at breakneck speed because this audience was so well prepared to understand how significant a sea change the PDS represents.
Afterwards I was talking with Paul Trevithick of the Higgins Project — which is now focused on delivering an open source personal data server — and he agreed that the PDS community is just shooting ourselves in the foot to keep using so misleading a term. So I’m moving immediately to “personal data service” and “personal data server” — the latter following the lead of Joe Johnston at SocialNori, the new open source project building social applications on top of the Project Nori personal data server. (Note: a key kernel of the Project Nori code is Markus Sabadello’s XDI4J (XDI for Java) libraries. If you haven’t seen Markus’ Three Visions video yet, don’t miss it — he draws a wonderful picture of how personal data servers, the federated social web, and personal apps are coming together).
This shift in terminology is particularly ironic given that Mydex has just published a seminal paper on the entire topic: The Case for Personal Empowerment: The Rise of the Personal Data Store. The lead author, Alan Mitchell, is one of the most articulate people on the planet about the potential for PDS. He was ably assisted by the rest of the Mydex team, including Iain Henderson, William Heath, and David Alexander (who have more mountain-moving news about Mydex and PDS coming later this month).
However the terminology shift doesn’t take one iota away from the paper, which explains (among many other things) why a PDS does not actually have to store the data. Also, just to be clear, a PDS can in fact store the data when a user needs it to, the same way web servers can and do store some of the content they serve. But nowdays that’s the exception, not the rule. Most web servers access the content stored in a database, whether local or remote, and often assemble content from many different databases.
A personal data server is no different: it will dynamically assemble (“pull” to use David Siegel’s term) your personal data as needed, from where ever it is best stored, and no matter whether it is operated directly by you or by an independent service provider. Either way it let you control and share your personal data where ever it may exist locally or on the Web.
Thankfully, ALL of these terms still fit nicely under the acronym “PDS”, so that’s the constant I’ll keep using.