Captain Data Modeler Chronicles: Prologue
By kristenh on Mar 30, 2008
It’s been a while since I’ve written here. The key reason is that I’ve made a transition from Captain Data Modeler to directorship of Sun’s Content Management Engineering department. I suppose the nice message is that a lot of the hard work to balance good data architecture and practical business need is what put me on the radar for promotion. Of course the downside is that now I have to peer longingly over a desk piled high with budgets, vendor contracts, and HR priorities in order to catch a glimpse of the bits and bytes in the distance. I do miss those bits and bytes, and how they would always ground me in the comfort of tangible, creative deliverables. There are days when I’m a bit jealous of my engineering team who gets to dive in and immerse themselves in the bits and bytes every day. Ah well, business needs first.
But on the bright side I get to pull the lens back and take a broader look at how we use data on the Web to put our strategies to work. In the past whirlwind year and a half I’ve overseen data flows from legacy data stores, ERP, isolated data silos and files from all sorts of footlockers and broom closets, and I’ve had to conduct that data into new Web site venues and features, low and high-volume e-commerce, unification of product documentation, community sites like BigAdmin, developer resource sites, and much, much more. The first thing that occurs to me, sitting at this lookout point, is that Sun has so much information that we somehow manange to squeeze outside our firewall through various tiny slits. We’re certainly ahead of the marketplace in opening up data to serve customers and partners, but we can do more, and I’m working to see that we do.
We all know that it’s now a much more collaborative marketplace, thanks to the Web. At Sun our marketplace contains some of the best brains in technology, and if we could open up more information in forms that they could easily digest, the possibilities are endless. The most obvious thing we need to provide is more Web feeds, in Atom and RSS, and it would be nice if we offered more data in JSON form, which is now one of the preferred inputs for mash-ups. In general we’d like to provide more content and data in source data formats such as well-defined XML and JSON. Right now too much of what we provide is in presentation formats such as HTML and PDF. And, in some cases, it is still all rolled up with the business rules that govern its current use.
But to get where you’re going it helps to remember where you’ve come from. I think data architecture for Sun’s Web content, while not perfect, is in pretty good shape to expand its usage as ambitiously as I want. There are some interesting lessons in how we tamed the data to that extent. I gave a presentation at XML 2007 (my co-presenter Uche Ogbuji was not able to make it for unfortunate personal reasons), covering some of the work we’d done to in data architecture, and focusing on some of the lessons learned for managing collections of XML. The presentation was very well received, and that gives me the impression that we’re ahead of the curve in what we’ve accomplished behind the scenes, and that this doesn’t manifest enough in what you see on Sun’s Web sites. My experience at XML 2007 encouraged me to discuss such things more often here, not just some of the neat things we’re doing inside the firewall, but more on how we plan to put to the service of Sun’s customer’s, Sun’s community and partners, and ultimately Sun’s strategy.