As I mentioned in my last posting, I spent a day in SF last Monday, participating in a mashup summit organized by Rod Smith and David Boloker from IBM. Many people -- Ross Mayfield, Jeff Nolan, SnapLogic and others -- have written about it, so I will not repeat a blow by blow account of what happened. But to me, two philosophies emerged... One, that believes that all/most action in the mashup space is only in the browser. I am not even sure I follow all the developments in the "browser as the ultimate battle ground for mashup dominance" debate, but it seems like a completely different perspective than where our enterprise customers are likely to end up.
That brings me to the second philosophy. Jeff Nolan articulated some of the points well, but again, let me give you my spin... Enterprise Architectures have evolved with reasonably clear separation of integration layers. Portals for integration "on the glass", ESB's and Process Servers for integration of "processes", and EII for integration of the "information layer." At some level, mashups is nothing but integration done simply. It follows that enterprise mashup architectures will evolve with these separations. And there is another excellent reason why this can be expected to happen. Rarely, in enterprise, the "new" replaces the "old." It adds to the old. So if mashups are this new shiny thing, they will not replace the traditional IT architectures, they will add to it. So EII will be added to, ESB will be added to, portals will be added to. Which means, the same separation will happen in the mashup layer.
And thus, logically, we are building our Info 2.0 layer -- to not replace, but to supplant, and to build upon, all of the enteprise's investment in data and information architectures. That is why I found myself reasonably disconnected with "all the action is in the browser" discussions at the summit, and I am sure I was not alone...
Another teaser of our Info 2.0 -- it is being built on the Zend Framework. Why? Because it provides some excellent support for XML (including the simple XML Content Store built by IBM to enhance deployment of XML persistence mechanisms such as DB2's pureXML). We will also port it to WebSphere soon.