June 2008 Archives

Google's presentation at the Enterprise 2.0 show in Boston pushed out more sound bites regarding cloud computing. They made the statement that innovation in enterprise applications during the next 10 years will happen on the Internet (in the cloud).  Certainly, company spokespeople like to put forth the vision at these conferences, but I doubt we will see the obsolescence of Microsoft, Oracle, SAP or other on-premise software during that same 10 year period. The capabilities of the cloud have from most analysts viewpoints, caught up with what you can do on the edge. The question is not about technology however, it's about who carries the asset on their books and how it is managed. There are many companies with rather significant investments in "legacy" systems. What may come to a surprise to this generation is much of that legacy business logic drives corporate differentiation and value creation. It can certainly be ported to new technologies (e.g., SOA hype), but why bother when most of these systems are already depreciated and continue to work? There are still those users who want to "see" the systems running on site, especially if they are communication systems or data centers. It's up to the business leaders to make the case for a change from the status quo. Lasting innovation is driven by market forces, not from the innovation itself.

HTML5 is an optimistic standards effort designed to bring all browsers, markup languages and plug-in APIs under one common industry framework. There has been an accelerated effort in technologies designed to make SaaS more robust by making web applications "behave" more like fat desktop applications. Concerns about connectivity, web response time and user experience have tapered the widespread adoption of these applications in the enterprise. Most success has occurred in the consumer/social web space. One promising development is the proliferation storage APIs. This facility creates a small database that is installed on the end user's client machine to enable access to application features normally operated while connected to the Internet. The one year old Google Gears is making some headway. MySpace has integrated Gears into its messaging application. Yahoo has introduced BrowserPlus in an effort to challenge both Google and Adobe Air. The idea is to have a web application accessible from the user's desktop much like a tray application or a native OS-based executable. One Yahoo demo allows users to edit photos on a desktop with Flickr before uploading to the web thereby increasing the speed and performance of such an operation. Most RIA APIs provide three things: a local database ("SQL Light"), a local object caching mechanism for images or web pages and thread pools to allow asynchronous tasks to occur in the background. These are the integral components of the architecture that enables a rich user experience. Go check out Buzzword.com for an Adobe example of a word processor written entirely in Flex. This brings us back to HTML5. Microsoft, Adobe and others (like CURL) are pushing ahead using some of the storage APIs from HTML5 but leaving other parts of the standard on the shelf. Apple has supported the Webkit open source project with Safari and has re-engineered their own site (removed Adobe Flash & PDFs) by using Ajax instead of proprietary alternatives. It will become increasingly difficult to try to adopt some kind of standard; HTML4 was probably the most successful. Innovation is very impatient with the standards process altogether I'm afraid. Being locked in to a proprietary approach may continue to inhibit the adoption in the enterprise. Most IT shops will choose to utilize a best of breed approach for specific RIA implementations in the short term.

The Facebook Open Platform is finally available a year after they announced it to developers. Back then, it was heralded as "Anti-MySpace" by opening up APIs to third party developers where MySpace had been closed.  Opening up any platform to a multitude of application developers is a two-edged sword. Many users' affection for the novel applications wear off and they begin to tire of the excessive spam associated with their promotion. Facebook says more than 24,000 applications have been built by over 400,000 developers. Google's OpenSocial took a different track by not requiring a special markup language (Facebook uses FBML) and that makes development much easier. Key partners for OpenSocial are AOL, Yahoo, Myspace, LinkedIN, Orkut, Salesforce.com, Plaxo and many others. The New York Times even uses it to allow users the ability to share articles with friends in their social graph.

 As mentioned here before, the key to success with social networking "openness" will be the ability to federate user privacy, profiles, preferences and the whole dimension of data and application portability. A new twist on social networking and a company thinking a little different is SkyDeck. They allow you to turn your phone bill into a map of your social network. The APIs utilize OAuth for secure API authentication and they've come out of the gate with a developer kit even though they are in closed beta. Is everyone racing for the bottom yet?