Recently, after reading Forrester’s latest research on mobile engagement, I recalled a piece of research that Frank Gillett and I wrote back in February 2001 called “Net App Platforms Emerge”. Re-reading this relic made me realize something -- smart developers saw the challenges of mobile coming years before it was on everybody elses radar. The story was simple enough -- client/server developers couldn't transition to a web-centric world, but web developers were expensive and hard to find and retain, and wireless apps were on the horizon but people were anxious about adding wireless to an already complicated transition. Keep in mind that in 2001, HP iPaq and Palm were market leaders, WiFi was in its infancy with the WiFi Alliance less than two years old and the release of the Danger Hiptop, with its totally cool swivel-retracting keyboard, was the buzz at tech conferences. Frank and I decided to focus the rest of the research report on mobile issue. There were some great Forrester reports that helped us out including "Many Devices, One Consumer" (June 2000), "Mastering Mobile Site Design" (November 2000) and "Internet Middleware" (July 1999). If you've been around Forrester as a client or an employee for a long time, these titles should definitely ring a bell.
What amuses me about the Net App Platform report now is some of the crazy lingo that we used to describe a problem that was years away from mattering to anyone. Yet, at the same time, there were some nuggets of truth buried in there. Our analysis went like this: As users adopt a growing array of Internet-capable devices even web-centric apps will come up short so you must make apps available anytime, anywhere, offer business services through machine-to-machine interfaces and weave application components into new services. Alas, today's platforms can't deliver the next generation apps because they lock apps to the PC browser, have no links between content creation apps and traditional development tools and lack critical Internet interfaces such SOAP, UDDI and RosettaNet. "What's the answer?" we heard the world cry -- The Net App Platform. This platform supposedly did three things.
- Separate interaction code from business logic. What the heck is interaction code? It was our term for the code that was written to allow users to follow different paths through the same application depending on their needs and the device capabilities. For example, the user needed access to the same core CRM application functionality but would have different tasks that could be accomplished when they were on a PC with a full browser versus a PDA (no network in 2001, of course) or a mobile phone (WAP, presumably). Different capabilities based on the device is nothing particularly earth shattering today but what Frank and I were trying to describe are two really key elements of mobile engagement -- context-awareness (knowing what capability the user needs at the moment they need to do something) and task-0riented design (the breaking down of a single process scenario into a series of unique and executable tasks).
- Support code and content collaboration. What on earth does that mean? In theory it sounds great. It will get harder to build great apps as the amount of rich content in them grows rapidly, so content creators and developers must collaborate in real time to bring these apps to life. While I'm no expert on content management systems, it does seem that as the years went by we did embrace some of this. For example, I'm writing a blog post in template that a developer has created to make it easy for me publish a post without having to ask a web developer to write a bunch of HTML for it. Do we collaborate? Not really. Do we need to? Probably not unless something goes horribly wrong.
- Embed Internet-native interfaces. When we wrote this report, the idea of cloud-resident, componentized application services was embryonic to say the least. XML was a pretty recent development and industry specific XML agreements like ACORD in the insurance industry we being discussed as ways to someday eliminate the need for EDI. In our report, we saw these interface efforts as a linchpin to being able to assemble services that lived "on the Net". Today, services-oriented architecture dominates enterprise development but, as I vaguely recall, the term did not exist in 2001.
We finished our report with the bold prediction that vendors would build Net App Platforms by 2005. Potential leaders included BEA WebLogic/WebGain, IBM Websphere/VisualAge, Microsoft Windows/Visual Studio, Oracle 9iAS/Internet Developer and Sun iPlanet/Forte. Clearly, our vendor predictions were off the mark. The question I want to ask everyone here is this: Has the basic problem outlined in this dinosaur of a report been solved -- that is, that different users need different capabilities on different devices wherever and whenever they need it? Give examples of applications that you think best illustrate success.