The death of applications?

Written By:
Published:
Content Copyright © 2011 Bloor. All Rights Reserved.
Also posted on: The Norfolk Punt

For much of my working life, I’ve been involved with “applications development”. But this seems, to me, to exemplify the essential, dysfunctional, siloisation of IT in practice.

Business users don’t want “applications” they want “business outcomes”—at least, they do until IT has “trained” them. They don’t, given the option, want to click on an application and decide what to throw at it, they want to click on something useful, like a machine or a store of data or a piece of information and initiate its behaviour or change its properties—to prod it and start it doing its thing—so as to produce a business outcome.

And yet IT remains application-focused, even though experienced users, even in IT, are more likely to click on a business object (such as an email or piece of data) to do something than to click the “start” button and look for an application.

Even in the brave new world of the smartphone there is a focus on “apps”—not, by the way, a new term but a term I remember using back in the 1970s for the things that my bank used on the mainframe to automate banking (perhaps this is significant!).

Now, Intel is suggesting (see Intel’s SoftTalk blog here) that it’s about time we put the users, not the applications, first. According to Doug Fisher (quoted in the blog) talking about tablets running Tablet User Experience for MeeGo: “There is a focus on the user, [The tablet] adapts and learns what users do. If you don’t read emails from certain people or follow their tweets, it learns. It puts the ones that are most important to you at the top. You can see how the user concept could be instrumented across all segments for consistent interaction across those user needs.”

This is not just a tablet thing, of course, and Google’s “priority inbox” on Gmail seems to be built around an aspect of the same philosophy (to say nothing of the Mac interface, partly because I’m not very experienced with it), but it is nonetheless welcome.

I see it as just another step towards IT generally worrying about “business outcomes” rather than just about cool, effective, efficient IT. Presenting a Music Panel which does play the music you want regardless of what application is needed to make it work, or a Friends Panel which communicates with friends whatever email or social networking application they use isn’t that much different to choosing an Acquire Customer panel to collect customer details and to fire up CRM and perhaps some real-time analytics to suggest a sales strategy.

Photo of Uli Darmschat © David NorfolkShould chatting to photographers on Flickr really be that different from talking to business partners on LinkedIn—and should both channels sometimes be open at the same time anyway? As Ulrich Dumschat of Intel puts it, “personalization of communication, doing business, social networking etc. is definitely the way to go” (Uli Dumschat is both a Product Marketing Engineer for Intel Software Development and an excellent guitarist). Should the interface to business IT be the Enterprise Architecture model or a subset of it, an interface that remembers the people and processes you mostly use. And should Social Networking be part of this?

There seems to be a certain synergy between all this and the “execution” (in a fairly wide sense) of a Business Process Model using something like Mendix.

But it also occurs to me that tools like uniface aren’t, or at least needn’t be, a million miles away from this idea either. if there is a difference it is probably around the idea of real or perceived lock-in, which i’ve looked at in my last blog.

For now, I’m just thinking about the importance of the “Application”. Some years ago, Martin Banks and I were part-owners of something called Application Development Advisor and it was, by most accounts, well received. However, perhaps we should have called it “Business Outcomes Advisor”, because assuring business outcomes from partly automated process is what modern IT should be enabling.