The concept of user interfaces with evolving levels of complexity has come up in several recent conversations I’ve had with companies involved in this space. The basic premise is this: a user’s needs change constantly throughout the ownership lifetime of their mobile device, therefore the interface should be capable of evolving to match those needs. The interface platform needs to teach itself to respond better to the needs of the user.

For example, a messaging application might start by guiding a user through the process of creating a video mail, but within a few weeks, it would start to remove elements of the ‘wizard’ interface as it recognised the user was becoming quicker and more precise in their actions.

This capability is already available to a certain extent in the desktop world. Windows XP initially hides system folders from the user and allows customers to choose between a ‘wizard’ inrerface or the traditional ‘power user’ control panel. However, it is a much greater challenge to effectively provide two or more tiers of complexity within the limited space of a mobile interface.

I am not fond of the industry’s penchant for using obscure terms to describe relatively simple concepts, but the best way I can think of to describe this idea is: intra-segment evolution.

The industry’s current focus on better understanding customer requirements and the proliferation of segments is absolutely essential and should continue. However, the idea of intra-segment evolution is to recognise that a user’s membership of a particular segment is by no means permenant.

It may change several times a day – e.g. from family man to busy executive to amateur sportsman – or it may evolve more slowly over the course of several months. For instance, a user may change jobs or relocate to a new country, thereby creating a new set of requirements. Or it may be something as a simple as removing ‘nannying’ self help features to allow technically proficient users to access applications more quickly as they become more familiar with a product.

It is a simple concept and, from a user’s perspective, it should be an almost invisible process. The difficulty, however, is that this kind of technology is actually very complex to implement. This is a real challenge not just for platform developers and UI designers but everyone in the value chain – as a fundamental issue of user experience, it touches everyone from chipset providers to brand consultants.

Self-learning and artificial intelligence are two extremely difficult concepts to deliver with even the most powerful silicon, let alone within the limited processing and memory footprints of mobile devices. There is also a very real danger that introducing this kind of functionality actually causes more frustration for the end user than ever before. I call it the ‘Paperclip’ syndrome: the desire among many Microsoft Office users to destroy the talking paper clip which is supposed to guide them through that application’s functionality, but in reality just becomes the target for the user’s anger!

There is nothing more frustrating than a system which claims to be trying to help or understand you, yet still fails to allow you to do what you want. The feeling of failure associated with not understanding an application becomes something much more: a sense that you’re being patronised.

Elements of intra-segment evolution already exist. At the most basic level, skinning the interface – possible with many mid-range and high-end devices, delivers a superficial level of customisation. Other solutions, such as Zi Corp’s Qix interface, go somewhat deeper. Qix works like a search engine on your mobile, scanning your contacts, applications and bookmarks so you can access them simply by pressing the first few letters of what you want right from the device’s idle screen. Pressing ‘6-2-7’ (M-A-R) would bring up my contact record (the first few letters of my name) and enable you to send me an email, text or edit my contact details. However, it would also present you with any bookmarks which included these letters – my homepage for instance…

The next stage is recognising which applications and contacts you access most frequently so that they are always presented top of the list.

From a hardware perspective, Wildseed’s SmartSkins offer deep customisation and the ability to evolve a single base handset into several types of device. It is unclear what the future holds for Wildseed after its acquisition by AOL, but I would hope that elements of this personalisation capability will remain.

I expect to see many of these ideas about personalisation and evolving interfaces addressed in the Day Two panel debate at MEX entitled: “Supporting user-driven personalisation.” The session is chaired by Julian Swallow of Mobrio and includes a keynote from Olof Schybergson of Fjord. Other panellists include Mark Stalker of Cognima, Mike Bell of RealEyes3D and Fabio Sergio of 3 Italia.