Browsing for Consistency

In the early days of personal computing every application had its own idiosyncratic user interface with virtually nothing in common with other applications, often including those from the same manufacturer. Yet most applications supported a very predictable core of common user tasks: loading and saving files, data entry, printing, navigation and exiting the application.

Eventually manufacturers did start to standardize and the popularity of graphical user environments somewhat forced their hands. These days it is difficult to find a desktop application that does not support common user tasks in a consistent and predictable way. While some might argue that this transformation stifled innovation, it did great things for usability. No more wondering what magical set of key strokes are needed to open a file or exit an application. No more wrestling with myriad user interface styles and techniques for even the most basic forms of navigation and data entry. No more spending hours learning how to use one application just to have to discard this hard-won knowledge when moving to the next.

But hang on, even for those of you who didn't experience the joys of 1980's personal computing, doesn't some of this sound just a little familiar? Regular internet users will notice that what we left behind in the desktop world almost ten years ago is the status quo for the web. We do wrestle with myriad user interface styles and techniques. We can spend hours learning how to use some sites with no general applicability of this knowledge to other sites.

The main problem is that the only reliable consistency between web sites is provided by the browsers. Yet, from a usability perspective, browsers have remained virtually unchanged from their inception. NCSA Mosaic Version 1 had back, forward, home and reload functions. (A stop button to abandon slow downloads was added in V2.) Fast forward about eight years and have a look at the navigation tools provided in the latest versions of Internet Explorer and Netscape Navigator: back, forward, stop, reload and home.

Compare this simplistic navigational interface to the exponential growth of complexity in web sites themselves. Many sites are huge by almost any definition of the term, with page counts running to five and six figures. Most sites offer top, side and embedded navigation, not to mention separate search and browse facilities. Yet the only thing that users can look to their browsers for help with is rudimentary linear navigation through pages already visited.

It doesn't have to be like this. After almost a decade of living with the web we know that most sites are hierarchical and split into a handful of major sections. We know that every web site has its own home page that may be preceded by a title page or introduction. We know that most sites have contacts and search pages, that e-commerce sites have checkouts and shopping baskets, that hardware and software manufacturers' sites have download and support pages. But there have been no attempts to provide basic facilities to support these kinds of consistencies from the browsers. It is up to individual developers to recognize (or not) the similarities between their site and others of its type and to build on these if they see fit. To me, this seems like no way to develop a user interface. Browsers themselves could offer a consistent user interface for navigating hierarchical site structure (with assistance from the sites themselves, of course). Browsers could quickly display navigation bars, site maps and "you are here" information after receiving data from a server that would be only a fraction of the size of a typical banner ad.

And why stop there? If we are trying to provide a consistent, usable, window on the web we could stop asking users for the same information in hundreds of different formats and in thousands of different locations. Browsers could maintain profiles that would hold delivery and payment information, traveling preferences, musical tastes, items already purchased and so on that would only be given out with users' explicit permission, rather than being scattered across hundreds of servers waiting to be compromised by the next security lapse.

Some of this has been tried before. Internet Explorer has a "Profile Assistant" that I eagerly briefed with my personal details a few years ago. To my knowledge that information was not requested by a single web site in all of the online transactions I performed (I've since stop bothering to supply it). Microsoft is also trying a new approach in this area. With its Hailstorm technology, it plans to offer centralized services to provide essentially the same kind of personal profile facilities that I am suggesting should be browser based. This may well have some advantages to users, but this is a politically and economically much more complex strategy since it relies on services provided by a company that is being routinely accused of anti-competitive behavior.

With the demonstrably poor usability record of the web, maybe its time to renew our interest in user-centered, rather than technology-centered solutions.

The Author

William Hudson is principal consultant for Syntagm Ltd, based near Oxford in the UK. His experience ranges from firmware to desktop applications, but he started by writing interactive software in the early 1970's. For the past ten years his focus has been user interface design, object-oriented design and HCI.

Other free articles on user-centred design: www.syntagm.co.uk/design/articles.htm

© 2001-2005 ACM. This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version was published in SIGCHI Bulletin, {Volume 33, July-August 2001} http://doi.acm.org/10.1145/967157.967171