The Objective Web

Scientific American has collected together a number of printed articles for one of their new online editions (URL below) entitled “The Future of the Web”. While some of the content has dated considerably since its first appearance, Tim Berners-Lee, James Hendler and Ora Lassila’s article on the semantic web still provides much interest and promise. Unfortunately, its promise is largely unrealized due to the almost complete lack of effect that XML – the primary mechanism for producing a semantic web – has had on end users.

The original idea was that the extensible markup language would be used to describe content so that searching could be done much more intelligently. Also, by using ontologies that meaningfully relate XML tags, software agents would be able to draw inferences which would previously only have been possible by knowledgeable users. An example given in the Scientific American article is that a program could readily deduce that a Cornell University address, being in Ithaca, must be in New York State and therefore should be formatted according to U.S. standards.

While this does indeed sound very enticing, the semantic web relies considerably on the kind of cooperation relatively unknown to commercial organizations. XML is being used as a technical solution to simple business-to-business communication problems, but there does not seem to be a sound motivational model for the large-scale use of XML and shared ontologies. Why should struggling .com’s bother? Aside from slightly better hit rates from search engines like Google (that are already doing a pretty good job using other approaches), there is not likely to be much return on investment, certainly in the short term.

The ontological approach also harbors some potential pitfalls for the semantic web. Much in the same way that relational databases force a very specific approach to data modeling, much to the chagrin of user interface designers, the predominant approach taken with ontologies is that of taxonomic hierarchies. That is to say they describe what things are, but not necessarily what they are for. This information of use is implicit in many everyday objects – we need say nothing further about what a potato peeler does, for example. But for many specialist areas, taxonomic hierarchies simply are not enough to help us reach our objectives. As well as a semantic web, we need an objective web – one that helps us to reach our goals, typically in the form of finding solutions to problems. Happily, we do not have to wait for the semantic web to take form to achieve this. All we need to do is to apply some lateral thinking to the way that we use hierarchies in web design. As well as taxonomic hierarchies, we could also provide users with objective hierarchies. So in addition to a product hierarchy organized by some notion of type (e.g. sweet potatoes would be listed under potato or root vegetables) we would also have hierarchies by application and other categories of more direct relevance to users’ poorly-formed goals. Returning to the sweet potato example, we might find them listed not only as vegetables under various methods of preparation, but also under puddings and pies.

This approach can be implemented now by any web site wanting to increase its conversion rate. In fact, Amazon.com already uses a similar approach when it comes to buying gifts. Several objective hierarchies are offered, including “by recipient”, “by event” and “by price”. Other web sites (less successfully) present hierarchies organized by type of user such as home, small business, large business, etc. Where these fail is in choosing discriminators which are not directly meaningful to users. A web site visitor may work for a very large company, but want a solution to a problem in a small branch or home office.

Taxonomies and ontologies are now getting a lot of attention, but we need to try to get these to work for users. When the semantic web does finally catch up, the stage will already be set for intelligent solutions that add real value to our current feeble attempts to mimic printed catalogs online.

Scientific American Special Online Issues: http://www.sciam.com/special/

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 34, July-August 2002} http://doi.acm.org/10.1145/543459.543471