As the number of non-PC devices accessing the internet increases, the ability to use a single-source of information that can be tailored to the device accessing it increases. The presentation outlines how this can be achieved, with particular focus on data gathering.

Firstly, let us assess why we need to gather data from users of a product or service. Feedback is essential if the quality of a product or service is to be assessed. Magazines often include surveys that users can fill in and return to publishers. In return for providing this information, the magazine may offer the chance to win a free subscription or some other enticement.

However, the drawbacks of traditional paper-based data gathering techniques are many and varied, starting with the cost of distributing the survey in the first place. This may include cost of printing the survey, posting it to a representative sample of respondents and including a pre-paid envelope to encourage users to return it.

When the organisation receives the returned surveys, responses are often double keyed into a database to avoid error. There is obviously a cost associated with this, and even then there is no guarantee that the information will be keyed correctly.

Even if the question is entered into the database with 100% accuracy, there is no gaurantee that the respondent has answered in a non-ambiguous way. For example, consider that a survey may ask the user to identify whether they are male or female. Depending on the answer, the survey may ask men to answer questions 2-6 and women to answer questions 7-10. There is nothing to stop a respondent indicating that they are male, and then answering the questions designed specifically for women. This will effectively result in a spoiled survey.

Web-based surveys can address these issues provided the architecture of the system is designed correctly. Firstly, distribution costs do not exist as such because the survey will be placed on a web server. There is no need to single- or double-key any responses as they are stored directly in a database for further analysis. Most importantly, web-based surveys can be designed to support question branching i.e., the next question presented to the respondent is dependent on the answer to the previous question. This removes all ambiguity from the responses given.

So where do wireless devices fit into this scenario? Forrester research predicts that by 2002, 50% of all devices connecting to the internet will be non-PC. This means that in order to attract a representative sample of respondents, anyone wishing to gather data over the web will need to have in place an architecture that can support collection from PCs and non-PCs alike. One of the best established non-PC internet devices is the mobile phone, and already it is possible to purchase a phone that allows you to connect to selected internet services using the Wireless Access Protocol (WAP). Information is presented in WML (Wireless Markup Language), a meta-language that adheres to XML.

The presentation will demonstrate how a single example survey can be used to gather information from a web browser on a PC and also to a mobile telephone.

The main technical emphasis is on WML and building applications which integrate with existing data and applications rather than on WAP protocols.