Two of the most important concepts in today's information technology environment are data warehousing and web-enablement.

But how does one effectively combine the building of large, "scrubbed" and summarised databases (i.e. the data warehouse) with an information distribution mechanism that offers a low-cost, easily maintained interface (that is, the World Wide Web)?

The key to doing so effectively, is the decision support system (DSS). A DSS allows users to access and focus on the information they need most with a minimum of effort. Creating a data warehouse allows the organisation to make the data available. Using the Web to access that data, through a web-enabled DSS, allows the IT organisation to make data available in a simple, easily maintained manner.

The most attractive feature of a so-called thin client interface is that it only requires a less powerful piece of hardware on each of the user's desk configured with a simple Web browser. The servers that these thin clients are connected to will provide the data and processing power.

Updating the data or enhancing the functionality of the DSS is done by making changes to the single, central server machine, not the machines on everyone's desktop. The need for the massive rollouts of a new piece of software, with all of the related support and installation headaches, is eliminated. So is training time, since users only need to know how to use their browser.

However, this concept of the "thin client" can create problems when it is implemented. A truly "thin client" lacks significant processing power and must rely on a server to handle all of the processing. This demands a very large amount of bandwidth.

Insufficient bandwidth reduces the user's ability to do his job effectively. It also places a large burden on the server, which may be accessed by a large number of users at one time. But this does not mean that distributed applications and the thin client have no relevance.

What is necessary, however, is making wise decisions on how the processing is distributed. Rather than a "thin client", it is often better to deploy an "intelligent client".

This is a client with sufficient processing power to enable it to handle some of the user's requests but still relies on the server for the most demanding tasks.

For example, if a user is attempting to subset the data they have been working with, rather than involve the server, an "intelligent client" handles the subsetting itself and only involves the server in the original data request. These intelligent clients, often implemented in a language like Java, make it possible to get the advantages of a distributed DSS without the limitations of a truly thin client.

In addition to providing a mechanism for users to access the data in the data warehouse, the Web can also provide users with information about the data warehouse. A data warehouse offers data from any number of sources providing a single, common definition of terms and concepts. The Web offers a powerful mechanism for making this data (metadata) available to users.

A powerful metadata tool would allow the users to search for all the elements in the data warehouse related to a concept or subject. This would ensure that their analysis is based on the most complete information available.