Dynamic Web Automation
Web automation is the process of making websites do useful things, usually involved with the collection, dissemination, and analysis of data. For example, a shopping cart page displays information requested by the customer and gathers a customer order.
L D Ness Inc. leverages exciting technologies, including AJAX (Asynchronous Javascript and XML) to provide a much smoother user interface while dramatically reducing server and bandwith loading.
Most web pages require a user to complete a form and press submit before seeing additional information. For example, a web page that provided data on each town and city in the US would have the user select a state and hit submit (or implicitly submit the form) before redrawing the entire form. Both the web server and the user experience pain during this operation, as significant redundant data is retransmitted and redisplayed.
Some web pages try to get around this problem by pre-loading the web page with all possible information. When the user selects a higher-level item, such as a state, such a page would use Javascript to display the desired data. This approach becomes unworkable when there is a large quantity of underlying data, as the initial page takes longer and longer to load.
We take a completely different approach to this problem. We have the web page interact with the server while the user is working with the page, requesting each piece of data only if and when it is required. In the city/state example, if the user chose New Jersey, then the server would provide a list of cities in New Jersey only, and the list would update on the web page. (The page would not be redrawn) If he or she requested a different state, then the server would provide that state. If the user then requested New Jersey again, the web page would not bother the server again, since it would already know the New Jersey cities.
We can use this approach to give rapid access to a vast store of data. We have developed a tree-view system based on this technology that facilitates data mining . Any database relationships can be represented in the tree. (We actually use a table to represent its structure) In a well organized tree, users can quickly drill down through many levels of data while incurring a minimum of server overhead or web traffic.