Old versus New Applications

Those of us who had some computer science classes typically treated the client (UI) as a minor part of an application. Most of the complexity and functionality existed in the server application and the database. The database was usually a relational database using SQL. Now, most applications have complex client code running in browsers. The server application may be divided up into several separate programs. The database stores JSON data.

Building an application from small programs. often written by others, means that one programmer can provide complex functionality. The use of JSON greatly simplifies the exchange of data between the client/browser, the applications, and the database. The price is that programmers are now system administrators. Much of effort is spent on configuring components written by others. JSON data structures, as well as XML data structures, tend to be poorly designed and are extremely difficult to process. Exchanging small amounts of data in JSON is easy, but QA and reporting suffers.

One good aspect of the new architecture, is that deployment environments such as Docker and Kubernetes now simplify installation. This is a major improvement and means that it is practical to install application environments rapidly.

In general, simple applications can be written quickly, but custom programming is difficult. The new tools make it easy to shuffle data between components, but complex processing is slow and difficult. Also, the new environment transforms programmers into system administrators. Much of the time goes to setting up virtual machines (VM's) and debugging complex components written by others.