I've learned yesterday at JavaSi conference that there are 9 millions Java developers in the world. Unfortunately for them, they are in majority employed by enterprises where they spend most of their days grinding forms and reports. Since this is not the most thrilling activity a programmer can do, they develop elaborate frameworks aimed at reducing manual effort while in reality just making everybody less productive. All enterprise software is in principle the same. The application collects some input from an user or imports some data from external data source, validates data for obvious errors, and stores the data in the normalized form to a relational database. When the data is required, the pieces of a document are first collected from the database and then presented in a myriad of different reports. Many years ago when I was still developing enterprise software also myself, I was taught to solve my application problems with the database. While this approach has served me and other enterprise developers very well, for too long the only type of database that we have known were the relational databases and we've become to reliant on them when solving our application problems. Therefore I think the solution to the problems of enterprise software development won't come from existing enterprise software vendors and developers, but it will come again from twenty-somethings for whom the M in LAMP increasingly means MongoDB and no longer exclusively MySQL.