Programming with ASLT and Metainformation

Abstract

The following paper is a short explanation and overview of the work of the team together with Mr. Wolke, Mr. Yermashov and Mr. Rasenack presented at the same conference. We do not prefer source code as the basic definition of an application. Instead we use the ASLT as the basic view. For a first understanding, this is source code parsed into a tree form and stored permanent. Changes on the application definition during development normally take place in the ASLT. Two converters, java2aslt and aslt2java produce the ASLT or the source code. A tree is a much simpler form to do automatic code insertion into an existing application. There are further views into the application. They can be used as input. They are synchronized with the ASLT basic view. The tree form can support a special form of “comments” called metainformation. This metainformation can be attached to each node forming a node of its own. So metainformation can be attached to a metainformation node producing nested metainformation. The concept is supported by tools to access the ASLT, for example to insert a node, to remove it, to change its contents, to validate the type or to evaluate the metainformation and process it, maybe during design time, may be during runtime. These tools are called metainformation processing tools (MIPTs).