?
Automatically programming api and driver sсripts method
Traditional object oriented way of programming includes objects and classes, where as result we have library of different modules, and scripts that use that library. Where group of the scripts developed for separate task. And even if that scripts are rather similar by their functionality they are different in part of values, functions that used from library and other features. That involves programming each script separately. Also you cant develop highly complicated script that will have all functionality, because in real life you will have problem with it support and further development, also it usually might be that each script is separate process on server side and it will take too much resource; or some functions can’t be in script when other do, that where programmer must do his architectural choice each time for each task. In this paper I reveal a good approach to solve this task. I split this task to architectural part and programming. Where architectural part solved as straightforward graph, and produces information for second, programming part. Programming part solved by using information from previous part and passes it to neural network, that as output produces functions that must be in scripts, and based on that information script will produce working group of script for particular task. Moreover, I experimentally compare traditional approach with this purposed one and reveal some essential shortcomings of traditional approach.