TRANSLATION OF ALGORITHM WITH NATURAL LANGUAGE-LIKE NOTATION TO PROGRAMMING LANGUAGE SOURCE CODE

• Main Author: KELVIN - NIM : 13514100 , ALBERTUS
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/25304
• Institution: Institut Teknologi Bandung

<p align="justify">Implementing algorithm to source code becomes difficult for programmers who have minimum knowledge about programming language syntax. This makes programmers cannot focus on problem solving. A system translating algorithm in natural language to source code could resolve the difficulty. The aims of this study were to find a method to build such translator, make algorithm writing becomes more flexible, and find a method to evaluate system performance. The method has 2 processes, namely translation of variable declaration and algorithm body. The former process aims for extracting variables and data type. The latter process aims for translating every algorithm line to source code. A collection of predefined syntax was applied on variable declaration, programming instructions, array and function call sentences. The syntax covers identifiers synonyms to increase writing flexibility. Functionality evaluation showed that the module for creating variable’s complete form in the first process resulted in more than one complete variables. The test case was combination of array declaration sentences. The module for converting function call to intermediate language in the second process resulted in irrelevant parameters. The test case was arithmetic operation with string concatenation function as the operands. This evaluation also showed that the algorithm writing was flexible. Performance evaluation showed that 10 respondents (100%) who are familiar with C+ + preferred using C++ when the syntax was and wasn’t provided. The top argument was natural language has ambiguity. Meanwhile, 10 respondents (100%) who are unfamiliar with C++ preferred using natural language when the syntax wasn’t provided. The top argument was having no knowledges about C++ syntax. When the syntax was provided, 8 respondents (80%) chose natural language. The top argument was natural language syntax was more flexible and C++ syntax was more complicated. Meanwhile, 2 respondents (20%) chose C++. The top argument was C++ syntax was more succinct.<p align="justify"> <br />