The Research Of Automation Technique Of Software Globalization To Large-Scale Legacy System

The demand for internationalized software products and products localized to a particular language and cultural environment has attracted more and more attention because of business globalization. Historically, software systems have been developed for the English speaking community. The internationalization for these existing systems which will produce new systems cater for other environment, has great commercial value. The internationalization and localization based on re-engineering mainly includes two steps, first, migrating the original ANSI source code to Unicode, and second, to migrate the locale-sensitive modules as local.One of the best methods for the code migration of existing software systems is to port the code to support Unicode standard which is a commonly used coding standard. Until now, there is not any automatic approach which can implement the code conversion efficiently. In this thesis, a technique based on lexical analysis is proposed to migrate the program automatically to Unicode. The suspicious code concept distinguishes between the code that needs analysis by context and the code that can be converted directly in the lexical level. Parsing and pre-processing the suspicious code first ensures the correctness of automatic code conversion by code converter in the lexical level. The main issue is the constructing of the suspicious rule set which should involve all the suspicious code in the system. This technique scans the program code twice: detecting, analyzing and manually converting the suspicious code in the first scanning, and auto-converting the remaining code in the second scanning.This thesis uses the satellite DLL technique to extract and organize the localized resources for the localization, implements the multilingual user interface, and develops a toolkit for the extracting of localized resources such as hard-coding strings. On the other hand, according to the different quality requirement, we divide the documents need to be translated into two groups: the group including documents such as user guideline which has low translation quality requirement, and the group requires to be translated carefully. The first resource group can be translated by applying machine translation which is with high efficiency. And for the second resource group, this thesis proposes a translation process, develops an automatic translation tool to implement the accurate translation of user interfaces, extracted hard-coding strings and so on.SecurityGM is a large financial analysis system which has 10 million lines of C/C++ code. Applied to SecurityGM, these kinds of techniques were proved to be available and highly efficient. We finished the internationalization and localization of SecurityGM successfully, and implemented a Chinese version of this system. These techniques can be applied commonly to C/C++ systems directly.