Ülevaade automaatprogrammeerimise hetkeseisust
Files
Date
2018
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Automaatprogrammeerimine või koodi genereerimine on teatud tüüpi arvutiprogrammide loomisviis, kus kood genereeritakse mõne tööriista abil, mis võimaldab arendajatel koodi kirjutada kõrgemal abstraktsioonitasemel. Selliste programmide rakendamine tarkvaraarenduse protsessis on hea viis programmeerijate produktiivsuse tõstmiseks, võimaldades neil keskenduda pigem käesolevale ülesandele kui implementatsiooni detailidele. Senises teaduskirjanduses on vaadeldud konkreetseid lähenemisi või meetodeid eraldi. Väga vähesed uurimustööd vaatlevad aga kogu valdkonna viimast taset. Käesolevas töös käsitletakse automaatprogrammeerimist olemasoleva kirjanduse süstemaatilise kirjandusülevaate meetodi abil. Töö teeb ülevaate teemaga seonduvatest algoritmidest, probleemidest ning uurmisvaldkonna avatud uurimisküsimustest ning võrdleb valdkonna hetketaset praktika hetketasemega. Vaaldeldud 37 asjakohasest uuringust tegelesid 19 automaatprogrammeerimise üldise määratlemise ja alateemadega. Kolmkümmend uuringut pakkusid välja konkreetse algoritmi või lähenemisviisi. Esitatud tehnikatest rakendati 2 praktikas. Viimasel ajal on automaatprogrammerimise fookus nihkunud programmide sünteesilt induktiivsele programmeerimisele, mille on põhjustanud läbimurded tehisintellekti valdkonnas. Mõistete ja alateemade määratlus on teadlaste vahel ühtne. Õigete spetsifikatsioonide sõnastamine ja piisava teabe andmine automatiseerimiseks on endiselt lahtine uurimisküsimus.
Automatic programming or code generation is a type of computer programming where the code is generated using some tools allowing developers to write code at the higher level of abstraction. Implementing these types of programs into the software development process is a good way to boost programmers’ performance by focusing on the task at hand rather than implementation details. Current literature on the subject reviews single approach or method. Very few of them are reviewing state of the art in general. This paper reviews the state of the art of automatic programming by overviewing the existing literature on the topic using systematic literature review method. The paper overviews approaches and algorithms of the topic, examines issues and open questions in the field and compares the state of the art to the state of the practice. Of 37 relevant studies, 19 addressed general definitions and subtopics of automatic programming. 30 presented specific algorithms or approaches. 2 of proposed techniques were implemented in practice. Currently, the focus of automatic programming shifted from program synthesis to inductive programming, caused by a breakthrough in artificial intelligence. Definition of the term and subtopics is consistent between scholars. However, formulating correct specification and providing sufficient information for automation is still an open research question.
Automatic programming or code generation is a type of computer programming where the code is generated using some tools allowing developers to write code at the higher level of abstraction. Implementing these types of programs into the software development process is a good way to boost programmers’ performance by focusing on the task at hand rather than implementation details. Current literature on the subject reviews single approach or method. Very few of them are reviewing state of the art in general. This paper reviews the state of the art of automatic programming by overviewing the existing literature on the topic using systematic literature review method. The paper overviews approaches and algorithms of the topic, examines issues and open questions in the field and compares the state of the art to the state of the practice. Of 37 relevant studies, 19 addressed general definitions and subtopics of automatic programming. 30 presented specific algorithms or approaches. 2 of proposed techniques were implemented in practice. Currently, the focus of automatic programming shifted from program synthesis to inductive programming, caused by a breakthrough in artificial intelligence. Definition of the term and subtopics is consistent between scholars. However, formulating correct specification and providing sufficient information for automation is still an open research question.