Research On Maintenance Mechanism Of Mobile Apps From The Software Ecosystem Perspective

With the rise of the mobile Internet,mobile applications,users,developers,and mobile software public platforms constitute a social-technical complex software ecosystem with continuous feedback and interaction.How to fully mine and utilize the data generated by entities such as users,mobile apps,developers,or dynamic interactions between them in the ecosystem to support the continuous and efficient maintenance of mobile apps has received significant attention from academia and industry,and urgently needs further research.In this article,from the perspective of the mobile software ecosystem,we continue to mining app’s multi-dimensional explicit or implicit requirements by data mining and natural language processing technologies from the internal data of the app,continuous crowd feedback data,and the interactive data of the developer in the open community,respectively.Meanwhile,the internal dynamic interaction mechanism of the software ecosystem is constructed and analyzed from the macro perspective of the entire software ecosystem.So as to realize the constant dynamic and efficient maintenance of the app.The main contributions of this article are as follows:1.App security maintenance based on minimum permission set identification.From the perspective of the internal permission of the mobile app in the software ecosystem,we realize the app minimum permission set identification from the permission data,the description information,and the API usage information of the app by using static analysis and collaborative filtering technologies.Then,we detect the app’s risk permission based on the minimum permission set of the app and recommend the reasonable permissions set for the developers,thus to achieving continuous security and reliability maintenance of the app.2.Continuous app maintenance driven by crowd review feedback.From the perspective of crowd feedback in the software ecosystem,we employ sentiment analysis,machine learning,and natural language processing technologies to mining the users’ explicit or implicit intentions based on the users’ continuous feedback data,such as ratings and reviews.Meanwhile,the multi-dimensional information such as user intentions,user sentiment,etc.,are comprehensively considered to calculate user review scores.Next,the review priority is divided and recommended to developers to support subsequent app maintenance and evolution.3.App maintenance and iteration that integrates user reviews and the wisdom of the developers.From the perspective of potential collaboration between users and developers in the software ecosystem,the source code files that potentially need to be modified are located through classification,clustering and linking of users review to establish the semantic links between users feedback and source code components.Thus the developers can complete the continuous and efficient maintenance of the app.4.The internal operating mechanism and health analysis of the mobile app ecosystem.From the macro perspective of the software ecosystem,we have built a complex dynamic interaction mechanism for entities such as users,apps,developers,etc.,and designed indicators to measure the health of the software ecosystem.Then,we simulate the dynamic interactions of various entities based on computational simulation experiments and analyzed how these entities affect the software ecosystem’s health.This not only helps mobile users better understand the software ecosystem’s operating status,but also provides a reference for developers and mobile platform managers to make app development,maintenance,and platform operation decisions,respectively.In summary,focused on the maintenance mechanism of the mobile app,this article has make use of the internal permission data of the app,the continuous user review feedback information,and the constant interactive data of the developer community to realizes data-driven app maintenance by data mining technology and collective intelligence.Further,the complex dynamic interaction behaviors in the software ecosystem are constructed which is used to analyze the health of the ecosystem by computational experiment simulation technology.This can provide a solid theoretical support for the continuous rapid and healthy iteration of the app in future.