This work presents the design, development, and evaluation of the Research Information Management System (RIMS), a web-based platform for managing student research in the Artificial Intelligence and Computational Science program at Chiang Rai Rajabhat University (CRRU), covering bachelor’s to doctoral levels. Previously, research was tracked with spreadsheets, e-mail, and paper-based feedback, which scattered submission versions and gave executives no program-wide view of progress. Key features include role-based access for five user groups, step-wise document submission, two-way advisor feedback, a seven-state research status workflow with full history, publication approval, email notifications, an executive dashboard, and a public research catalogue. A review of five university research-tracking systems and three general-purpose tools was conducted to identify gaps and guide the design. Development followed the ISO/IEC 29110 Basic Profile with three iterations, producing 33 use cases and 318 traceable software requirements. The system passed all 792 test cases and 927 automated tests, served up to 200 simulated concurrent users in a local test environment without failed requests, and achieved an overall user satisfaction of 4.51 out of 5 (SD = 0.68) from 32 evaluators. The results show that the proposed system centralizes research data, keeps every submission version in one place, and gives executives and administrators a transparent, auditable view of research progress.

This study presents the Automated Tax Deadline Notification System via LINE for Accounting Firms Using Event-Driven and Dynamic Rule-Based Architecture, a web-based system that calculates tax filing deadlines from configurable rules and notifies staff and supervisors through LINE when task-state events occur. Exploratory interviews with a supervisor and two accounting staff at an accounting firm in Chiang Mai, Thailand identified problems that recur in every monthly and annual filing cycle: misremembered deadlines for clients with non-standard fiscal years, spreadsheet formula errors and public holidays ignored in manual date calculation, and no central view of filing status. Tax rules are stored as data, so supervisors can add or modify rules without changing source code or redeploying the system. A daily scheduled event triggers rule evaluation and task generation, and an alert is sent only when a task meets a state condition, such as being unfiled seven days or one day before its due date; a task still unfiled on or after its due date is escalated to the supervisor. Development followed ISO/IEC TR 29110-5-1-2:2011 over 24 weeks. All 134 functional and non-functional test cases passed, including tests confirming that edited rules take effect in the next calculation without a restart, and User Acceptance Testing (UAT) with three real users across 21 scenarios met the 100% pass criterion. The supervisor’s time to configure tax rules fell from 300 s to 120 s over three trials. Because firm-specific knowledge is held as data, the design can be reused by other firms and for other deadline-driven compliance tasks.