Automated Transformation of Delphi Applications to .NET Architectures: A Framework-Based Approach

The rapid evolution of modern software development frameworks has rendered many legacy systems increasingly difficult to maintain and integrate within contemporary IT ecosystems. This thesis presents a framework for the automated transformation of Delphi applications into modern .NET architectures. Motivated by the specific needs of Witteveen+Bos, the project addresses the challenges associated with migrating internally developed Delphi software, often insufficiently documented and increasingly unmaintainable due to developer turnover, to the widely supported C# language and .NET environment. The research introduces and systematically compares three distinct migration methodologies: a fully local Large Language Model approach, an internally hosted LLM API strategy, and a hybrid solution combining GUI migration through rule-based parsing (ANTLR) with LLM assisted logic conversion. Each methodology is evaluated across multiple dimensions, including migration efficiency, critical code handling, applicability across projects, tool reliability, output compliance with .NET standards, data confidentiality, and migration security. To ensure robust evaluation, the project employed real-world test cases from Witteveen+Bos's internal software portfolio and conducted testing to benchmark each method. Results demonstrate that the LLM APIs based approach is the preferred one in most cases. The hybrid ANTLR + LLM method also shows promise, and the Local LLM has some performance issues due to the lack of hardware resources. This thesis also emphasizes the effect of prompt engineering on the quality of the result, detailing iterative strategies for refining LLM inputs to enforce output accuracy and consistency. Additionally, the work presents practical solutions for handling legacy libraries, ensuring data confidentiality, and maintaining security throughout the migration process.

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2025-07-18
DOI
https://doi.org/10.5281/zenodo.16096683
Primary Topic
Software Engineering Techniques and Practices
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Automated Transformation of Delphi Applications to .NET Architectures: A Framework-Based Approach

Zenodo (CERN European Organization for Nuclear Research)
Software Engineering Techniques and Practices
article

Automated Transformation of Delphi Applications to .NET Architectures: A Framework-Based Approach

article en

Abstract

The rapid evolution of modern software development frameworks has rendered many legacy systems increasingly difficult to maintain and integrate within contemporary IT ecosystems. This thesis presents a framework for the automated transformation of Delphi applications into modern .NET architectures. Motivated by the specific needs of Witteveen+Bos, the project addresses the challenges associated with migrating internally developed Delphi software, often insufficiently documented and increasingly unmaintainable due to developer turnover, to the widely supported C# language and .NET environment. The research introduces and systematically compares three distinct migration methodologies: a fully local Large Language Model approach, an internally hosted LLM API strategy, and a hybrid solution combining GUI migration through rule-based parsing (ANTLR) with LLM assisted logic conversion. Each methodology is evaluated across multiple dimensions, including migration efficiency, critical code handling, applicability across projects, tool reliability, output compliance with .NET standards, data confidentiality, and migration security. To ensure robust evaluation, the project employed real-world test cases from Witteveen+Bos's internal software portfolio and conducted testing to benchmark each method. Results demonstrate that the LLM APIs based approach is the preferred one in most cases. The hybrid ANTLR + LLM method also shows promise, and the Local LLM has some performance issues due to the lack of hardware resources. This thesis also emphasizes the effect of prompt engineering on the quality of the result, detailing iterative strategies for refining LLM inputs to enforce output accuracy and consistency. Additionally, the work presents practical solutions for handling legacy libraries, ensuring data confidentiality, and maintaining security throughout the migration process.

Zenodo (CERN European Organization for Nuclear Research)
Reduced inequalities
Openalex Percentile: Top 45%
Software Engineering Techniques and Practices
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Automated Transformation of Delphi Applications to .NET Architectures: A Framework-Based Approach · Zenodo (CERN European Organization for Nuclear Research) (2025) | TGRS Research Map | TGRS