Low-Code Development platforms (LCDPs) are becoming increasingly popular during the last few decades. They offer increased productivity for developers and lower complexity, making software development accessible for non-professionals. This research outlines a lack of platforms for professional developers, though. Such platforms could take advantage of existing and battle-tested software development technologies, centred on de-facto standard Version Control System (VCS) Git. Nonetheless, a systematic review of the most popular LCDPs (section 2.1) demonstrates that the architecture of a transparent (exposing source models / generated code) LCDP can be improved to offer even better integration with Git.Therefore, the aim of the following research is to develop a technology, which combines textual code with its visual representation in a way to enable editing the same language expressions both textually and graphically, with textual code being the native representation of the source model, adding first-class integration with VCS Git.The research method is the following: • By analysing existing competitors (section 2.2) and choosing core technologies for the solution (3.1) an overall architecture of the solution is developed (section 3.2),• The solution named Langium Model Server (LMS) is implemented using task-list language example (LMS integrated with Diagram Server as a Source Model server, section 3.3),• LMS is then validated with a set of manual tests (Appendix C).Finally, the implemented software is demonstrated (section 4.1), and its features and limitations are discussed (section 4.2). It has been discussed, that:1. LMS represents a core technology for an open-source LCDP, demonstrating the best usability with VCS Git among competitors,2. LMS manifests novel features by combining rich text editing of a domain-specific language (DSL) with rich diagram editing of a visual representation, with textual language serving as a source of truth for LMS models.In the future, LMS can be further enhanced by being extracted from the demo example into a generic library, reusing Langium-based grammar to generate its infrastructure, and integrating with the most recent EMF.cloud Model Hub technology to further expand its capabilities and standardize its API.
| Date of Award | 21 May 2024 |
|---|
| Original language | English |
|---|
| Supervisor | Joan Lu (Main Supervisor) & Vladimir Vishnyakov (Co-Supervisor) |
|---|