Objectives: We set out to create a Hypertext Rapid Application Framework (HTRAF) that would allow non-programmers to produce data-driven web-based applications in a matter of hours, with little technical skill and at low cost, putting web-tools within reach of most research programs. To meet this ambitious goal, HTRAF had to (1) include a Visual Application Builder (VAB) that non-programmers could use to build and modify applications, (2) be sufficiently flexible to pull data from diverse data sources, and (3) combine the simplicity of meta-data-driven development with the ability to create custom overrides to meet specialty requirements.
Methods: We built HTRAF as a jQuery-based JavaScript framework that can access relational databases and other data sources, pull data into HTML page elements, and allow the data in these different elements to interact. HTSQL is the primary method for retrieving data from relational databases, selected because a single HTSQL URL can retrieve both the data and the meta-data that allows HTRAF to display the data properly. A Visual Application Builder (VAB) allows non-developers to drag and drop data-enabled objects into HTML pages, and to set attributes of each element.
Results: We used HTRAF to rapidly develop the front-end tools for a sophisticated decision support system for funding decisions built on top of a data warehouse. HTRAF delivered better-than-expected results. We were able to build seven prototype tools, with three complete designs for each tool (a total of 21 screens), in under two weeks. We achieved similar speed with applications built on top of a research-data management system. Modifications to the tools following user-testing were incorporated in a matter of hours. Data queries used by the tools could be modified in real-time by updating the HTSQL used to access the data. Non-developers were able to use the VAB to build simple tools. We have not yet attempted to have non-developers build complex tools.
Conclusions: The early results with HTRAF are encouraging and indicate that the technology can be extended to allow typical research lab staff to create and modify powerful, data-driven web applications with minimal web development support. Currently HTRAF still requires someone with web development knowledge to complete the creation of a complex tool, however, non-developers trained to use the VAB and HTSQL can already quickly build simple tools that are immediately useful. HTRAF is open-source, royalty-free, and can deliver value quickly in any organization that wishes to build web applications on top of a relational database. Natural extensions include supporting non-relational data sources. HTRAF is a promising framework that can transform how researchers access and use data in data-intensive fields such as multi-disciplinary autism research.