The final build time metric in the “Discovering The Best Cross-Platform Framework Through Benchmarking” whitepaper measures the total amount of time required for the framework to produce the final build for the benchmark application. The whitepaper evaluates two frameworks supporting multi-platform desktop application development: Delphi and Electron.
This is the fourth in a series of blog posts looking more closely at each of the 26 individual metrics used in the study, and how Delphi and Electron each fared on these metrics. The first can be found here.
Benchmark Category: Developer Productivity
Developer productivity is the measure of effort and code required for developers to complete typical development tasks. Productivity directly impacts product time-to-market and long-term labor costs so tools that increase developer productivity have substantial impacts on business timelines and bottom lines. Productivity can be realized in two distinct ways – reduced coding requirements due to native libraries, and IDE tools like code-completion and visual design.
IDEs with greater library breadth generally result in fewer lines of code per application and produce a clean, lean codebase that minimizes opportunities for bugs or maintenance problems later in the product life cycle.
Benchmark Metric 4/26: Final Build Time
Final Build Time: Total hours required to “speedrun” the application using a known solution. This measures the number of actions and volume of code required to complete the full application by an expert developer with perfect knowledge of a working solution. Productive frameworks reduce development time on repetitive but slightly altered tasks.
Delphi Score: 5 (out of 5)
Electron Score: 3 (out of 5)
Once complete, the Electron application was “speedrun” in half the time of the Delphi application despite requiring almost twice as many lines of developer-typed code. This is largely because Delphi’s IDE provides visual application development [P2] through drag-and-drop components, decreasing the complexity of GUI creation at the cost of increased time configuring components.
Delphi exhibited strength in other areas, however. Its database and network code composed only 46% of developer-typed lines compared to 61% for Electron, a clear indication that Delphi’s FireDAC database library and network tools abstract those operations better than Node.JS, reducing developer effort and opportunities for error.
Overall, similar results in the initial development phase made Delphi and Electron appear equivalent. This conclusion changed after amending the specification to add internal unit tests. Once the contractors sufficiently understood the test requirements, the Delphi application was modified and