Pregunta de entrevista de AgileEngine
The test task asks to write the code finding "similar" HTML elements in given files. There is no definition of "similarity" in task description; they expect "similarity from the human point of view". There was 4 files provided (4 test cases). The task description explicitly says "Target completion time is 2 hours. We would rather see what you were able to do in 2 hours than a full-blown algorithm you’ve spent days implementing" and "remember that working software is the main goal so something simple that works is generally better, than a complex unfinished solution" They also asked to write detailed readme and to provide executable file. It was allowed to use libraries.
Respuesta de la entrevista
Anónimo
I provided a solution covering the given test cases using generic string similarity algorithm (from a library). Based on this solution my application was declined. I got a feedback about my solution. There are few interesting things in this feedback: "Application is not able to parse additional html cases" - how should one know about their "additional" cases?! "Some methods are too long, it would be good to decompose them" -checked the solution again. There is no method longer than a screen. "Everything is in a single file." -how is this compatible with the time restriction and the hint to write "something simple" ? "... expect to see an interface with implementation... Data model doesn't exist." -the solution is a 200 lines long program (including detailed comments for every method) just doing the job. There was no requirement to write a Java lib or to implement the Service provider interface! "It would be good to have tests" -actually JUnit tests are part of the solution