The TypeScript Tax
Eric Elliott
14.8K127

Great article.

I observed the exact opposite from your conclusion however.

Our team was on a large multi-year, multi-million dollar project that started with a core group of around 5 developers and ended with a globally distributed team of over 20 developers. We used pure JavaScript for 6 months while ramping up the team, then added TypeScript in an effort to accelerate delivery. We have very empirical evidence that demonstrates a clear 4x boost in velocity (time from a story being accepted and defect-free code delivered) as well as a significant reduction in defects. We noticed that:

  1. We had fewer defects overall
  2. New team members were able to ramp up faster
  3. We had less collisions with different teams working in parallel

Based on that success, our consulting firm began leveraging TypeScript in other projects and we found similar benefits. Looking across a 5 year span of projects it is clear the TypeScript projects had higher velocity and lower defect rates than the pure JavaScript ones.

We didn’t break down the success into refined categories and try to score them but I suspect the ratings are deceptive and would come out differently based on who is doing the rating. I encourage anyone reading this to consider the evidence but at the end of the day, you won’t really know until you try. I suggest doing a 4 week spike integrating TypeScript to see. TypeScript generates clean, well-formatted JavaScript so at the end of the spike if you don’t see ROI then you can just check-in the JavaScript files and continue on your way. Most companies I’ve worked with have never looked back.

The ROA (return on aggravation) for JavaScript was just too high.