Chow #256 – Leveraging WebAssembly

The Challenge:

Your team is excited about the potential benefits of WebAssembly. Your team wants to try it in their application. How, as a team member, do you approach this?

Suggested Answer:

Prioritized Goal:

We need to identify where we should improve rather than doing a random walk. Hence as a team, we start our journey of identifying the wildly important Goal from a list of Goals. A SMART Goal with a tricky “m”.

We work towards collecting doable Usecases, by creating a relative score matrix with the following tentative vectors:
a) customer delight thanks to quicker response – Positive
b) reduced load to servers – Positive
c) reduced defects due to improved testability – Positive
d) complexity in modernizing within a reasonable time period – Negative

This should help us to reason and prioritize the Usecase that has the potential to offer us the right benefit within a time. Of course, We baseline the “value of the current situation” for future comparison which is use-case specific.

Leverage every idea:

On a number of occasions, we need to apply more than one strategy to reach our Goal.

If I were driving the developer of selected use-case, then

1. I will not hesitate to assess if my design is right. Design and algorithm make the cornerstones of an elegant and efficient solution. If I have the right design and algorithm in place, it is adequate for me to squeeze the best possible performance.
2. If the problem is with DOM Manipulation and re-painting the canvas, there are lesser chances for WebAssembly to help me here.
3. I will look for areas that involve heavy I/O operations. I will check if I can decouple that aspect to a combination of WebWorkers and WebAssembly. This may make the main thread available to converse with the User which may improve the customer delight eventually.
4. I will apply all the possible ideas like JavaScript, WebWorker, and WebAssembly judiciously to improve the situation.
5. I will listen to my team for out-of-the-box ideas and steal them shamelessly.
6. I will look if the bottleneck is shifted to a different place because of the small improvements, to avoid solving a problem that no longer exists (“phantom problem”)