Fullstack Web Assembly

The other day I was thinking about the Shape Detection API and the possibilities that it presents for web developers, and I was also thinking about what the progressive story is for web developers in the sense that this API is still in development, not fully landed in Chrome and for a long time is likely not to be supported natively on the web-at-large.

A traditional understanding of Progressive Enhancement is to manage rendering and logic on the Server so that we have full support across the web platform, when JS is available try to emulate the API if the API doesn't exist. This is a relatively sane model that we can all pretty much agree on.

My terrible mental model is: Server → JS (+ Web ASM maybe) → Web API; and it is this bit on the left and in the middle that is interesting to me because in an ideal world we would share the code between the client and the server.

The Shape Detection API is pretty simple and it belies the complexity that lays beneath. There is a huge amount of logic that goes on in any form of image processing and it is certainly non-trivial and to support the Facial Detection part of the Shape Detection API progressively the model might look like:

  1. Try to get Open CV on the server via a C-binding into nodeJS
  2. Try to get in Open CV on to the client (emscriptem/wasm) so that we have the same API on client and server.
  3. Try to create some wrapper around Open CV and the target API.

For example the OpenCV binding

cv.readImage("./examples/files/mona.png", function(err, im){
  im.detectObject(cv.FACE_CASCADE, {}, function(err, faces){
    for(let face of faces) {
      im.ellipse(face.x + face.width/2, face.y + face.height/2, face.width/2, face.height/2);

vs the API for the Web.

var faceDetector = new FaceDetector();
  .then(faces => {
    ctx.lineWidth = 2;
    ctx.strokeStyle = 'red';
    for(let face of faces) {
      ctx.rect(face.x, face.y, face.width, face.height);

The logic for outlining the detected face is the same, and the API if you squint is nearly the same, but the API's are obviously different.

I believe it would be relatively easy to provide a mapping between the "Shape Detection API" and the OpenCV API. You could then have this binding shared between web client that has no native implementation and then the server binding too.

As far as I understand it, the NodeJS team are waiting for the v8 team to land support for WASM and once it is in, then it is likely to be a done deal.

I am excited by Web Assembly for many reasons, one of them is that I think we have a great chance of having one "binary" that can run on the client and the server that is easy to deploy and maintain whilst at the same time allowing us to have consistent results across server and browser.

I lead the Chrome Developer Relations team at Google.

We want people to have the best experience possible on the web without having to install a native app or produce content in a walled garden.

Our team tries to make it easier for developers to build on the web by supporting every Chrome release, creating great content to support developers on web.dev, contributing to MDN, helping to improve browser compatibility, and some of the best developer tools like Lighthouse, Workbox, Squoosh to name just a few.

I love to learn about what you are building, and how I can help with Chrome or Web development in general, so if you want to chat with me directly, please feel free to book a consultation.

I'm trialing a newsletter, you can subscribe below (thank you!)