Can you explain this a little more? Do you load the rasterized image using a base64 url, or do you store specific drawing instructions in JavaScript (lineTo, m

Can you explain this a little more? Do you load the rasterized image using a base64 url, or do you store specific drawing instructions in JavaScript (lineTo, moveTo, etc)? For the vectorized icon, do you load SVG inside canvas, or just display SVG directly? Also, how do you fall back for older browsers that do not support Canvas or SVG?