Here's a summary of differences of HTML5 Canvas and Scalable Vector Graphics (SVG).
What's the Difference Between Canvas and SVG
Canvas is bitmap based, SVG is vector graphics tech. (this basically means, SVG graphics can be resized without getting fuzzy.)
The language of Canvas is low level and procedural, while SVG is declarative and higher level. (you can think of them as imperative lang vs functional lang) In procedural way, you give commands to draw as if telling a paint brush where to move. In declarative style, you state your graphics shape and size. ⁖ circle(‹center›, ‹radius›)
with Canvas, once a image is drawn, it's dead image, like paint on wall. With SVG, it's live: changes in code or object parameters are rendered in real-time.
Canvas is relative new, part of HTML5. Canvas is invented by Apple in 2004 as proprietary tech, and is well supported in Apple's mobile devices. SVG is some 10 years old. All current browsers supports SVG, including IE9. (A light version of SVG is supported for years in mobile devices) Canvas is currently the hot topic due to HTML5.
SVG also supports bitmapped images.
What's Canvas Good For?
Because Canvas is very primitive, so it's fast. It is suitable for bitmapped games, such as DOOM, Quake, where all you want to do is painting the screen. (but as of , there's no robust 3D graphics (WebGL) support yet)
For casual games, board games, interactive illustration, charts, SVG is much more powerful and 10 times less code.