New Idea on Visual Thumbnail

[Xinan]

Hi all!

I've come up with a new approach on the Visual Thumbnail of Projects.
Since either rendering it in the browser or on the server has its drawbacks, why don't we combine these two approaches?

My idea is that, we try to get the thumbnail from the browser side by capturing the output canvas. If it fails (program has not been run), we fall back to use server side rendering.
This can ensure that most of the time, the thumbnail is just as what the student sees in the output canvas. And even if the program is not run, we can also create a thumbnail for it. Since usually student would run their program to test before saving, so there will not be too much load for our server. In this way the two approaches complements each other and we will be able to overcome each of their challenges.

I'm working on a proposal based on this approach. Any suggestions?

Thanks!

Best,
Xinan