At some point in the evolution of the web someone invented the idea that you could leave the protocol off of a URL in a <script> element and it would inherit the protocol of the page it was referenced within.
This is probably an effort to be compatible with HTTPS.
It shows up in embedded tweets. When you get the code from Twitter, at runtime, it's of the form:
Here's the problem. The page is a file:// page.
So the address doesn't work.
I don't really want to rewrite the page so it loads it over HTTP, but if I have to I guess I will.
Hard to search for the answer on the web because, well how do you describe it?
Update: The workaround.