Reading this article, I understand that erasure codes are a potential solution to the data availability issue in sharding. That is, a light node can have higher certainty regarding the contents of a block when erasure codes are used.
However, I couldn't find a straightforward explanation of the concept. Can someone shed some light on it?
