Earlier quoted context omitted.
My understanding of holographic algorithms is that they make a primitive out of a matching algorithms that could be executed in constant physical time with holographic storage. Actually, from what I understand, that's not the point at all! What is "holographic storage" btw?
http://en.wikipedia.org/wiki/Holographic_data_storage The whole point being that the reference signal can be projected into the storage and match found in a single pulse. Holographic algorithms deal generally with algorithms based on interference, so they might not directly relate to a physical storage device but the principles are similar.
The point is that this paper is about algorithms for classical (e.g., not quantum or anything special) computers. That's why it is so interesting -- it's kind of like quantum algorithms but will work with standard computers.