Time to play devil's advocate and do a thought experiment. What if we created a site containing every word and every conjugation in a way that could be linked to. This work is clearly not a copyright violation. What if we then wrote a "compression" program that took any text file and converted it to a file containing only links to this site. Few would argue that is not a copyright violation. Yet the resulting file contains nothing of the original, it is just a link. If linking to a whole text with a single link is not a copyright violation, at what point between that and our "compressed" file does it cross the line, and why?

Or, put another way, if links had been copyright violations from the start, would we think that it was a sensible and logical characteristic of copyright?