Hacker News Reader

Comment by edmundsauto | Hacker News Reader

It looks more like a ToS claim; they aren't claiming to own the lyrics. They aren't objecting to Google scraping their site for the index -- they object to Google/LyricFind scraping the lyrics, removing hidden watermarks to prove provenance, and displaying the lyrics in whole on SERPs, without attribution.

The removal of watermarks, after repeated contacts and the WSJ article, indicate that Google/LyricFind acted in bad faith.

The best part is they encoded the watermark so that Watermark #1 spelled out REDHANDED and Watermark #2 spelled out GENIUS, using a pretty clever method that survived multiple data transfers/transformations.

View on HN

I mean, from my reading of the complaint, the waterproofing method was rather crude. They just simply changed either the apostrophe character or the space character to different characters that are visually the same.

Any system that wasn't actively trying to clean up "trouble" characters would leave it alone.

If the watermarks are just text oddities, wouldn't it make sense that they were removed during some sanitization process?

They weren't removed at first - that's how Genius confirmed the lyrics were being lifted. They were only removed after the WSJ article.

Then Genius embedded a second watermarking scheme, and -- drumroll! -- Google's results only removed one of the two sets (the ones they knew about).

Ok - but what's the issue with that?

I don't think anyone is debating at this whether or not Google was in fact copying the genius lyrics, just whether or not that was in fact breach of contract or otherwise a violation of the law.

Changing fake apostrophes to real apostrophes after they found out their (in their mind legally acquired) data had them is just plain sensible, fake apostrophes break functionality like search for users (both ctrl-f and possibly more sophisticated search engines like google.com). They add edge cases to any software that wants to use them for text to speech or whatever. And so on. Literally the only use case fake apostrophes don't impact by at least adding an edge case is humans reading song lyrics that they have already found.

The issue is intent, did they remove the watermark as a matter of input sanitization, or did they remove it to conceal their behavior. Genius claims it was the latter.

Reply on news.ycombinator.com