>They need to publish their full article content for search crawlers to read At first, I thought it might be possible to solve this by getting search engines t

>They need to publish their full article content for search crawlers to read At first, I thought it might be possible to solve this by getting search engines to standardize on a vector format that they could accept for protected content. So the crawler sees a 300-dimensional vector that effectively gives a semantic summary of the document. But then I thought content providers could achieve a substantially-similar effect by just serving their documents to search engines in scrambled (e.g. alphabetized) form. They could still provide normal headlines to get them clicks. BUT THEN I thought it would be a really cool and bizarre problem to circumvent this by attempting to devise a method for finding the most-probable original document given its alphabetized version.