With Google aiming to make the robots.txt exclusion protocol a standard, they proposed some changes and submitted them the other day. Now, Google updated their own developer docs around the robots.txt specification to match. Here is a list of what has changed.

  • Removed the “Requirements Language” section in this document because the language is Internet draft specific.
  • Robots.txt now accepts all URI-based protocols.
  • Google follows at least five redirect hops. Since there were no rules fetched yet, the redirects are followed for at least five hops and if no robots.txt is found, Google treats it as a 404 for the robots.txt. Handling of logical redirects for the robots.txt file based on HTML content that returns 2xx (frames, JavaScript, or meta refresh-type redirects) is discouraged and the content of the first page is used for finding applicable rules.
  • For 5xx, if the robots.txt is unreachable for more than 30 days, the last cached copy of the robots.txt is used, or if unavailable, Google assumes that there are no crawl restrictions.
  • Google treats unsuccessful requests or incomplete data as a server error.
  • “Records” are now called “lines” or “rules”, as appropriate.
  • Google doesn’t support the handling of elements with simple errors or typos (for example, “useragent” instead of “user-agent”).
  • Google currently enforces a size limit of 500 kibibytes (KiB), and ignores content after that limit.
  • Updated formal syntax to be valid Augmented Backus-Naur Form (ABNF) per RFC5234 and to cover for UTF-8 characters in the robots.txt.
  • Updated the definition of “groups” to make it shorter and more to the point. Added an example for an empty group.
  • Removed references to the deprecated Ajax Crawling Scheme.

The big changes are (1) GoogleBot will follows 5 redirect hops (which we knew in 2014), (2) there are no crawl restrictions if unavailable is greater than 30 days, (3) unsuccessful requests=server error, (4) there is a 500 KiB size limit and (5) it supports URI-based protocols.

Here are some additional answers:

Forum discussion at Twitter.

Original Source