Because web sites won't be moving to ipv6 (only) any time soon, and the limiting factor here is residential connections, not Google?
Are you proposing that Google should crawl ipv6-only web sites? And then serve those results to whom, exactly? (I don't think there is any reliable way of knowing whether an HTTP client that's connecting to you via ipv4 is also capable of ipv6?)
> Another effect of the default policy table is to prefer communication using IPv6 addresses to communication using IPv4 addresses, if matching source addresses are available.
If only v6 is available there is no point for them to include the site in their general index.
If both v4 and v6 access is available, what is the extra value of doing the indexing using v6 rather than v4? (The indexing process itself would add very little to the global usage of v6, comparatively speaking.)
Because otherwise we don't need ipv6. Google advertised the use of ipv6 heavily. If everything still works fine with ipv4 and NATs, why should we ever switch? It's a serious question, I know that the ipv4 space is somehow exhausted, but I can still get very cheap ip addresses (one is included in every $5 vps at major providers).
If even Google with its huge network of crawlers doesn't see the need to communicate via ipv6, why should there be any demand?
I guess they could also move this checking to the user agent (because they control Chrome) to have the check performed before the user is actually waiting for results and milliseconds matter. Introduce some ugly but pragmatic HTTP header, like "Accept-IP: v6". :)
So, this way they could serve v6-only results reliably to clients that can access either both v4 and v6. (People who can only access v6.. that must be a very special kind of people.)
Are you proposing that Google should crawl ipv6-only web sites? And then serve those results to whom, exactly? (I don't think there is any reliable way of knowing whether an HTTP client that's connecting to you via ipv4 is also capable of ipv6?)