Well I learned something that's not in the article. I thought browsers requested a domain lookup (gethostbyname()) and basically got back a "zone file" which w

Well I learned something that's not in the article. I thought browsers requested a domain lookup (gethostbyname()) and basically got back a "zone file" which would have the cnames in. So, I was confused, when people complained about Twitter forcing a domain lookup "on the wrong domain" as I was assuming this would at least cache the domain lookup: It's the right domain, of course, but the lookup for an address on a subdomain includes the subdomain and then gets the cname directing wherever. It always confused me that dig/nslookup didn't seem to provide all the info. They can, using nslookup 'ls example.com' or 'dig example.com -t AXFR' but the server in general refuses to serve the zone file (seemingly for security by obscurity reasons). So, for example, if the browser looks up example.com it doesn't get that there is a cname from www->example.com . It only gets that relationship from looking up "www.example.com". So, TIL, and now results provided by dig/nslookup on the command line make more sense!