[ih] Funny how things work out

Dave Crocker dhc at dcrocker.net
Mon Feb 1 13:19:30 PST 2021

On 2/1/2021 12:55 PM, Jack Haverty via Internet-history wrote:
> Hi Dave,
> I agree that the DNS was never intended as a searching mechanism.  It
> has always done "lookup", converting strings into IP addresses.
> That's a view from the Technology side.  I was thinking as an end User,
> where the History of the Internet looks somewhat different.

While users often misunderstand exact lookup from approximate search, I 
think the issue has always been present.  I don't think there is 
anything different now vs. 30 years ago, regarding this point.

>  From the User side, DNS also, at first, provided two other mechanisms
> useful for Users - Delegation and Organization.  Delegation solved Jon's
> problem, spreading out the work for managing the namespace to multiple
> people and organizations.  Organization provided a means of structuring
> the namespace so that it made some sense to Users.

I'd distinguish administration (assignment) from operation (query) and 
note that your observation about delegation and organization applies to 
each.  Separately...

> As a User, I knew that a school site likely ended in .edu, a company
> site in .com, a US government site in .gov, etc.  So instead of
> "searching", a User knowing that old simple DNS organization of the name
> space could guess that MIT was mit.edu, UCLA was ucla.edu, yahoo was
> yahoo.com, Social Security was ssa.gov, etc.   Names were predictable,
> guessing was reliable enough.

edu had rigorous enforcement, but with problematic interpretation, as I 
recall.  mil and gov, were straightforward and firm, albeit ethnocentric.

com, net and org were never rigorously enforced. So they provided a 
somewhat useful heuristic, but not nearly as good as people tended to think.

> That organizational structure broke as the Internet grew, and name
> collisions became common (like my examples).  

I got a personal domain name about 25 years ago.  crocker.com/net/org 
were all already taken and not by Steve.  So we've been in this scaling 
artifact ever since the Internet went mass-market.  By some metrics, 
that is longer than a human generation.

>  But its utility to Users
> has decayed over time as the Internet grew. 

To the extent it ever had the attribute or search utility you are 
describing, it hasn't had it for a very, very long time.

> Since Marketing cares mostly about Customers (Users), that's what leads
> me to wonder when the marketing forces will recognize that paying lots
> of money to "protect the brand" in the DNS namespace is still worthwhile
> -- except to the companies charging for the naming rights.

Not sure whether it's been obvious to everyone, but we keep seeing 
serious demonstrations that humans (and organizations) can be highly 
resistant to facts.  (My current summary is that we tend to say that 
humans are intelligent, where the correct assessment should be that 
humans are merely capable of intelligence.)


Dave Crocker
Brandenburg InternetWorking

More information about the Internet-history mailing list