🚀 go-pugleaf

On Fri, 24 Jan 1992 09:24:48 EST <IGOLDSTE%ALB9370@SED> said:
>On 23 Jan 1992 Valdis Kletnieks writes:
>> All 617,000(*) of them?  Where are you going to *put* a list that big?
>
>While sending and maintaining a list of that size is not pratical, a list
>of the sub-domains (ie mit.edu, Apple.COM, Compuserve.COM...) would be
>helpful.

This is more doable - there's only some 4,000 second-level domains now,
and only 50 or 60 new ones a month...

>Another idea would be for some (one) site to maintain, as best as posible,
>this large list and have some "whois"-like command to query it by mail for
>the site you'd be looking for.

NIC.DDN.MIL keeps a *partial* list - the difficulty in trying to keep this
list up to date caused them to give up once there were 5,000 hosts on the
Internet.

It may be an apochryphal story, but I am told that when they ran the
'DNS shaker' that produced the count of nodes in October of 1990, it
ran all weekend, and they estimated that the number of nodes changed
by several hundred *DURING THE TIME IT TOOK TO RUN THE PROGRAM*.
The Internet is now twice as big, so....

The whole *reason* for the Domain Name System was because it became
impossible to update and maintain a centrally-controlled hosts listing.

The sheer size of the Internet is such that you're unlikely to get anything
much better than the 'whois' service that you mention.

In addition, there is a *strong* cultural factor here - the Internet
functions on a different mindset than Bitnet does.  On the Internet,
it is assumed that you don't *NEED* a list of all nodes, because for
any given resource that you are interested in, if you are authorized to
contact the resource, you are given its name.

If you have a userid you can TELNET to, you're told the machine name when
you get the userid.

If you want to e-mail to somebody, you either get the e-mail address from
a CC: of mail, from netnews, or phone the person and ask...

There's one unaddressed problem, and that's "I remember Joe Smith has
an account on what-the-heck-wazzit.podunk.edu, but..." -- however, this is
better addressed via other means (say, a modified NSLOOKUP that does
fuzzy matching on a zone transfer)....


                                  Valdis Kletnieks
                                  Computer Systems Engineer
                                  Virginia Polytechnic Institute