ICANN/DNSO
DNSO Mailling lists archives

[ga]


<<< Chronological Index >>>    <<< Thread Index >>>

Re: [ga] Critics say "there's no shortage of dotcoms"


On Wednesday, April 11, 2001 6:22 PM (AEST)
Gavin Stokes <gavin.stokes@autodesk.com> wrote:
Subject: [ga] Critics say "there's no shortage of dotcoms"

> Patrick, when you say:
>
> > Shortage?    There is no shortage of dotcoms.
> > Inflation?      Irrelevant.
>
> What do you mean by "dotcoms"?  If you mean domain names (and that's the topic
> here), it appears that a lot of people disagree with you.  After all, aren't
> we talking about adding TLDs (in part) to alleviate this perceived shortage?
> And here's a clip from an article on CNet:
>
> "Capitalizing on the shortage of domain names, the Kingdom of Tonga--the only
> inhabited South Pacific island never to be colonized--has begun hawking the
> ".to" designation to Netizens."
>
> And if artificial scarcity and inflation are irrelevant, I guess we can clear
> the books of all those wordy antitrust laws.

Hi Gavin

You have said a number of things which indicate that you are not well informed
on the subjects you have been discussing.  This is not meant as a criticism,
just a statement of fact.

And I do not really wish to enter in overlong debates on issues that have been
discussed at length on this and other lists.  Just one or two points are worth
clarifying.

(1)    I said "dotcoms" because the general thrust of some people's argument is
that there is a shortage in that gTLD.  Certainly there is no shortage of domain
names at the second level.  Of course a number of these are "taken" already.

(2)    You say that "people disagree with me".  People who understand the
subject don't.  From memory 67 characters are allowed including the .com which
is four characters.  That means you can have 63 alphanumerics before the dot.

Since there are 26 alphabetics, ten numerics and a hyphen, that's a BIG number.

On this issue I'd refer you to the thread "Calculation of Domain Name Bounds" on
Verisign's Domain-Policy mailing list last year.  See the Michael Krieger and
Russ Smith postings below (with my initiating email) as typical examples.

The URL for the archives is http://lists.netsol.com/archives/domain-policy.html

(3)    Giving people "more choice" does not mean there is "a shortage".

(4)    You didn't explain why you think "inflation" is relevant to the perceived
shortage of domain names.  The "antitrust laws", whilst interesting, are not
relevant to "inflation" either.  Please don't try to confuse the issue.

Regards
Patrick Corliss

----- Original Message -----
From: Michael M. Krieger <MKRIEGER/0005975596@MCIMAIL.COM>
To: <DOMAIN-POLICY@LISTS.NETSOL.COM>
Sent: Wednesday, November 29, 2000 8:41 AM
Subject: Re: Calculation of Domain Name Bounds

   The "exact" number of possiblities really doesn't need calculation
as it's more than enough not only to give each electron in the universe
(which I vaguely recall as something like 10**40) it's own IP address,
but even a private channel from each electron to every other one.

   E.g., using Bill's 255 limit, and assuming the maxiumum number of
"."s (which maximally constrains the choices), we can count the number
of addresses of alternating dots and single letter subdomains.  This
is 26**127, i.e., 26 to the 127th power.  Here we've only counted
names of the form  .x.y. ...  .z.   (127 single letter subdomains between
128 "."s.

   Already an out of the universe number, as is the number of "labels"
that can be done just at the 2nd level in, say, 60 characters, namely
something between 26**60 and 27**60  (fuzzed by the slight limitations
of using "-"s).

  Of course most of these are unmemorizable nonsense strings.  But
even if one were to limit oneself to say, the 100,000 words in the OED
of 10 letters or less (I'm guessing here about the '100,000' just to be
able to give an approximate example) - and only use 2nd level domains,
that would be 6 "blocks" with 100K possiblities in each, i.e.,
(10**5)**6 = 10**30, which ain't a shabby number either.  And these DN's
would be much more mnemonic.  Take a look at the passwords on the free
AOL disks to see the rich turf of pronounceable strings of just
10 or so letters).

  All of which long ago led me to suspect that there are enough domina
names in .com to go around if we relax a bit about wanting the "perfect"
one.

   Michael

   Michael M. Krieger
   Krieger & Nunziato llp
   UCLA Computer Science
   Editor, "New Matter" -
      Intellectual Property Journal of the State Bar of California

   mkrieger@mcimail.com

----------
Date:     Tue Nov 28, 2000 10:11 am  PST
From:     Patrick Corliss   MBX: patrick@QUAD.NET.AU

TO:       DOMAIN-POLICY     MBX: DOMAIN-POLICY@LISTS.NETSOL.COM
BCC:    * MICHAEL M KRIEGER / MCI ID: 597-5596
Subject:  Calculation of Domain Name Bounds

William X. Walsh wrote:

>  The fact is that those of us telling you this are trying
>  to do you a favor.

Hi William

Can you do me a favour, please?  I'd like to ask you a couple of technical
questions.  Maybe others can help?  It's a bit of theory.

(1)    If a TLD can have a 2LD etc, how many levels can you go to?  Is the
limit physical or logical?  In other words, counting "DOT" how many dots are
allowed in a multipart domain name.

(2)    If the 2LD can go to 63 or 67 characters, does this restriction apply
to 3LDs etc?  If every level can have 63 characters, then what's the maximum
string length counting "DOT" and including all of the other dots.

(3)    Given the physical limit set out about, do any of the files have a
physical limit.  I remember using TPF in Qantas where block size was 1 or 2 K.

(4)    Using the number, hyphen, characterproblem, what is the maximum number
of domain names that could possibly exist?

Thanks William & Others

Best regards
Patrick Corliss

----- Original Message -----
From: domainiac <domainiac@HOME.COM>
To: <DOMAIN-POLICY@LISTS.NETSOL.COM>
Sent: Wednesday, November 29, 2000 10:03 AM
Subject: Re: Calculation of Domain Name Bounds

>36 + 36*36 + 36*38*36 + 36*38*38*36 + 36*(38^3)*36 + ... +
>36*(38^249)*36 - [all combinations that have two dots in a row
>which I would have to dig out a combinitorics book to calculate]

On second thought the second part is not that difficult.  You have 249
characters but the first and last cannot be "."-" so you are left with 247
"interior" characters.  Now think of two dots in a row as one fixed item.
You now have this fixed item plus 245 variable characters (and a variable
length of the string)

1*(38^245) + 38*1*(38^244) + (38^2)*1*(38^242)+ ... (38^245)*1 +
1*(38^244) + 38*1*(38^243) + ... (38^244)*1 +
1*(38^243) + 38*1*(38^242) + ... (38^243)*1 +
...
1*(38^2) + 38*1*38 + (38^2)*1
+ 2(38)
+1

= 246(38^245) + 245(38^244) + ... 3(38^2) +2 (38) + 1

-----

number of .com host names (I think) = [36 + 36*36 + 36*38*36 + 36*38*38*36 +
36*(38^3)*36 + ... +
36*(38^249)*36] - [246(38^245) + 245(38^244) + ... 3(38^2) +2 (38) + 1]

Russ Smith







--
This message was passed to you via the ga@dnso.org list.
Send mail to majordomo@dnso.org to unsubscribe
("unsubscribe ga" in the body of the message).
Archives at http://www.dnso.org/archives.html



<<< Chronological Index >>>    <<< Thread Index >>>