Re: [wg-c] Possible solution to lock in?

["'Kent Crispin'" <kent@songbird.com>]

On Sat, Nov 20, 1999 at 08:41:38AM -0800, Roeland M.J. Meyer wrote:
> > Behalf Of Kent Crispin
> > Sent: Friday, November 19, 1999 1:21 PM
> >
> > On Fri, Nov 19, 1999 at 02:02:16PM -0500, Harold Feld wrote:
> > > Why isn't the solution to the lock-in problem simply
> > > getting a new name, while keeping the old one, and gradually
> > > migrating off?
> >
> > Because it doesn't work.
> 
> Oh?

Thank you for your carefully considered knee-jerk.

> > You are amazon.com.  Your domain name is bookmarked in a hundred
> > million browsers, search engines, linked in a million pages around
> > the world.  NSI says, hmm, amazon, we are changing to a usage based
> > model of charging because it is more fair to the little guy, and our
> > measurements show that your domain is very frequently hit, and
> > therefore your domain name will cost you $100,045.39, next year.
> > This amount would cover 100,000/35, or about 2800 years of
> > registration at the old rate of $35/year.  On the other hand, those
> > millions of references that are embedded in the web won't clear out
> > for years, and it will be much longer than one year for amazon.biz to
> > become anywhere near as well-known as amazon.com.
> 
> First off, there is not even ONE instance proof of this scenario, even
> historically.

It was expressly presented as a hypothetical, to illustrate that, 
contrary to Mr Feld's claim, there are possible scenarios where his 
scheme doesn't work.  It was also purposely exaggerated, to make the 
point.  

> Therefore it is sheer theoretical speculation.

Instances of similar activity are legion.  Though I hate to mention
it, the absurd charges for pre-registrations on the part of a few
CORE registrars are very close to what we are talking about.  [Most
of the CORE registrars did not charge for pre-registrations, but
there were a few that did.]  In another variation, we see domain 
speculators...

> Generally, the
> Internet does flat-rate billing. At the higher access speeds, it is mostly
> flat-rate. Only the telcos are even equiped to do volume-based billing. How
> is a registry going to know what a site's hit-rate is?

Who said anything about a "site's hit-rate"?  I was talking about 
dns lookups -- you know -- the stuff that is reported in named 
statistics already?  It would require very little work to instrument 
named so that it dumped dns "hit rates"

> Actually, I'd
> personally love to know because I have an immediate need to be able to do
> this.

I would be happy to do the small amount of custom programming 
required, at my normal rates...

> As regards marketing required to inform public of the new name, it is done
> all the time. Even for some famous brands. Current instance proof is TCI
> cable, where they are changing their name to AT&T cable. It's all a matter
> of proper handeling. Yes, it does cost money. However, note that this was
> done voluntarily by AT&T, they decided that the effort was worth the cost.
> Anyone having a big enough business to afford $20M ad campaigns won't even
> blink at the cost.

You are making the same mistake that Mr Feld made -- TCI cable's 
current name has not been cut off.  They can make this transition 
precisely because they have control of the old name.

[...]

> > In fact, of course, amazon.com has the legal resources to fight such
> > an egregious move.  But a registry can increase profits by 50% by
> > simply raising the already cheap price of domain names for their
> > locked in customers.
> 
> Actual fact is, registration costs are going down over time, not up.

Not willingly.  NSI has fought lowering their prices tooth and nail.

> > The problem is that there is an *enormous* disparity between the
> > value of a domain name to a registrant on the one hand, and the
> > actual cost of maintaining the registration on the other hand.
> > Wannabe registries would love to move to value-based pricing, and
> > will use any opportunity to move in that direction.  Consumers
> > welfare, on the other hand, is served by cost-based pricing.
> 
> We've had this discussion before. Market forces alone will force commodity
> pricing,

How do domain speculators stay in business, then?

>  which is bad enough, but it is at least fair to the operators. You
> are willing to set some arbitrary cost level (which is inherently
> non-auditable) and overlay the whole mess with a thick layer of bureaucracy
> which will cost more than the price of registration.

Nope.

> What I am saying is that, it is your scheme that doesn't work.

It works GREAT.  See Nominet.  No "thick layer of bureaucracy".  
Prices cheaper than NSI, in spite of the fact that it provides more 
service than NSI.  Better service than NSI.  Extensive competition 
between registrars.

-- 
Kent Crispin                               "Do good, and you'll be
kent@songbird.com                           lonesome." -- Mark Twain