September, 14 1997
From: Benjamin Azevedo
To: notice-97-02@gtld-mou.org
Subject: A New Concept: sTLD (Standard TLDs, externally managed)
Summary: sTLD Standard TLD, externally managed
.i International Telephone Numbers based Addressing
.tm International Trademark TLD
.EAN International EAN/UPC based Addressing
Other sTLD's examples
Multiple addressing: Domain Aliasing
Immediate action: Issuing sTLD's and new gTLD's
Dear Sirs,
I've been since last June dedicated to the problem of new gTLDs.
I've made some postings in specific discussion lists,
I've been listening to the ongoing discussion,
and I also researched previous proposals.
I suppose I found a missing link in TLD creation.
I soon arrived to a new concept that will certainly make it easier
to adopt a radically different direction in this matter.
A concept that requires no technical change, eliminates conflicts,
might be started immediately and adds a lot of functionality.
During this time I've tested the concept against other ideas
and against common objections.
It seems to respond properly to all those circumstances.
Many previous proposals arrived very near but did not be
recognized as good solutions, mostly because of the current
focus on "generic" TLD.
The best of these gTLD are close to some examples of sTLD.
Let's explain what is the new concept.
Generic TLD are sets of names, not linked to any external
rule, assigned in a first comes first serve basis.
They are no more than (suffixed) nicknames.TLD
Conflicts arise exactly because of this: there is not any
basis for someone get a domain name, except asking for it first.
But people start using them as if they were "companies' names".
But there are far more companies with the same choice name
than different gTLD's to properly serve them.
Trademark owners soon start fighting against domain name owners.
I am sure these are two different things.
Trademark owners do own a name as a mark within a product class, within
one or more geographic regions.
Domain names owners do own a name as domain within a specific TLD.
Neither one owns the NAME itself, but rather that NAME within a
very specific name space.
Separated name spaces.
As well as the same name that is your trademark may also
be the trademark of someone else in another class or country.
And your name.com may well have a pair on name.net.
This seems very clear, except for someone that doesn't
get the domain name (in .com) that matches with his trademark.
And don't have another proper gTLD to use.
There is just one ".com". Everybody prefers to be there.
In some cases a ".net" or ".org" domain solves the problem.
But in many case doesn't.
As I said, the problem is that so called "generic TLD" namespace
has been allocated without any relation to anything.
It has been exactly as it is named, "generic".
So it is just a set of names.
And they are a very poorly used namespace as although it's a long
alphanumeric address, people just want to use a few meaningful
and preferably short addresses.
What a waste of name space addressing !
My new concept is to add sTLD (Standard TLD's)
opposed to "generic" gTLD's.
I define sTLD as a TLD that IS based on an externally existing
and widely accepted STANDARD.
It's absolutely a surprise that all existing standards hadn't yet
be used as international TLD.
They do not require any structural change in Internet dns.
It's just a matter of choosing as TLD already resolved naming spaces.
A sTLD will be created in exact match with the standard it mirrors.
All and every conflict is dealt by the existing standard board.
Indeed we should expect NO conflicts.
These standards provide many thousands/million unique codes.
Let me give some examples to make things clear:
1) Telephone sTLD ".i"
Maybe called ".i", that stands for "i"nternet enabled telephone
number.
Any advertise or business card would just bring the telephone listed
with a trailing "i" to indicate that it might be used for web addressing
as well.
Like John Doe
+1-212 234-5678 i
WebAddress: www.1212-2345678.i
I think this would be a very popular and useful sTLD.
And it's specially easy to adopt in conventional communications,
as ads, catalogs, directories, and regular phonebook software.
Researching for variations of this idea, I've found propositions as
.num gTLD by M Schultz (Internet Draft RFC 1591)
and also
.bus / .res gTLD by Ed Tsunoda (gTLD Comments Sep 13, 1997)
Both address the same objective, but are slightly US-oriented,
missing
the mandatory country-code, in order to make it really standard.
What made proposals like .nom gTLD not acceptable is that they were
just
another gTLD, and it was not reasonable to start any new gTLD, as many
other
were also asking for it.
Now I am stablishing a new type of TLD, sTLD, that is different from
old gTLD, and might well be added, independently of gTLD discussion.
This ".i" sTLD is not another new proposed regular gTLD, but rather
the start of a new class of TLD's: namely sTLD.
Viewing it as a standard TLD (sTLD) makes it easier to accept it.
Each telephone number in the world is a unique identifier,
without any conflict.
This is true if taken with the whole expression
Country + area + exchange + sequential or simply
Country + area + local number
It might be established that "-" are meaningless in this sTLD, so we
can use them more freely for easier reading.
A US number +1 (202) 234-5678 would correspond to:
1202-2345678.i or 1-202-234-5678.i
A Brazilian number (55) (21) 234-5678 would correspond to:
5521-2345678.i or 55-21-234-5678.i
And for instance BT's UK number +44 1179 217721 would correspond to:
441179-217721.i or 44-1179-217721.i
Please note that this sTLD, as well as any other sTLD,
it is NOT required any technical change for its immediate adoption.
And no conflict is to occur.
We have about 150.000.000 unique, pre-assigned, telephone numbers
operating in the world.
And more important yet, a whole structured search structure, in the
form of directories, directory assistance, Internet telephone search,
etc.
If you want to know Internet address, just get the phone number.
Or get through your directory assistance service, company's Internet
enabled telephone number.
It's really very appropriated.
A new gTLD that starts with its domains pre-registered.
Without any conflicts.
And with multiple sources for information services.
And the same applies in large scale to other sTLD as well.
One objection is that many people would not want to have their
domains known as a number.
To exclude this objection, please look for "Multiple Addressing:
Domain Aliasing", later on this note.
This telephone domain may point (as an alias) to effective named gTLD
domain.
2) Trademark sTLD ".tm"
This is a widely discussed proposal, that I also do believe hasn't
yet be adopted (internationally) in part because of the until now
accepted notion of "generic" TLD.
Some countries successfully use variations of it.
To match with Trademarks actual structure, each country would have
its own .tm.country sTLD.
At international level, it would be created a ".tm" sTLD
that would be used solely by those that have a internationally
recognized
mark. It should be also noted the class(es) in which it is recognized.
Nomenclature for class may be choosed to be numeric codes or names.
For those internationally recognized trademarks, it would be allowed
to use corresponding .tm domain name. Let's list very generic examples:
benetton.class25.tm coke.class04.tm
benetton.cloths.tm coke.beverages.tm
benetton.tm coke.tm
The assignment directly under .tm only would be made in the case of
clearly undisputed names.
Should it exist more than one occurrence of a given name, then
http://name.tm may show a clickable list, within corresponding classes.
This domain would constitute itself a valuable source of trademark
information too. Easy and convenient checking of registered marks.
At country level it should be promoted the use of similar
criteria so that trademarks owners do have a proper way to correspond
their marks with Internet addresses and need not fight for a name.TLD.
We should provide also additional information at upper levels in
sTLD: such as having http://www.tm showing available classes.
Objections against being those domains too long, should be answered
that it's barely a direct reflex of trademark hierarchies.
See also "multiple addressing" that offers a relief against these
problems.
3) EAN sTLD ".EAN"
Almost every industrial consumer product is worldwide identified by
its
EAN/UPC code.
This code uniquely identifies not only the product but also the
manufacturer.
Each manufacturer has a unique code.
What if using eancode.EAN you get an information page on that
product?
And using manufacturer.EAN you get the manufacturer page?
Again we have this way a very easy tool to locate a site/page.
Using a "key" that is right in visitors hand, like EAN code.
The same for info about your diet softdrink, or anything else.
Just enter the eancode.EAN to have exactly the information you want.
4) Other sTLD examples
The strength of the concept of sTLD is that it's very broad.
We have hundreds (or more) of established standards in the world that
might well serve as comfortable Internet addressing.
These standards are the result of endless work in corresponding
standard management institutions.
It would be very difficult to have a registrar able to solve
all the aspects that these boards around the world do.
That is WHY the gTLD problem seemed so difficult and full of
conflicts.
All of we were forgetting to use already established conventions.
A few more examples would be sufficient in this concept presentation.
4a)ISBN code for Books
Pointing to a standard info page with author/editor links.
4b)ISSN code for Periodicals
Permitting direct access to publications,
and eventually specific issues,
without regarding which "times" or "post" you're looking for.
4c)Airlines (IATA/ICAO) indicators and Airports designators
Direct access to United through united.air
regardless that it is not united.com
4d)NYSE bond tickers
Get to the site of your preferred companies,
just using your investor tags.
4e)At country level use IRS codes to get companies' info
Most countries adopt codes like this, that appear in many
documents, and even on product tags.
4f)ISRC for Music CD's
Yes each CD has it's unique code, although today it's
in most cases hidden in CD's inner ring, very hard to read.
There are hardly limits for this concept.
And it is also a strong incentive to wider use of these valuable
standards.
Maybe after establishing these sTLD, standards codes stop being
printed in small fonts.
Maybe you get more easily the ISSN code of your newspaper and
magazine
or the ISRC for your new music CD.
5) Multiple Addressing - Domain Aliasing
Many of these sTLD would better work as "subsidiary" aliasing
addressing for companies sites and inner pages, depending on the case.
They will mostly act as an alias for corporate sites or pages within
these sites.
In order to make clear this concept, I want to emphasize that
by multiple addressing I am referring to have a request to a domain name
lead the caller to this domain, or to another domain, or any page in
another domain.
This option would be set in name server.
For instance if we create a 074101-791013.EAN that correspond
to Fujitsu 100Mb disk for Zip Drives, Fujitsu will not need to set
a new site for it, but just make it load proper product page on
Fujistsu site.
If visitor use just the manufacturer code he will get Fujitsu's
site.
http://074101-791013.EAN -> Fujitsu ZipDrive 100MB Disk page
http://074101.EAn -> Just an alias for Fujitsu site
Direct access, no need to search for it.
Please note that searching has been very useful, but had
the side effect that made us all live together with this
uncomfortable routine of trying to guess a domain name,
or trying to guess appropriated keywords to have a manageable
search result.
With these sTLD, and many others, people will starting being able to
go directly to related site without having to make any guess,
without any search.
Just use what you know: their phone number, product EAN code,
the brand mark you remember.
6) Immediate action - Issuing sTLD's and new gTLD's
My proposal is to create immediately several sTLD,
using widely accept international standards as basis for them,
before creating any new gTLD.
Suggested names for some sTLD in this proposal are the most natural
ones.
Some objections may arise against these sTLD names themselves.
Such as ".i" to be a single letter sTLD.
Or ".tm" to be already the ISO3166 country code for Turkmenistan.
Exact sTLD might be changed, or otherwise, some previous conventions,
as .tm.
Only after established the use of sTLD, and as a consequence current
pressure over gTLD is lowered, we could start thinking about creating
new ones.
Maybe several ones indeed.
Differently from sTLD that have natural assignment of domain names,
gTLD do not have and need special regulations, in order to prevent
misuse of newly added gTLD.
It is not desirable that current domain owners in ".com" feel
pressed to get corresponding domains in new gTLD.
This would make adding new gTLD non-sense, except for lucky
registrars.
It should not be this way.
Either should we, not now, but in a near future, when it's arrived
the time for adding new gTLD:
5a)Charge for the right to manage a new gTLD, permitting
registrars to
make afterwards a pre-registration auction to sell for the best
bid
names that candidates do not want to risk loosing in regular,
first
comes first serve basis.
5b)Allow so many gTLD that nobody would think about making
corresponding
registrations in every gTLD.
As gTLD are fresh new name space, just gave them away is a sort of
wrong doing, as if individual names are considered valuable assets, it's
easy to realize that a fresh new gTLD is a highly valuable asset.
They should be auctioned accordingly.
If they are to be issued in small quantity, each one should be sold
at premium price.
If they are to be issued in large quantities, each one would be less
expensive.
In any situation, money paid for these "licenses" should be used
in promoting Internet progress, either providing end user additional
means of accessing it (specially in development countries),
or investing in Internet structure,
or still rewarding outstanding contributions to Internet progress.
This auctioning is not necessary for sTLD, as they should first to
show wide acceptance in order to qualify for a sTLD.
It's not like a gTLD that has no intrinsic meaning other than to
be a set of nicknames, with only chronological precedence. Like current
.com/.org/.net gTLD's, and most of proposed new gTLD's and alternative
proposals as well.
I am finishing a complete work about the sTLD concept, with other
examples and specially the answers to foreseeable objections.
As you might see later, the more interesting is about why adopt
numeric addresses if domain names were there exactly to mask IP
addresses.
Seems to be a good point but there is a strong reason why this is
not a step back, but rather a new path. I'll show you. Very soon.
Should you have a suggestion or objection, please give me a
feedback, through email or snail mail, as follows.
Benjamin Azevedo
Caixa Postal 10870
Copacabana
Rio de Janeiro RJ
22022-970 BRAZIL