NOTICE-97-02
ACTION: REQUEST FOR COMMENTS
CLOSING DATE: October 13, 1997
SUBJECT: Review of new generic Top Level Domains (gTLDs)
B. GENERAL QUESTIONS
1. Number of gTLDs
Generally, how many new gTLDs do you think should be created?
[0 __] [1-10 XX] [10-50 __] [50-150 __] [150 or more __]
---------
> Note: the above answer is for the first year only. Experience with the
> initial release of new names will help answer whether more or less
> should be created in subsequent years.
C. CHOICE OF gTLDS
1. Review of IAHC-Suggested gTLDs
How would you rate each of the following gTLD names selected by the IAHC?
a) .firm (intended for businesses, or firms), check one of the below:
[strongly approve __] [approve XX] [no opinion __] [disapprove __] [strongly disapprove __]
b) .store (for businesses offering goods to purchase), check one of the
below:
[strongly approve __] [approve XX] [no opinion __] [disapprove __] [strongly disapprove __]
Special note: There have been many suggestions that .store be replaced by
.shop which has significance in multiple languages and, unlike .store, has
related verb and noun meanings. Do you think .store should be replaced by
.shop?
[yes XX] [no __]
> Shorter is better and if .shop is more meaningful in more languages,
> then I would say go for it.
c) .web (for entities emphasizing activities related to the World Wide
Web), check one of the below:
[strongly approve __] [approve XX] [no opinion __] [disapprove __] [strongly disapprove __]
d) .arts (for entities emphasizing cultural and entertainment
activities), check one of the below:
[strongly approve XX] [approve __] [no opinion __] [disapprove __] [strongly disapprove __]
e) .rec (for entities emphasizing recreation/entertainment activities),
check one of the below:
[strongly approve __] [approve XX] [no opinion __] [disapprove __] [strongly disapprove __]
f) .info (for entities providing information services), check one of the
below:
[strongly approve __] [approve XX] [no opinion __] [disapprove __] [strongly disapprove __]
g) .nom (for those wishing individual or personal nomenclature, i.e., a
personal nom de plume), check one of the below:
[strongly approve XX] [approve __] [no opinion __] [disapprove __] [strongly disapprove __]
Special note 1: Since this gTLD is for individuals and registrations could
number in the millions, if not hundreds of millions, there is naturally
concern about the scaleability of a personal name space gTLD. There have been
suggestions that in order to improve the scaleability of the .nom name space,
a numerical identifier should be added to the .nom gTLD from the outset (e.g,
.nom1, .nom2, .nom3, etc.) Do you agree?
[yes __] [no XX but see below]
> I believe that this would be a particularly good TLD in which to start a
> much needed trend towards multi-level gTLD names. In particular, most
> persons in the world (although of course not all), have multi-part
> names one part of which is a "family" or inherited name. Of the other,
> non-inherited or "given" name parts, there is usually one commonly used
> part (the "first name" in some cultures). I would propose that only two
> level delegations be name in .nom. For example, the most natural such
> name for myself would be Donald.Eastlake.nom. To make this really work
> well, there should be some greater barrier to creating new second level
> labels while it should be easier to create third level ones. Possibly
> a financial differential, which would make the more common addition of
> a new first name particularly inexpensive, or a requirement to send in a
> paper application with a birth certificate if you want to create a new
> family name in .nom or the like.
> I'm inclined against having numbers in the .nom domain names, but not
> very strongly. If numbers were to be included, they should be a
> separate label, just below .nom. For example, Donald.Eastlake.3.nom.
> In that case delegations would actually be at the fourth level. (Above
> I am refering to the delegation to the individual. Of course there
> could be internal delegations in the .nom TLD above the outside
> delegations for the convenience of operating the TLD.) This additional
> label could also be used to accodate the few people with a single name.
> For example, smith.0.nom to distinguish the delegation point to the
> person known by the single name smith from john.smith.nom,
> john.smith.2.nom, or whatever for the two-part-named people with family
> name smith.
2. Creation of gTLD Indicating Adult-Oriented Content
There is often the suggestion made that a dedicated gTLD indicating the
equivalent to a "red-light-zone" would be useful in isolating and filtering
adult-oriented content from minors on the Internet. There are differing
opinions on how effective this would be. The most common suggestion is the
request to create a gTLD .xxx. Do you think that a sort of "red-light-zone"
gTLD should be created?
[strongly approve __] [approve __] [no opinion __] [disapprove __]
[strongly disapprove XX]
If you strongly approve or approve, do you think this should be the gTLD
.xxx?
[yes __] [no __].
If you have replied "no", what gTLD would you suggest in place of .xxx in
this category? ______
> Excuse me, but saying I don't think a '"red-light-zone" gTLD' should be
> created does not imply that I think there is a better TLD name for it
> than xxx. Therefore I have no answer to your second question above.
> It is absurd to believe that there is a binary decision as to what is
> "adult-oriented" or the like. Standards differ widely in the global
> arena. The early existence of such a TLD would result in enormous
> pressure to force content providers to censor themselves by moving to
> this TLD, endless hassles about just what criteria should be used,
> attempts to criminalize the offering of even vaguely indecent content
> anywhere else, etc. In short, it is a recipe for disaster. And even
> if a simplistic "xxx" label were a good idea, there is no strong reason
> for it to be a TLD. It could be anywhere in the domain name. In fact,
> most "network nanny" filtering programs *already* look for xxx anywhere
> in the domain name. Even a company that wanted to be so labeled would
> probably have one or two web pages that just gave their street address,
> phone number, etc., that there is no reason to hide. So they should be
> able to use www.company.store for that and ...xxx.company.store for
> their adult stuff, for example. This is substantially more flexible
> than a TLD.
> In short, self-labeling "adult" content by TLD is a bad idea likely to
> lead to censorship and forced speech. Systems like PICS are the right
> way to label things. And even if you do want to use domain name to
> label adult content, there is no need for a TLD as it is more flexible
> to put xxx elsewhere in the domain name.
3. Other Suggestions
Please make any other suggestion concerning the names of the new gTLDs
not covered above.
> Below are three numbered suggestions specifically related to the names
> of new gTLDs. However, I would first like to say that how names are
> allocated within the gTLDs is more important than the names of the
> gTLDs. I can understand that with NSI continuing to run virtually
> unrestricted one level registration in .com, .net, and .org, you would
> want some similar TLDs under the gTLD-MoU to compete head on. But in
> the longer run, I think there is a strong need for gTLDs in which (A)
> there are name collision rules and/or (B) there are two level names
> registered. To expand on these:
> (A) One purpose of names in a gTLD is to be mnemonic and not trivially
> confused by other names which are phonetically or typographically
> extremely close. In particular, I believe there should be gTLDs in
> which an applicant for a name which is a near miss with one or more
> existing names in that gTLD should need the existing name holders
> approval (unless the name is very short). A near miss could be
> defined as a SOUNDEX conflict with an existing name or a name which can
> be formed from an existing name by a single typograhic character
> insertion, deletion, or swap of adjacent characters. (The typograhic
> criteria is probably too harsh for names under 4 or 5 characters and
> the SOUNDEX criteria is probably too harsh for names under 2 or 3
> charcters.) A disadvantage, of course, is that this limits the number
> of names that can be added to a gTLD without permission from earlier
> registering name holders.
> (B) Registering two level names seems like a natural way to avoid the
> sort of extreme congestion that .com is experiencing. In fact, many
> entities are registering what is almost a two level name in .com, it's
> just that since only single level name registration is provided for,
> they are using a hyphen or the like between the two parts of their name
> instead of a DNS label separating "." . It just seems very natual to
> register things like Acme.Enterprises.tld. Note that combining this
> with A above, by enforcing the non-confusion rules on the second level
> label and on the third level label but only for the sets of third level
> labels with the same second level label, squares the number of names
> that can be added without permission from earlier registering name
> holders. However, I believe the two level name resigration idea, which
> I am by no means the only person to have thought of, has merit by
> itself. See my comments above in reference to .nom.
> (1) Name collisions: To avoid confusion among TLD names that are >2
> characters long, they should be choosen so that no two have the same
> SOUNDEX code. Domain names are sometimes spoken over the telephone
> and/or remembered and confusingly similar names should be avoided.l
> (None of the suggested names conflicts.)
> (2) Avoid 1 character names as well as two character names: Some
> possible technical steps to handle a very large root zone would be made
> easily if all 1 character TLDs were available. And of course two
> character names are reserved for use in connection with ISO 3166
> country codes. (None of the suggested names violates this.)
> (3) Avoid names of countries, cities, geographic entities, etc. Using
> such names will, at least for prominent names, risks severe conflict.
> (None of the suggested names violates this.)
_________________________________
[END]
Donald
=====================================================================
Donald E. Eastlake 3rd +1 978-287-4877(tel) dee@cybercash.com
318 Acton Street +1 978-371-7148(fax) dee@world.std.com
Carlisle, MA 01741 USA +1 703-620-4200(main office, Reston, VA)
http://www.cybercash.com http://www.privacy.org/ipc