[bc-gnso] BC comment on singular plural

Smith, Bill bill.smith at paypal-inc.com
Fri Aug 16 20:03:06 UTC 2013

I agree.

I'm not really comfortable letting consumers decide. However, by attempting to make these "similarity" judgements, ICANN is introducing complexity into an already overly complex system. As an engineer (computer scientist) I see this as akin to catching a design flaw during testing or after shipment. It's really expensive to correct an error that could have been avoided by better/different design.

In the case of the new gTLD program, a fundamental aspect of the program is choice and availability of (largely) unrestricted strings. The Community was surprised by the number of applications, perhaps by an order of magnitude. This has stressed a system/process that would have found it difficult to handle even a few hundred new names. But here we are.

If ICANN is going to make rational, reasoned decisions regarding similarities, etc., it almost certainly will need to slow down and consider "similarity" in a broader sense. Mary Jo's list would be a good starting point but I doubt it represents a complete list of considerations.

The new gTLD program fundamentally morphs the DNS from a taxonomy to a market. If that wasn't obvious before, it should be now. Our struggles with this new market come not from its competitive nature but rather how best to "regulate" an otherwise open market in order to ensure the greatest good internationally.

I don't have a good answer for how to accommodate GAC advice, address similarity concerns, or any of a host of other issues that will no doubt arise in the future. If the new gTLD program follows a traditional technology curve, we may will see 25,000 or more applications in the second round. With no limitations on what can be applied for, that's a distinct possibility in my mind.

On Aug 16, 2013, at 12:40 PM, "J. Scott Evans" <jscottevans at yahoo.com<mailto:jscottevans at yahoo.com>>

I think Ron makes some valuable points.

Sent from Yahoo! Mail for iPhone

From: Ron Andruff <randruff at rnapartners.com<mailto:randruff at rnapartners.com>>;
To: 'Smith, Bill' <bill.smith at paypal-inc.com<mailto:bill.smith at paypal-inc.com>>; 'Mari Jo Keukelaar' <mj at johnberryhill.com<mailto:mj at johnberryhill.com>>;
Cc: 'bc - GNSO list' <bc-gnso at icann.org<mailto:bc-gnso at icann.org>>;
Subject: RE: [bc-gnso] BC comment on singular plural
Sent: Fri, Aug 16, 2013 7:05:05 PM

I appreciate Mary Jo’s analysis as well.  The reality is – like or not – as much as the pundits see little dilution in .COM whether 100 or 1000 new gTLDs arrive on the scene, one has to also realize that the English Internet is the global Internet insomuch as English is the unofficial language of business on a global scale.  So the problem we are dealing with is an ‘English’ problem. Which begs the question, how do we – as the stewards of ICANN – find an appropriate solution?  I, for one, am very uncomfortable with Bill’s last comment (Consumers will eventually settle the matter, at least that's the theory.)  Not pushing ICANN to establish policy on this matter is tantamount to tacitly agreeing to force consumers to buy and park names that they have no interest in – or will simply point to their domain (web site) of choice.  Wasn’t that a problem we were hoping to dispense with by way of opening up the domain space?

My two cents…


Ron Andruff
RNA Partners

From: owner-bc-gnso at icann.org<mailto:owner-bc-gnso at icann.org> [mailto:owner-bc-gnso at icann.org<mailto:bc-gnso at icann.org>] On Behalf Of Smith, Bill
Sent: Friday, August 16, 2013 13:50
To: Mari Jo Keukelaar
Cc: Smith, Bill; bc - GNSO list
Subject: Re: [bc-gnso] BC comment on singular plural

Mari Jo,

Excellent post. No easy answers… except for ICANN to abandon what has become and will remain a hopelessly complex and expensive transition of the DNS from a taxonomy-based system to a market-based one.

Where there was once order, we will now have "competition". "Dot chico" and "dot chica" can compete between themselves and with "dot child", "dot children", and a myriad of semantically similar but linguistically distinct TLDs. In a market-based system this is goodness. It will result in an increased pool of available names. Typically an increase in supply without a commensurate increase in demand results in lower prices.

Over time, we may see consolidation of TLDs as competition winnows out the less fortunate in this expansion. We will then have name collisions at the second level which will bring an opportunity for yet more competition and profit making. Regretfully, consumers (in the large) won't really benefit because as a rule they better understand simplicity, order, and stability. Change is not necessarily bad, but it can impact consumers in significant ways, not always positive.

For me, "doing the right thing" would have been a rather more orderly expansion of the gTLD space. Given that that train is far out of the station, we are left with how to deal with the complexity of an arbitrary expansion. Part of the complexity in the "current new" system is the policy decision to avoid similarity. Making such decisions requires judgement and weighing a considerable number of factors in order to arrive at a "common sense" decision. ICANN (the community) works best with a smaller number of factors and ICANN (the corp) prefers to make no judgements.

If we are committed to the market-based approach, perhaps the best path forward is to ignore similarity altogether and allow the market to decide. I say this only partly in jest given the time we have all invested in this discussion and the apparent arbitrary decisions we are seeing in the English-only singular/plural cases. There seems no point in endlessly debating similarity if in the end an arbitrary decision will be rendered. Better to let the market decide and save everyone a great deal of time.

Consumers will eventually settle the matter, at least that's the theory.


On Aug 16, 2013, at 10:08 AM, Mari Jo Keukelaar <mj at johnberryhill.com<javascript:return>> wrote:

It is worth noting that this discussion is unfortunately referred to as "singular/plural" as a consequence of our Anglo-centrism; even there the discussion does not take into account Germanic plural forms in English - e.g. child/children or other northern European English adoptions like wife/wives or mouse/mice, where the SWORD algorithm would likely fail to pick up enough in the line of similarity to provide some evidence of a likelihood of confusion, let alone something impermissible. I, for one, have never been confused by whether "wife" and "wives" mean two very different things, and I hope I've managed to clearly convey this concept to my husband.

As for the subject of what is "common sense," we may all have common sense but we do not have a common language. I would be curious to know what "common sense" dictates, for example, when discussing masculine and feminine forms in Romance languages. In Spanish, chico and chica are as different as night and day. Any boy and girl would tell you the same; they all have cooties. If one doesn't know what "chico" and "chica" mean, however, - which is likely true for the majority of internet users - is the answer, as a policy matter, different?

The answer is very different if you speak Spanish, but the "majority language" of internet users is most likely not Spanish, nor is it English.  If we are seeking to make policy which is generally applicable to the global internet, does it matter whether "chico" and "chica" mean two different things to one language group, while appearing to be pretty much the same thing for everyone else?   If one speaks neither English nor Spanish, "chico" and "chica" probably appear to be more similar than, say, "thing" or "things", which are at least different string lengths.

Quite frankly, to me, all of the Arabic, Hebrew and Chinese IDN TLDs look very much alike since I am unfamiliar with those scripts. Given my ignorance, and applying the principle of “I doubt that I am alone,” shouldn't we also be asking "confusing to whom?"

Unless we are seeking to make a limited policy on "English Nouns With Latinate Plurals" we are going to have to become very cunning linguists indeed, as the next round will find us conjugating verbs and making adverbs out of adjectives while the entire non-English speaking world patiently waits for valuable policy attention to be paid to some issue which might be relevant to them.

As with the repeated "what about us?" comments from users of pictographic writing whenever the subject of "single character domains" came up, we should be sensitive to how the "singular/plural issue" is defined, because we are only addressing a subset of English nouns.

Is the goal here to devise a general rule addressing "things which are similar across the board" or are we embarking on an exercise which will eventually lead to a policy compendium addressing the peculiarities of every human language, guided by a "common sense" standard? If the latter, that is going to take considerable time and resources. As for “doing the right thing” I am hard pressed to know how to build a policy or a standard around such an ill defined concept. It reminds me of obscenity being something I know when I see it.

Mari Jo

From: owner-bc-gnso at icann.org<javascript:return> [mailto:owner-bc-gnso at icann.org<javascript:return>] On Behalf Of Smith, Bill
Sent: Wednesday, August 14, 2013 7:03 PM
To: Andy Abrams
Cc: Mari Jo Keukelaar; jscottevans at yahoo.com<javascript:return>; bc - GNSO list
Subject: Re: [bc-gnso] BC comment on singular plural


I don't see this as a trademark issue. Rather it is an issue of common sense and "doing the right thing". It seems we lose on both counts.

On Aug 14, 2013, at 1:34 PM, Andy Abrams <abrams at google.com<javascript:return>> wrote:

Hi Mari Jo,

Thank you for your input - I think we're actually opposite each other in the other pending car/cars case, so it's good to get other perspectives.  You raise some very valid points, and I think the ICDR will likely agree with your legal reasoning.  But speaking only for myself, my concerns are not based on the application of trademark law at the top level.  Rather, as an Internet policy matter, I respectfully believe that allowing generic singular-plural TLDs will be confusing to consumers and bad for businesses that wish to operate websites on one of the two TLDs.  Imagine a case where a legitimate business is faced with negative PR, or worse, scams/phishing, etc. because half of their prospective customers mistakenly go to used.cars instead of used.car.  Secondly, simply as a matter of process, if is literally impossible to win a string confusion objection unless ICANN had already placed the strings in contention (thereby making the need for an objection moot), it would have been nice to have that guidance prior to everyone expending money and resources on these proceedings.  I guess we'll learn very soon whether the remaining singular-plural decisions follow suit...



On Wed, Aug 14, 2013 at 12:41 PM, Mari Jo Keukelaar <mj at johnberryhill.com<javascript:return>> wrote:

To be clear: the decisions reached in car/cars and hotel/hotels are actually more in keeping with present interpretation of trademark law than they are dissimilar since plurals of generic words are not "confusingly similar" to other generic words under traditional trademark analysis. We should avoid applying terminology from trademark law out of context, in what are contests between strings that are not themselves trademarks.

In these "generic word" disputes, the objecting parties have tended to make arguments based on trademark confusing similarity analysis. In that analysis, though, the trademark has already been determined to be distinctive. So those circumstances involve comparing something to something else which has established distinctiveness. The point is to avoid erosion of what are already distinctive marks. That reasoning just does not apply to generic terms.

This has been true for years under the Uniform Domain Name Dispute Policy. No confusing similarity was found between “Tire Discounter” and “Tire Discounters” despite the difference of only a single letter in Tire Discounters, Inc. v. TireDiscounter.com<http://tirediscounter.com/>, NAF Claim Number:

FA0604000679485 (“[b]ecause the mark is merely descriptive, small differences matter”).

Nor was confusing similarity found between the “The Suit Warehouse” mark and a “SuitWarehouse” domain name because the presence of descriptively common words necessarily restricts the “confusingly similar” analysis very closely to exact identity. The Men's Wearhouse Inc. v. Brian Wick d/b/a Defaultdata.com<http://defaultdata.com/>, NAF Claim Number: FA0208000117861.

The results have been no different under the Lanham Act. In a similarity analysis between “ENTREPRENEUR” and “ENTREPRENEUR PR” in two internet domain names, the 9th Circuit declined to find confusing similarity, noting, “[i]n the Internet context, consumers are aware that domain names for different Web sites are quite often similar, because of the need for language economy, and that very small differences matter.” Entrepreneur Media, Inc. v. Smith, 279 F. 3d 1135, 1147 (9th Cir., 2002).

Additionally, the notion that terms lacking distinctiveness are not subject to "confusing similarity" analysis is also true outside of the US, as the High Court of Australia opined in Hornsby Building Information Centre Pty Ltd v. Sydney Building Information Centre Ltd, [1978] HCA 11:

“There is a price to be paid for the advantages flowing from the possession of an eloquently descriptive trade name. Because it is descriptive it is equally applicable to any business of a like kind, it’s very descriptiveness ensures that it is not distinctive of any particular business.”

Of course if you are dealing with established distinctive marks then, yes, "Verizons" will be confusingly similar to "Verizon". But the scope of "confusing similarity" has always been delimited by the distinctiveness of the mark under analysis.

That a panel, applying ordinary principles of application of the trademark term "confusingly similar" would decline to find such similarity between two generic words is neither surprising nor unusual. If this was an unexpected result, then one might just as well ask how lawyer.com<http://lawyer.com/> and lawyers.com<http://lawyers.com/> manage to exist under different ownership (along with cars.com/car.com<http://cars.com/car.com>, house.com/houses.com<http://house.com/houses.com> dog.com/dogs.com<http://dog.com/dogs.com>, and many other generic words).

Furthermore, and perhaps even more importantly, under the ICANN rules we are not merely looking at "confusing similarity" but rather, a higher standard of "impermissible" confusing similarity. Comparing these string objections to outcomes in trademark cases is the proverbial mixing of apples and oranges because the generic strings involved are not, as a threshold matter, even trademarks. The lack of distinctiveness in generic words drives the result that they are not entitled to the broad swath of protection afforded that which would be impermissibly confusing.

ICANN has done much to protect trademark interests and we can all argue as to whether they have done enough or not.  Here, however, we are not even talking about trademarks; we are dealing with two generic terms.  The panels did not lose sight of that and I wonder why we sit here bashing them for applying the Policy we all approved and the law over which we’ve little control.

Mari Jo Keukelaar, M.A./J.D.
Name Administration, Inc.

From: owner-bc-gnso at icann.org<javascript:return> [mailto:owner-bc-gnso at icann.org<javascript:return>] On Behalf Of Andy Abrams
Sent: Tuesday, August 13, 2013 7:09 PM
To: Steve DelBianco

Cc: bc - GNSO list
Subject: Re: [bc-gnso] BC comment on singular plural

Update: the first singular-plural decisions have come in.  Both singular-plural decisions have gone against a finding of string confusion (our car/cars objection against Donuts, and a Hotel Top-Level-Domain S.a.r.l. v.Booking.com<http://booking.com/> B.V. for hotel/hotels).  In the car/cars decision, the Panel stated: "It is true that
the ICANN visual similarity standards appear quite narrow, but it is not the role [of] this Panel to substitute for ICANN’s expert technical findings."  In the hotel/hotels decision, the Panel similarly stated: "I find persuasive the degrees of similarity or dissimilarity between the strings by use of the String Similarity Assessment Tool, that ICANN did not put the applications for .HOTEL and .HOTELS in the same contention set."  In other words, the early results suggest that the ICDR may give complete deference to ICANN's earlier refusal to essentially find any instances of string confusion, no matter how close the strings.


On Thu, Apr 11, 2013 at 12:50 AM, Steve DelBianco <sdelbianco at netchoice.org<javascript:return>> wrote:
Here's what we just told the Board at the Public Forum, on behalf of the BC

ICANN’s String Similarity Panel was to place into contention sets any strings that create a possibility of user confusion.

But in late February ICANN published contention sets that did NOT include 24 pairs of singular-plural forms of the same string (English and Spanish)     Sport(s) Loan(s)    Web(s)    Game(s)  Hotel(es)

Risks of allowing both singular and plural TLDs for the same word are well understood.
-precedent for the next round
-ICANN looking pretty ridiculous

What’s not understood is how it happened and what we can do about it.

First response is to ask if the panelist follow GNSO Policy on confusingly similar.

Second response is “Chong”  ( Chinese for “Do-over” )
-Do-over on just these 24 pairs
- WIPO Mediation Rules, Article 1 says, “Words used in the singular include the plural and vice versa, as the context may require.”

Guess we could correct the Guidebook (plurals are confusingly similar)

String Confusion Objections on 7 of these pairs are in the hands of the ICDR rightnow.  If ICSR does the right thing and finds these pairs should be contention sets, The Board can apply this rule to ALL 24 pairs

Failing that, there’s Formal Reconsideration.

We all worry about threat from inter-governmental groups just waiting for ICANN to stumble.

We have enough vulnerability to stumble with so many unknowns in the new gTLD launch.

No need to add to our vulnerability with this self-inflicted wound

Andy Abrams | Trademark Counsel
Google | 1600 Amphitheatre Parkway, Mountain View, CA 94043
(650) 669-8752<https://www.google.com/voice#phones>

Andy Abrams | Trademark Counsel
Google | 1600 Amphitheatre Parkway, Mountain View, CA 94043
(650) 669-8752<https://www.google.com/voice#phones>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mm.icann.org/pipermail/bc-gnso/attachments/20130816/dce5970a/attachment.html>

More information about the Bc-gnso mailing list