Re: [soac-mapo] charter and mission

[Richard Tindal <richardtindal@xxxxxx>]

Hi Milton,

Thoughtful points.  I agree with you that defining a standard is the hardest 
part.   

I think removing MAPO entirely is a legitimate option for us to evaluate.    
This would presumably need GNSO endorsement as their original Implementation 
Guideline would need to be changed.  

To put my cards on the table --- I'm not wed to any form of solution.      As a 
TLD applicant whose string won't be affected by this rule (a boring and 
uncontroversial string) my motivation is to get agreement on a reasonable rule 
so the TLD program can move forward.    My definition of reasonable is whatever 
the large majority of people can live with.  

Let me poke at one matter of detail in your note though.   When you mention the 
illegality of censorship,  isn't that only in the context of Government action? 
  Are there rules about private US corporations and censorship?  

Richard 


On Jul 12, 2010, at 1:16 PM, Milton L Mueller wrote:

> I find all this talk of groups of a group vetting specific applications to be 
> weirdly disconnected from reality.
>  
> Lets be honest and direct: if you propose something like that you are 
> advocating a form of arbitrary censorship, which when applied is likely to be 
> illegal in many jurisdictions – just as the absence of such censorship would 
> be illegal or objectionable in others. Let’s stop pretending that this 
> problem doesn’t exist, OK? No organizational arrangement or method of 
> selecting people can get around the fact that standards differ across 
> countries.
>  
> So who the group is a secondary matter. What matters first and foremost are 
> the criteria used to censor or block names. Once you have the criteria, who 
> applies it or how it is applied is a matter of implementation. An important 
> detail, but a secondary matter.
>  
> So what should the criteria be?
>  
> The original DAG list is quite good in some respects, in that it is narrowly 
> drafted. The problem, however, is that it is virtually impossible for a TLD 
> string to “incite” anything, based on what the term “incitement” actually 
> means when used as a legal standard.* You would have to show that the mere 
> creation or implementation of a TLD string (not the content associated with 
> it) was responsible for generating violent, lawless action, creating racial 
> discrimination or sexual abuse of children. This would be very difficult, if 
> not impossible.
>  
> Another, even more serious problem is that since incitement is already 
> illegal, it is obvious that you don’t need TLD restrictions to stop the kind 
> of conduct or problems that are targeted. In other words, if someone applies 
> for a TLD with <your favorite racial epithet here> and starts using it to 
> incite, it would be illegal as can be under existing law in virtually every 
> country I can think of.
>  
> So others are suggesting even broader standards – the “something I find 
> objectionable” standard Tindal mentioned. This is clearly unacceptable. We 
> can – and must - rule out any such “standard” immediately. To begin with, 
> such a standard is literally illegal in about 2/3 of the Internet world 
> (contravenes both the U.S. Constitution and the European Convention on Civil 
> and Political rights). It may be inconsistent with other countries national 
> constitutions as well. Therefore, there is literally no group you can select, 
> or any method of selecting a group, that can implement such a standard in a 
> lawful way.
>  
> * Some of you may have been present at the GNSO discussions in Brussels when 
> a colleague of mine asserted that it was impossible for a tld string to 
> incite and someone disagreed, saying “.kill the jews” was an example of such 
> a string. Many people at the time thought this was a very clever comeback, 
> and seemed to think the point had been refuted. But actually the opposite 
> point was proven. The person who uttered those words was not inciting 
> violence against Jews. He could never be prosecuted for uttering those words. 
> He managed to say “kill the Jews” without anything bad happening and without 
> anyone. Q.e.d. We simply don’t need ICANN rules to protect ourselves against 
> that kind of incitement. We are already protected against it by more more 
> carefully drafted, narrowly applicable laws.
>  
>  
>  
> From: owner-soac-mapo@xxxxxxxxx [mailto:owner-soac-mapo@xxxxxxxxx] On Behalf 
> Of Richard Tindal
> Sent: Monday, July 12, 2010 12:25 PM
> To: soac-mapo@xxxxxxxxx
> Subject: Re: [soac-mapo] charter and mission
>  
> I agree with Avri, however I do think the expansiveness of the definition 
> (the standard applied) is related to who is making the decision.      
>  
> The tighter the standard the more comfortable I am with a smaller group 
> making the decision.     The broader the standard the more I want to see a 
> large group making it.  
>  
> These are really the 2 issues every approach must determine   (actually,  
> every approach except the one proposed by Milton -- which is remove MOPO 
> entirely).
>  
> 1.     What standard will apply?;
>  
> 2.     Who will make the decision?
>  
> It seems to me 2. is a little easier to decide that 1.      For 2. we could 
> have the IO,  a Panel,  a suite of judges (per the DAG) or we could even have 
> ICANN SO/ ACs make the decision.   There are pros and cons to each.
>  
> For me 1. is harder.   The standard could be very broad (e.g.  'something I 
> find objectionable' )  to very specific (e.g.  a list),    or anything in 
> between.  
>  
> By way of background,   the current DAG has a attempted a sort of hybrid 
> standard of specificity.     It has three, fairly specific standards:  (1)  
> incitement to violent lawless action  (2) Incitement to or promotion of 
> discrimination (race, color, gender, ethnicity, religion, national origin) ; 
> and (3) Incitement for child pornography or other sexual abuse of children.   
>    I haven't read any deep concerns about these standards.
>  
> The fourth standard in the DAG is more expansive though - (4)  contrary to 
> norms recognized under principles of international law   (note:  my bold 
> type).     This is the DAG standard that seems to be most objectionable to 
> the GAC and some others.
>  
> RT
>  
>  
>  
>  
> On Jul 12, 2010, at 6:51 AM, Avri Doria wrote:
> 
> 
> 
> Hi,
> 
> an intersting idea.
> 
> since an IO already exists, how about an IO with an advisory panel?  This 
> would have difference sorts of specialists, and depending on the kind of 
> string.  Again like a RSTEP.
> 
> I have a problem with the expansiveness of our definitions however. and the 
> fuzzing of the boundary between problems with a string 
> (.burn-a-vampire-today) that is an incitement in itself, and a string 
> (vampirelove) that might contain unsavory and unacceptable content.
> 
> a.
> 
> 
> 
> 
> 
> 
> On 12 Jul 2010, at 08:56, Antony Van Couvering wrote:
> 
> 
> Thanks Jon you did wonderfully :-)
>  
> An IO is one person, hence a lightning rod for conspiracy theories.  Also, 
> one person, even of the utmost probity, has biases a broad-based group would 
> not (though a group might have other biases).  That's why a panel might work 
> better.
>  
> On the other hand, the IO position already exists.  
>  
> Sent from my handheld.   
>  
> On Jul 12, 2010, at 7:55, Jon Nevett <jon@xxxxxxxxxx> wrote:
>  
>  
> I can't speak for Antony, but I think that the approach he was taking issue 
> with was the one Evan mentions directly below and not the IO process.  If 
> not, I will.  I think that the IO objection process would be easier to 
> implement than a banned list or a pre-application advisory process.  First, a 
> list would require a whole lot of needless debate about names that no one 
> would have applied for during the process.  Second, a pre-application 
> advisory would not be able to take into account the applicant, the string, 
> and the intended use.  If it did, it would give these applicants an unfair 
> advantage over other applicants that might be in a grey area on other issues 
> (e.g. trademarks on the top level, string similarity, etc.).  Finally, we 
> shouldn't be too worried about applicants (and their investors) who apply for 
> a name that they know will be highly controversial.  They obviously will know 
> that going in and there already is a partial refund available in DAG4 if they 
> see that!
>  their application got caught up in an objection process and they choose not 
> to proceed. 
> 
>  
> Thanks.
>  
> Jon
>  
>  
> other option raised during the GAC/At-Large meeting was inspired by the 
> trademark clearinghouse. There could be an advisory process through which TLD 
> applicants would know -- in advance of approval -- whether their string was 
> likely to be blocked by countries once implemented. Based on that advice, a 
> TLD applicant could withdraw or continue, knowing ahead of time that their 
> string could cause problems being seen in some jurisdictions. An advisory 
> process rather than an objection one preserves free expression, while 
> ensuring that applicants (and their investors) are aware of national 
> obstacles that may lie ahead.
>  
>  
>  
> On Jul 12, 2010, at 3:49 AM, Evan Leibovitch wrote:
>  
>  
>  
> On 12 July 2010 03:28, Antony Van Couvering <avc@xxxxxxxxxxxxxxxxxxxx> wrote:
> The issue with this approach is that the string itself may not be the issue 
> -- I would contend that in most cases it would be the combination of the 
> string and the applicant.  There is nothing wrong with the string "boy," for 
> instance.  But there's a big difference between .boy operated by the Boy 
> Scouts and .boy operated by NAMBLA.
>  
> On the contrary, that's the strength of the At-Large proposed approach. By 
> putting such issues in the hands of the Independent Objector and offering 
> sufficient leeway, context can matter as much as the literal string itself. 
> Unanticipated problem applications can be objected to "on behalf of the 
> public interest" by the IO rather than depending upon some external body 
> (who? The Boy Scouts? A government? The Catholic Church?) to object to a 
> NAMBLA-run registry ".boy".
>  
> (Of course none of this prevents NAMBLA from purchasing second-level domains 
> under .boy, but that's a different issue :-P)
>  
> Under a similar scenario mentioned in one of the meetings, even everyone's 
> favourite example of  ".nazi" might be acceptable if it was proposed purely 
> for academic study. But this, like the NAMBLA one, is a rhetorical device 
> rather than a real-world proposal that will have to be confronted. That's the 
> nice thing about having an IO process that's not tightly restricted ... that 
> when real-world problem applications come up, we have a process suitably able 
> to deal with it.
>  
> (In more recent versions of the DAG, certain undesirable restrictions were 
> put on the IO that would inhibit the role's effectiveness in performing such 
> a duty. However we can recommend changes that remove such limitations.)
>  
> - Evan
>  
>  
>  
>  
> 
>