DNSO Mailling lists archives


<<< Chronological Index >>>    <<< Thread Index >>>

RE: [ga] percentages and numbers of cases

> From: Harald Tveit Alvestrand [mailto:harald@alvestrand.no]
> Sent: Monday, April 09, 2001 5:42 AM
> I think we agree on the problem (no factual knowledge) and on 
> the available remedy (statistics); what follows below is mainly about

YEs, agreed.

> At 01:11 09.04.2001 -0700, Roeland Meyer wrote:
> >Also, one of the most damning arguments is that court cases are
> >puclicly available documents and, as you indicate here, UDRP 
> cases aren't.
> I think you are mistaken on both counts :-)
> court documents are not always public, due to privacy concerns.
> UDRP cases are publicly documented.

I maybe a little US-centric here, but court decisions are always public
unless specifically, with justification, sealed. That seal must itself be
defended or it will not stand. WRT UDRP, I am taking the queue from your

> http://www.icann.org/udrp/proceedings-stat.htm gives some 
> summary stats.

I thank you for the link. I wasn't aware of it.

These are of limited usefulness. There is also a time component that is
completely ignored (how many cases over what time? How long from initial
filing to decision?)  There is also the issue of quantitative data (How many
cases per complainant? How many per defendant? How many cases per panelist?)
These questions are only trivial examples. With some work, we can have
detailed questions that the statistics should be able to answer.

> What I think will not be public is WIPO's evaluation of the 
> performance of their panelists.

Management stats don't just evaluate individual job performance. They also
evaluate performance of the system. I agree that individual panelists are
internal WIPO issues, for now. But, they should become internal ICANN

> > > (BTW, wrt the "someone, somewhere" above: reports of problems
> > > that have
> > > domain names and WIPO/NADF/xxx docket numbers are quite a bit
> > > easier to
> > > discuss than "someone once reported that they had a problem".....)
> >
> >Neither of them are very interesting except where there are 
> many of them
> >with the same symptoms.
> But a claim that there are many of them with a symptom looks 
> SO much better 
> when it quotes at least a few of those that have that symptom....

What I had in mind was a repository of abtracts coupled with an intelligent
search engine. Google is one and Alta Vista is another. Most of the
remainder have a lot of "manual process" in their operations. But that is
not nearly enough. The arbitrators should also be required to enter
statistical values about their cases. Just what these values should be, is
to be determined by analysis of representitive cases.

Every case has metrics points in common with all other cases. What this
means is that the entire UDRP process comes under management control.
Panelists can now be graded and decisions can be compared for consistancy,
on an objective basis. We can also start profiling "good faith" and work on
imnproving the process. Moreover, it can be published so that the process
becomes transparent.

This message was passed to you via the ga-full@dnso.org list.
Send mail to majordomo@dnso.org to unsubscribe
("unsubscribe ga-full" in the body of the message).
Archives at http://www.dnso.org/archives.html

<<< Chronological Index >>>    <<< Thread Index >>>