Logo

Michelle DeSmyter's Personal Meeting Room - Shared screen with speaker view
Sue Schuler
01:49:01
Welcome to the meeting of the DAAR Discussion Group. Please announce your name before speaking for purposes of the transcript. Please mute your microphone when not speaking to help maintain sound quality. Thanks
Maxim Alzoba
01:52:57
sharing too much will lead to situation, where ICANN compliance will use it against us
Kristine Dorrain (Amazon Registry)
01:55:12
https://docs.google.com/document/d/1KNVd0-Z9lUatSIAIswm52XnTAL3iF5W4qKIAWaXkuq8/edit
Kurt Pritz
01:57:00
I think you would need to show the arithmetic formula needed to calculate this
Kristine Dorrain (Amazon Registry)
01:58:45
Feel free to make a suggestion, Kurt!! Even right in the doc.
Kurt Pritz
02:08:21
One thought of two: The amount of harm is proportional to the total number of abused domains. (If Big Registry drops abuse 10% that is a better outcome than if Small Registry drops abuse 10%)
Kurt Pritz
02:09:55
Second thought: Abuse “performance” is not directly related to to percentage. Smaller registries should be expected to have more volatily.
Maxim Alzoba
02:11:09
it should be called somehow neutral, like coefficient this or that
Kristine Dorrain (Amazon Registry)
02:11:29
well, I think I'm only stream of consciousness... :)
Dietmar Lenden - SKY
02:15:06
I'm sorry I have to leave now - look forward to seeing this discussion going forward in the email section
Kristine Dorrain (Amazon Registry)
02:15:15
thanks Dietmar
Kurt Pritz
02:18:09
@ Jim: That is sort of where I was going - Ranking by total number of abuses or Ranking by percentage of abuses? The answer might be somewhere in between
Maxim Alzoba
02:18:36
I am against ranking
Kurt Pritz
02:18:53
Regarding the Rate of Change written in the document:“Persistence” might be a negative measurement of rate of change (as I understand the definition), So, Persistence = (# names from month 1 appearing in month 2) / (month 1 total names)
Maxim Alzoba
02:18:54
even given quite low percentage we have
James Galvin (Afilias)
02:26:16
I’m editing text but it’s not showing on the screen.
Sue Schuler
02:27:10
working in the background, so paused the share...…...sorry
Kristine Dorrain (Amazon Registry)
02:27:26
no problem... :)
Kurt Pritz
02:28:40
I am concerned about comparing with the “day before” as that will tend to establish a de facto goal (standard) that domains are taken down in a day
Kristine Dorrain (Amazon Registry)
02:30:10
agree. we don't want to set that expectation.
Kristine Dorrain (Amazon Registry)
02:30:28
We'd want to see how long a name persisted maybe?
James Galvin (Afilias)
02:31:10
that’s a valid point but I don’t know how to avoid it. frankly, most reputation providers update in near-real-time so we’re already pushing it by doing a daily calculation.
Kurt Pritz
02:34:43
There are two ways to go about this type of calculation:(1) Calculate a sort of “time-to-die” for each abused domain(2) Calculate a percentage of “persistence” percentage after an arbitrary time period. I still like the idea of setting it at a month and addressing those - but signaling that we will work with ICANN to continually reduce that time.
Maxim Alzoba
02:35:45
+1, whatever is there - ICANN will try to 'improve over time'
Kurt Pritz
02:37:17
If we set the time frame at a day would likely result in a very high percentage that will be difficult to reduce. 12 hours for a registry to receive understand the report; time to investigate; time to notify the registrar