The call began with a discussion about the purpose of the two groups. Chas defined them as one for the developers and the other for marketing. Mark suggested that it would be useful to evaluate the marketing group. This would be helpful in particular when Steve Olshansky will begin working with early adopters.

There was discussion about particular individuals who could be involved in MWe2eD. Russ recommended bringing in some of the people who have talked about generating diagnostic data such as the Abilene NOC. Chas asked if these calls should be reporting calls or used for deeper discussions. In that light Russ asked about how dependencies of an application might serve as a guide. He gave example how certain events need to come from directories or authentication. This led to more discussion about how to get various groups engaged in conversation.

Chas suggested we think about having a focus where an identified group or individual representing an area would be invited to a call. They would be given a list of questions and topics on diagnostics to consider before the call.

1. Diagnostic Tools that generate data. These are known tools like OWAMP and traceroute. We need to talk with people about the applications.
2. Areas or Applications (e.g., authentication, authorization) that would need diagnostics. The challenge is to get folks on the call and ask them what tools would they need down the road? These would represent Service & Application builders.
3. A larger area, new stuff with log and NetFlow consumers who would be doing the Analysis of data. These would be the folks who have been vocal in past looking for diagnostic tools.

With the focus areas in mind, Mark suggested think of a box with applications on top. To one side there would be tools that produce data. Below would be tools that analyse or correlate various different sources of data.

Chas aggreed that when you want to get information from a group you should work with the end user in mind.

Refering to the first group -- existing tools and known methods like OWAMP or traceroute. The main question would be how could these tools be enhanced. Whereas with the second group there are no existing tools (other than ad hoc scripts and such), so the question is what kind of diagnostic tools do they need.

There was more discussion confirming the relationships of the focus areas. Russ pointed out that it appears that this is a linear process where different things happen at different stages and we're helping to identify where gaps are. Mark added the fact that we're also working with different people / users and we're trying to extract data and then be able to correlate and analyze it. The challenge is to figure out how to get the different groups to work well together.

The group broke the elements into two major foci with sub groups:

1. Diagnostic applications with requirements that are met by
   1. Existing tools
      1. OWAMP
      2. Traceroute
      3. uses ad-hoc scripts, unknown tools
   2. New tools needed in specific domain
      1. Shibboleth
      2. LionShare
      3. Ad-hoc methods
2. Diagnostic data producers
   1. Log and netflow producers
   2. Analysis
      1. consumers and producers

The discussion began using wiki as shared whiteboard to map out the matrix as discussed above.

After some discussion Chas said he thought Russ's idea of getting a small group together to address particular issues of diagnostics was the corrcet direction. That would help keep the focus tight.

[AI] Chas and Mark will edit the wiki page a taxonomy diagram of what will be used to move forward. Groups to talk with and affinity topics for discussion.

Chas said he would be going to Penn State to meet with the people there working on LionShare and Shibboleth to learn more from them.

Mark worked on the diagram a bit more. He was trying to clarify and identify those who would be involved. With a bit more discussion the groups were identified as: end users, help desk, operators, diagnostic tool developers, and researchers. Then the question is who would be the consumer or the producer of the diagnostic data and tools in these different communities.

Next steps will be to talk with people who respresent the different areas identified. Conversation should be in specific domain. Which groups should be selected initially?