I have been thinking about how this group is progressing - partly because I am retiring and the Brussels meeting will be my last. It seems to me that the group is slogging it out, and making progress. but slowly I do think the framework doc is pretty good. It probably should be given to a technical writer to clean it up and ask logic questions, but it is basically sound. The connection doc is well started, but in my mind still needs a fair amount of editing and hashing out of technical concepts. It can work out, and perhaps if I am gone it will go faster. I wouldn't be surprised :) I do have a suggestion that the group might think about at Brussels. The Fenius protocol software that was demo'd at GLIF does most of what the NSI protocol will need to do. It does not include an NSA "federation" agent. The limited capability it does have as an agent is in the "superagent" which call other Fenius servers. I believe that the interface is close to what is provided in Argia, DRAC, Harmony and OSCARS, though perhaps with only a subset of some interfaces. I suggest that this group take Fenius, OSCARS, Argia, and DRAC implementations of the basic commands that Fenius supports, compare them, create a "standard version 1) by taking the best of each and declaring it a standard. I think the commands in Fenius are Reserve, query, list, release -- these might be all that is supported in version 1. If this gets implemented by multiple implementers the group could then upgrade the standard based on implementation and operation history. It is likely this would get implemented by GLIF GOLE pilot to get some operational experience. Something to think about.. John