Re: [nm-wg] immediate future of current schemas

[Warren Matthews <warren.matthews@oit.gatech.edu>]

Tanya makes a very good point. It's a bit chicken-and-egg. I think more
implementations will allow us to understand what is missing. But that
drives change, and it seems that is delaying implementation.

How much of Mark's to-do list can be swept under the carpet for the
"new" schema? Can we get a date when the "old" schema will be frozen?

Dan's notes from last weeks call indicate a new schema draft will be
available soon. Perhaps we should delay this or at least recommend the
frozen old schema is used for now. Then we can get on with some
development and provide feedback.

Sorry I don't get to attend meetings. I'm sure this stuff has been
discussed. 

I'm willing to help if anyone out there wants to work on a client to
make requests (for data and tests) from the PIPES and/or AMP services. I
even have some suggestions if someone has a student with time on their
hands.



On Wed, 2004-10-20 at 16:05, Tanya Brethour wrote:
> > I don't have such a problem with change. I think we need to iron out
> > what is the best solution and not be tied to any bad ideas for the sake
> > of backwards compatibility.
> 
> The best solution does need to be ironed out, but isn't one of the main
> points of the new schema that we have learned from our mistakes with the
> current schema? Or are we still going to be exploring new ground with this
> new schema?
> 
> At some point we need to decide that the new schema is stable. Up until
> then... you can ignore backwards compatibility. However, I suspect that
> many of us won't be jumping onboard with the new schema until that
> happens.
> 
> My main issue is that there are many more interesting aspects of our
> project then how to represent metrics. We do want to and find it necessary
> to collaborate with other networking measurement infrastructures, but if
> we spend all of our time reimplementing the schema in our backend.. then
> it really prevents us from making any progress at all. 
> 
> I'd really prefer not to spend another 1-2 years with each infrastructure
> having a different version of the schema implemented and therefore making
> it impossible to interact. Its just gotten to the point where things are
> stable enough to rely on successfully interpreting the response from
> another measurement database.
> 
> -Tanya
>