Alternative proposal

[Roy Williams]

I agree that the emphasis must be on making simple things simple and 
correct. Once we have a community of stakeholders from the astronomy 
community, we can enrich the structure.

My experience from earlier lives is that complex software infrastructure 
often fails. During development and deployment, users find other ways to 
get what they want. Simple tools for simple jobs can be allowed to grow 
organically. How do I find out which image services can deliver cutouts? 
How do I see how galaxies look in these two surveys? Why doesn't this 
advertised service actually run?

Here is my take on metadata distribution. Yes the service or service 
provider is the original source. In real life, I have this relationship 
with several registries. There is a way for me to change my data by 
request (i.e. publishing interface), and there can also be request 
notices to fill in current information (i.e. getCapabilities). If these 
requests are ignored, then the old data is assumed correct.

Roy