[standards-jig] LAST CALL: Service Discovery (JEP-0030)

Adam Theo theo at theoretic.com
Sat Dec 7 00:16:59 UTC 2002


After reading the disco spec now, and twisting my brain around the 
problems being argued about here, I'm thinking perhaps we should pause 
disco's advancement into the Council. At the very least it should 
undergo alot more public discussion. There appear to be alot of people 
who are displeased, confused, or apprehensive (which is where I'm at) 
about Disco, and I believe the best way to get Disco passed and accepted 
by the community and developers is to ease people's fears and spend some 
effort explaining why Disco is good. Besides, perhaps in the process 
Disco could improve as long as the authors keep open minds about ideas 
from new developers.

I don't know about pausing it for addressing to be better worked out, 
but I do think iot should be paused for pub/sub to be worked out. I see 
alot of the requirements that are layed out in Disco can be solved by 
pub/sub, freeing Disco to be alot more flexible and useful than it is 
now. I specifically mean the restriction that Disco cannot contain any 
details about children, thereby forcing the requesting client to send 
dozens if not hundreds of more IQ requests to each child returned in the 
parent's list in order to find what they want. This would turn into a 
disaster which earlier drafts of Disco avoided by allowing that info to 
be held by the parent and sent back in the main response. Pub/sub can 
solve this connundrum, allowing Disco to act more like earlier drafts in 
that respect.

Joe Hildebrand wrote:
> I guess the question is, do we need to pause disco development while we
> solve the addressing issue?




More information about the Standards mailing list