multiple URI elements per Service in XRDs

Johannes Ernst jernst+lists.danga.com at netmesh.us
Fri Jan 20 17:19:31 UTC 2006


This discussion seems to say that we need to think about this from  
two angles:
1) What does the YADIS spec need to define so it is clear what is and  
isn't a conforming implementation
2) What are implementors *required* to do, and what is only good  
implementation practice.

It appears to me, that there is not requirement (aka MUST):
  - that an implementation faithfully follow the priority attributes  
at all
  - that an implementation try out more than one of the URIs provided  
in a Service element
It is only considered good implementation practice.

Is that true?

It would make sense to me, because of many reasons such as the  
following:
  - we can't / don't want to require implementors to support https as  
well as http
  - we can't / don't want to require implementors to support all  
possible service types -- which means that by definition, they get to  
pick which ones the consider, never mind the priorities
  - we don't want to prevent implementors to code local rules such as  
"I prefer talking to foo.example.net instead of bar.example.com  
because I know that bar.example.com has just been taken over by the  
transsylvanian mafia and this user might not know that yet"...

If this is right, we have some options, such as
  - be clear in the spec about the entire algorithm as intended
  - touch on a random subset of the entire algorithm (which is the  
current state, I guess)
  - drop all mentioning about processing algorithms from the YADIS  
spec, and just leave a note in its place saying that "best practices"  
are defined in the upcoming OASIS documents.

Thoughts?


Johannes Ernst
NetMesh Inc.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: lid.gif
Type: image/gif
Size: 973 bytes
Desc: not available
Url : http://lists.danga.com/pipermail/yadis/attachments/20060120/d88f526d/lid.gif
-------------- next part --------------
  http://netmesh.info/jernst






More information about the yadis mailing list