Difference between revisions of "Talk:Discovery Change Proposal-2"
From Earth Science Information Partners (ESIP)
Line 17: | Line 17: | ||
# Define third attribute, with appropriate namespace. | # Define third attribute, with appropriate namespace. | ||
## As long as using namespaced attribute, it does not break in the [http://validator.w3.org/feed/ W3C Feed Validation Service]. | ## As long as using namespaced attribute, it does not break in the [http://validator.w3.org/feed/ W3C Feed Validation Service]. | ||
+ | |||
+ | == Pros and Cons of proposal #2 for DCP-2 (dereferenceable URIs) -- [[User:Rozele|Rozele]] 15:24, 9 August 2011 (MDT) == | ||
+ | |||
+ | * Positives of dereferenceable URIs | ||
+ | ** Keeps the specification simple in the sense that only two attributes are used (rel and type) | ||
+ | ** Capable of extended semantics (extensible beyond the current use cases) | ||
+ | * Negatives of dereferenceable URIs | ||
+ | ** We will need a specification (rather, vocabulary) for the content at the URI | ||
+ | ** Complicates feed readers by requiring (potentially many) dereferences of URIs | ||
+ | ** Breaks feed elements when host of URI is down |
Revision as of 15:24, August 9, 2011
Combined with incubator DCP from Matt Cechini -- Clynnes 10:22, 17 June 2011 (MDT)
We grafted the description from Matt Cechini's incubator DCP on OPeNDAP link_type with this DCP.
However, the namespace remains in opendap.org. This is chosen as a namespace because it is more authoritative than ESIP, i.e., the keepers of the DAP standard.
From telecon discussion of handling alternate service/data links in granule-level Discovery responses -- Hook 15:12, 9 August 2011 (MDT)
- See Discovery Telecon 2011-08-09 for meeting minutes on discussion.
- To continue discussion here and vote on the preferred approach.
- Possible ideas so far:
- Original DCP-2 idea: rel attribute value of http://xml.opendap.org/ns/DAP/3.3#.
- "rel" could be dereferencible URI. but is fragile.
- Could leverage the fragment part of the URI
- Use space-delimitation in "rel" URI
- Define third attribute, with appropriate namespace.
- As long as using namespaced attribute, it does not break in the W3C Feed Validation Service.
Pros and Cons of proposal #2 for DCP-2 (dereferenceable URIs) -- Rozele 15:24, 9 August 2011 (MDT)
- Positives of dereferenceable URIs
- Keeps the specification simple in the sense that only two attributes are used (rel and type)
- Capable of extended semantics (extensible beyond the current use cases)
- Negatives of dereferenceable URIs
- We will need a specification (rather, vocabulary) for the content at the URI
- Complicates feed readers by requiring (potentially many) dereferences of URIs
- Breaks feed elements when host of URI is down