|
[
Permalink
| « Hide
]
Patrick Durusau added a comment - 06/Jul/09 01:52 PM
Will be found in part 3.
DISCUSSION REQUIRED: The RFC 3987 relative IRI reference is not the IRI counterpart of the relative-path URI specified in section 17.5.
This is related to the other discussions of ODF 1.0/../ODF 1.1 section 17.5. 1. The relative IRI reference in RFC 3987 has the following problematic statement in its definitions section: "An IRI reference may be absolute or relative. However, the 'IRI' that results from such a reference only includes absolute IRIs; any relative IRI references are resolved to their absolute form." (We have some work to do with regard to how absolute forms within a package are to be construed. 2. RFC 3987 is based on RFC 3986 for what a URI is, not the now-obsolete RFC 2396. There is this useful provision that needs to be aligned with how IRIs may be used to reflect fragments in ODF XML documents of the same package: "Note that in [RFC2396] URIs did not include fragment identifiers, but in [RFC3986] fragment identifiers are part of URIs." 4. The RFC 3987 relative IRI reference includes IRI forms that begin with "//" and "/" (and "@" too) as well as other characters, including ".". There is some additional subtlety in RFC 3987 around knowing the scheme. Although the relative IRI references have no explicit scheme, there is a scheme and it needs to be specified for how that (implicit) scheme works inside of a package. There is more that has to be defined about that. RECOMMENDATION Resolve this in conjunction with everything else having to do with ODF 1.0/1.1 section 17.5 and relative bits of 17.1 (e.g., manifest:full-path and its relationship to Zip package content item names and how content items are located). This means We could continue to defer for resolution in ODF 1.2, and then figure out what to do in rationalization of ODF 1.0/IS 26300/ODF 1.1. I agree to Dennis that a relative path-reference is not the same as a relative IRI reference.
The term relative path-references means a relative path-reference as defined in the last sentence of the last but one paragraph of section 4.2 of RFC3986. A proposal how to resolve this has been submitted in THIS IS NOT AN OASIS ODF 1.0 STANDARD DEFECT.
In ODF 1.0 Errata CD05 We need to punt this to later along with N1078:35 ( The current rejection comment in ODF 1.0 Errata CD04 should be replaced and N1078:35 should have an equivalent statement that this requires ODF 1.1 alignment to be remedied. |
|||||||||||||||||||||||||||||||||||||||