|
Note that in the schema we say:
<element name="dc:language"> <ref name="language"/> </element> Elsewhere: <define name="language"> <data type="language"/> </define> That just looks odd to me. Shouldn't that be the standard regex? Regarding the resolution of the issue: We should just remove the full paragraph:
"The manner in which the language is represented is similar to the language tag described in [RFC3066]. It consists of a two or three letter Language Code taken from the ISO 639 standard optionally followed by a hyphen (-) and a two-letter Country Code taken from the ISO 3166 standard." The schema says that the attribute value is a "language". This is defined in chapter 17 already properly. Regarding Patrick's comment. "<data type="language"> references the W3C "language" datatype. This implies the standard syntax. Where is no need to repeat it. should delete the paragraph and rely on the external reference to schema datatype
Correction will appear in OpenDocument-v1.2-part-1-cd02-rev02.
There seems to be confusion here about the fix needed in ODF 1.2 Part 1 versus the defect report against ODF 1.0 and what is called for in the proposed Errata update.
In ODF 1.0, language is a custom datatype, not the W3C Schema datatype. Furthermore, under the custom data types list of ODF 1.0 section 16.1, it says """ - language A language is the same as a W3C schema language data type, except that the values a [sic] described by [RFC3066] rather than [RFC1766] """ It goes on to define (in schema lines 17524-17528): """ <define name="language"> <data type="token"> <param name="pattern">[A-Za-z]{1,8}(-[A-Za-z0-9]{1,8})*</param> </data> </define> """ Which basically allows all the sub-sub-sub-languages one might ever want and matches the syntax specified in RFC3066. I find it peculiar that there are far many more attributes defined to have value of type languageCode where "" - languageCode is a language code in conformance with [RFC3066], as specified in [XSL] """ yet we have this schema (lines 17534-17537): """ <define name="languageCode"> <data type="token"> <param name="pattern">[A-Za-z]{1,8}</param> </data> </define> """ in which language sub-tags are nowhere allowed. This is peculiar because fo:language has no such restriction in [XSL]. With respect to the Defect Report against ODF 1.0 and IS 26300:2006, I believe the correct response is to make the following Errata statements:
""" Replace the second paragraph of 3.1.15 with the single sentence "The language is represented using a language tag in accordance with [RFC3066]." """ This is a factually true statement and it removes any question concerning "similar to." The schema definition, with the definition of the custom "language" datatype is completely consistent with the simpler, sharper statement. In ODF 1.2 the definition of <dc:language> is as a value of language datatype and in 18.3.15 the provision includes the values of [RFC3066] and more. The result is syntactically the same as with the xml:lang attribute of [XML1.0].
There is no ODF 1.2 transposition issue for this defect and the resolution in ODF 1.0 Errata 02. |
|||||||||||||||||||||||||||||||||||||||||
 |
|
Bug
Applied
Major
ISO/IEC JTC 1/SC 34 N 1078 : DEFECT REPORT NUMBER JP2-1
Should fix in Errata 02 as well as ODF 1.2