gbuehler
commented
3 years ago In ISO 19115 there are 4/many main areas where the ordering differs from the printed standard In ISO 19115-1 MD_Metadata there are two XML elements that differ: resourceScope (developed to replace hierarchy and heirarchyLevel) DataDictionary has it at the end whilst the XSD has it much earlier resourceLineage DataDictionary has it second from the end whilst the XSD has it a few elements earlier EX_VerticalExtent there are two XML elements that differ: verticalCRS DataDictionary has it second from the end whilst the XSD has it at the end verticalCRSId DataDictionary has it at the end whilst the XSD has it second from the end SV_CoupledResource resource DataDictionary has it second from the end whilst the XSD has it at the end operation DataDictionary has it at the end whilst the XSD has it second from the end In ISO19115-2 all four of the XSD files have miss-ordered elements, some worse than others
As for implementation issues, there are two concepts to consider. Easy now and hard later or the other way around. At each new edition of the standard implementors make a decision as to whether or not to move to the new material. The older versions of the XML schemas are (now) retained for users to continue to access if they have older material that can/should/will not be transformed into the newer version. Moving from ISO 19115 to ISO 19115-1 there were XSLt transformers generated that allowed the older XMLs that complied with the implementation developed in ISO 19139 to bridge the gap to ISO 19115-3. An update was undertaken to progress XSDs for ISO 19115-3 Revision1 to Revision2, but no transform was developed. The change from the current ISO 19115-3 to the proposed new ISO 19115-3 implementation (all XSDs in a 19115-3 directory -> material labeled with the origination standards) can take two paths. 1) minimal change - only change namespace URIs and schemaLocation Migrating older XML to new revision could be undertaken by relatively simple code (PERL, Python, C, etc) 2) realignment elements to data dictionary - change as above plus element order changes URIs and schemaLocation changes in XML as above PLUS a few additional scripts (XSLt would be the least efficient). The extra code could be developed in parallel to the URI and schemaLocation scripts. NOTE: XSLt solutions only work if the XML can be considered as files. Many systems store the attribute values in databases, and generate the XML only when exporting the material. In these cases the import is usually done using a order agnostic XSLt and the extract script should be readily adjustable. It is possible that this is why an XSLt for the Rev1 to Rev2 for XMLs was never published.
I believe that the decision is the responsibility of the ISO 19115-3 review project, but without multi-lateral support it would be disastrous.
Yours faithfully Evert Evert Bleys 4 Tudor Place HUGHES ACT 2605 Australia email: ejbleys@gmail.com Mob: +61 (0)411 483 876 Land: +61 (0)2 6281 1773 Skype: ejbleijs@gmail.com
Justification: It has been brought forward by XMG that the xsd files as they associate to part 3, all four of the XSD files have miss-ordered elements, some worse than others Proposal: Discussion must occur or a decision made to reorder the xsds to fit the data dictionary, both for this standard and part 1