| | 1 | = Proposal number : Components and Composites = |
| | 2 | |
| | 3 | || '''Date''' || 2008/08/05 || |
| | 4 | || '''Contact(s)''' || Ted Habermann || |
| | 5 | || '''Last edited''' || [[Timestamp]] || |
| | 6 | || '''Status''' || draft || |
| | 7 | || '''Assigned to release''' || to be determined || |
| | 8 | || '''Resources''' || Some may be available || |
| | 9 | |
| | 10 | == Overview == |
| | 11 | Large metadata repositories include significant amounts of information that is repeated in many metadata records. This includes on-line resources, contact information, citation information, instrument information, platform information, ... Our current system at NGDC allows us to store this sort of information once in a separate group of records, to edit this information once, and to include it as a “component” in a larger metadata record (a “composite”). Our repository currently hold thousands of components, some of which are included in hundreds of composites, so this capability is critical for us. |
| | 12 | |
| | 13 | Our overall concept is that metadata “records” be viewed as collections of components that are held together by xlinks rather than as a single blob of XML in a database or a file. These xlinks would point to RESTful web services that would provide the component content in response to a URL. We hope that the MD_Identifiers that are included in numerous locations in the ISO 19115 and 19115-2 standards may provide a mechanism for linking these, although we would need to add new MD_Identifiers to these standards to make this work. For example, neither CI_OnlineResource or CI_ResponsibleParty packages currently include identifiers and both are clear candidates for componentization. |
| | 14 | |
| | 15 | |
| | 16 | === Proposal Type === |
| | 17 | * '''Type''': GUI Change, Core Change, Module Change |
| | 18 | * '''App''': !GeoNetwork |
| | 19 | * '''Module''': Not sure about this one yet |
| | 20 | |
| | 21 | === Links === |
| | 22 | * '''Documents''': |
| | 23 | * '''Email discussions''': |
| | 24 | * '''Other wiki discussions''': |
| | 25 | |
| | 26 | === Voting History === |
| | 27 | |
| | 28 | == Motivations == |
| | 29 | See Overview |
| | 30 | |
| | 31 | == Proposal == |
| | 32 | We propose to test this idea on a simple case that involves a CI_OnlineResource and a CI_ResponsibleParty, one of the simplest Component/Composite relationships in 19115. We will |
| | 33 | |
| | 34 | * straightforward stuff |
| | 35 | * create a !GeoNetwork Template that includes just an On_lineResource |
| | 36 | * add a MD_Identifier to that template |
| | 37 | * populate !GeoNetwork database with some On_lineResources with identifiers |
| | 38 | * create a !GeoNetwork Template that includes a ResponsibleParty and CI_Contact |
| | 39 | * need some ideas stuff |
| | 40 | * figure out how to reference the On_lineResources using an xlink from the CI_Contact |
| | 41 | * figure out how to resolve the xlink so that we can index the On_lineResources. |
| | 42 | * search for a contact based on content of the On_lineResource |
| | 43 | |
| | 44 | === Backwards Compatibility Issues === |
| | 45 | This solution, as proposed, will only work in the context of ISO metadata standards |
| | 46 | |
| | 47 | == Risks == |
| | 48 | This solution, if it works, may suggest useful extensions to the ISO 19115 standard. That standard is up for review during 2008, so this might be a good time for such suggestions. |
| | 49 | |
| | 50 | == Participants == |
| | 51 | * Ted Habermann and his group at NOAA's National Geophysical Data Center |
