Talk:Parent

=Proposal: Internal site= A parent is a technical aggregation of two or more data sets.

If several data sets should be combined for a certain reason, e.g. to be made citable through a single citation, those child sets are aggregated in a parent set. This is just a technical process as a prerequisite to give several data sets ONE citation and DOI. The DOI of the childs remain as a technical DOI, the DOI of the parent set may become a citation DOI as part of the official data publication.

On the web a parent set is presented with its list of childs, also providing the opportunity to download all childs in one zip-archive.
 * A parent set can not contain data sets with static URL.
 * Child sets can not belong to more than one parent set.
 * The definition of a parent is NOT required if just one data set belongs to the publication/data citation.
 * For an easier parent definition, the citation of the supplements reference should be related to each child set. When the parent is defined, in the child sets the reference will disappear under Reference(s): and instead will appear in Citation: as In Supplement to: ... (Example: )

Case 1: The supplement to a publication consists of more than one set; in this case all data sets related to the reference of the publication are grouped in a parent set. The parent set is also related to the reference and has the status supplementary data
 * Example of a parent set as Supplement including 9 childs:

Case 2: A number of data sets are defined by the PI as a citable entity (e.g. could be all meteorological observations of one expedition). The data sets are grouped into a parent, an extended abstract is added, the parent proofread by the authors. After completion the parent is published by setting the status published & citable.
 * Example of a parent set including 10 unpublished child sets:

The number of childs grouped in a parent is nearly unlimited. The following example includes 333 childs, has a total size of 8 MB and can be downloaded in one zip-archive. To formulate a query to find all childs of a parent, ask for parent:datasetID, e.g.
 * http://www.pangaea.de/search?q=parent:715004

Import
A parent set is technicaly aggregated by the curator in the following steps:

1. Open DATA SETS and choose New Parent

2. A new parent import dataset window is opened, go to References

3. Add the Child datasets (Add/Remove) into the References (PANGAEA data sets) list

4. Click on Replicate meta info from childs, some fields will be filled with entries depending on the related sets. Modify metadata as appropriate, e.g. add technical comments related to all data.

5. Go to Details and add an abstract. If the data are supplementary to a paper, the abstract of the paper should be copied; for data publications an extended abstract is mandatory.

Add an export file name in the format:
 * one author: Author_Year
 * two authors: Author1-Author2_Year
 * more than 2 authors: Author1-etal_Year

6. If the data sets are a supplement to a paper, go back to References. Set the relevant reference in References (literature & data) to supplemet to. A checkbox opens and ask, if authorship of child sets and parent set are uniqe and contains all authors of the reference. It is standard to list all authors of the paper in the data sets authorship

7. Go to Basic and add a short but significant title describing the content of the parent.

If the parent is a supplement to a paper set status published & citable.

Save the parent import


 * always check validity in web view,
 * always ask author(s) for review.

! Any official data publication of PANGAEA must be submitted for final review to the editor-in-chief !

=Proposal: External site= A parent is a technical aggregation of two or more data sets.

If several data sets should be combined for a certain reason, e.g. to be made citable through a single citation, those child sets are aggregated in a parent set. This is just a technical process as a prerequisite to give several data sets ONE citation and DOI. The DOI of the childs remain as a technical DOI, the DOI of the parent set may become a citation DOI as part of the official data publication.

On the web a parent set is presented with its list of childs, also providing the opportunity to download all childs in one zip-archive.
 * A parent set can not contain data sets with static URL.
 * Child sets can not belong to more than one parent set.
 * The definition of a parent is NOT required if just one data set belongs to the publication/data citation.
 * For an easier parent definition, the citation of the supplements reference should be related to each child set. When the parent is defined, in the child sets the reference will disappear under Reference(s): and instead will appear in Citation: as In Supplement to: ... (Example: )

Case 1: The supplement to a publication consists of more than one set; in this case all data sets related to the reference of the publication are grouped in a parent set. The parent set is also related to the reference and has the status supplementary data
 * Example of a parent set as Supplement including 9 childs:

Case 2: A number of data sets are defined by the PI as a citable entity (e.g. could be all meteorological observations of one expedition). The data sets are grouped into a parent, an extended abstract is added, the parent proofread by the authors. After completion the parent is published by setting the status published & citable.
 * Example of a parent set including 10 unpublished child sets:

The number of childs grouped in a parent is nearly unlimited. The following example includes 333 childs, has a total size of 8 MB and can be downloaded in one zip-archive. To formulate a query to find all childs of a parent, ask for parent:datasetID, e.g.
 * http://www.pangaea.de/search?q=parent:715004

=Rekursion (Uwe 2008-06-23)=

Eine Funktion f(x) ist so definiert, dass ihre Definition sie selber enthält, also z.B. f(x):=x*f(x) Um also den Wert von f(5) berechnen zu können führt man folgendes aus:f(5)=5*f(5*f(5*f(...))). Ohne Abbruchbedingung ist das natürlich sinnlos. Mit Parents und Childs ist das so: da gibt es zum Beispiel eine Rekursion beim Ausdrucken des Zitats: Da Parents und Childs im Endeffekt ja strukturell dasselbe sind (Datensätze) ruft die Funktion zum Ausdrucken der Datensatz-Citation eines Childs dessen Zitat-Ausdruckroutine rekursiv selbst auf. Zunächst produziert sie alles Child-Spezifische. Diese Funktion stellt dann fest, dass es einen Parent gibt (ID_dataSet_parent). Und was macht sie?: sie ruft einfach die exakt gleiche Funktion des Parents auf, welche das Zitat ausprintet und schreibt das hinter ein "In:". Hätte jetzt der Parent selber wieder einen Parent würde das so weitergehen. In:... In:... In:... Die Abbruchbedingung ist hier, dass irgendwann die Kette aufgrund eines fehlenden Parents unterbrochen wird.