revise the interpretation of production dates which have only an "Earliest Date"

[Conal-Tuohy]

There is an issue with the production date fields as stored in Solr; they are assumed (sometimes incorrectly) to be date ranges which, if the end date is not specified, is understdood to mean a period which remains ongoing.

For instance, this object http://data.nma.gov.au/object/139881#
(a postcard) has a Creation Activity that begins on 1937-02-06 and is still in progress today, i.e. the stored temporal_date value is [1937-02-06 TO *] which is wrong.

In some cases we have long-lasting events, but my feeling is that for production events, we should assume that unless the EMu record includes both an earliest and a latest date, that any "earliest date" given should be interpreted as the single point in time of the creation (i.e. that the creation is treated as instantaneous, rather than taking place within some period of time).

Thoughts?

The EMu source data is here:

root@nma-dev:/mnt/emu_data/full# xmllint --xpath "/*/*[irn='139881']" 2018-12-05_19-30_objects_86311_FULL.xml

<record>
  <irn>139881</irn>
  <TitObjectTitle>At the Entrance to St. Kilda Road, Melbourne</TitObjectTitle>
  <TitObjectType>Object</TitObjectType>
  <TitObjectNumber>1986.0117.6741</TitObjectNumber>
  <TitObjectName>Postcards</TitObjectName>
  <PhyDescription>A postcard showing a black and white photograph of a street scene with vehicles, with a message written on the back. Text on front bottom left hand corner reads "No. 23. At the Entrance to St. Kilda Road, Melbourne". The postcard is addressed to "Miss Ste[---?] / Havenden / Hammond Avenue / Croydon / N.S.W."</PhyDescription>
  <AccAccessionLotRef>1849</AccAccessionLotRef>
  <WebReleaseDate>28/03/2008</WebReleaseDate>
  <InwardLoan>no</InwardLoan>
  <PhyRegistrationHeight>88</PhyRegistrationHeight>
  <PhyRegistrationWidth>138</PhyRegistrationWidth>
  <PhyRegistrationUnitLength>mm</PhyRegistrationUnitLength>
  <PhyMaterials_tab>
   <PhyMaterial>Paper</PhyMaterial>
   <PhyMaterial>Ink</PhyMaterial>
  </PhyMaterials_tab>
  <AcsCCStatus>Public Domain</AcsCCStatus>
  <AcsAPI_tab>
   <AcsAPI>Public</AcsAPI>
  </AcsAPI_tab>
  <AdmPubishWebNoPassword>Yes</AdmPubishWebNoPassword>
  <AdmDateModified>26/03/2018</AdmDateModified>
  <ProductionPlaces>
   <ProductionPlace>
    <ProPlaceRef_tab.irn>45</ProPlaceRef_tab.irn>
    <ProPlaceType_tab>Place made</ProPlaceType_tab>
   </ProductionPlace>
  </ProductionPlaces>
  <AssociatedPlaces>
   <AssociatedPlace>
    <AssPlaceRef_tab.irn>80140</AssPlaceRef_tab.irn>
    <AssPlaceType_tab>Associated place</AssPlaceType_tab>
    <AssPlaceNotes_tab>addressed to Hammond Avenue, Croydon</AssPlaceNotes_tab>
   </AssociatedPlace>
   <AssociatedPlace>
    <AssPlaceRef_tab.irn>45</AssPlaceRef_tab.irn>
    <AssPlaceNotes_tab>photograph taken of St Kilda Road</AssPlaceNotes_tab>
   </AssociatedPlace>
  </AssociatedPlaces>
  <ProductionDates>
   <ProductionDate>
    <ProDate0>06/02/1937</ProDate0>
    <ProEarliestDate0>06/02/1937</ProEarliestDate0>
    <ProDateType_tab>Date of correspondence</ProDateType_tab>
   </ProductionDate>
  </ProductionDates>
  <OnDisplay>no</OnDisplay>
  <HasEmuImage>yes</HasEmuImage>
  <WebMultiMediaRef_tab>
   <image>
    <media_irn>56127</media_irn>
    <media_mimeformat>jpeg</media_mimeformat>
    <media_mimetype>image</media_mimetype>
    <res640px>
     <image_path>56/127/rg_56127.640x640.jpg</image_path>
     <image_width>640</image_width>
     <image_height>420</image_height>
    </res640px>
    <res200px>
     <image_path>56/127/rg_56127.200x200.jpg</image_path>
     <image_width>200</image_width>
     <image_height>131</image_height>
    </res200px>
   </image>
  </WebMultiMediaRef_tab>

[Conal-Tuohy]

A related problem is object 112121 which is a string bag which it claims is still in use, though clearly it is not.
https://data.nma.gov.au/object/112121#

[f27wood]

This looks like a data issue and ideally we would review all of the dates with this issue. I have managed to put together a list of records with production dates that need reviewing - there are around 100 of them.

In the meantime, yes I think the assumption about the production date is what should be used.

It also raises the issue around the other dates (associated dates), which I think you should leave as they are for now and we will also review these.

[f27wood]

Hi Conal, speaking to Karen, we are going to review the problem dates - so hold off on doing anything until we have done that.

[Conal-Tuohy]

Following discussion with @f27wood, agreed to modify the interpretation of such dates for the purposes of searching only (i.e. the Solr search fields would be points in time, rather than date ranges stretching off into the indefinite future)