The fault is not in our databases, but in ourselves: messy data, metadata, and interoperability

The fault is not in our The fault is not in our databases, but in ourselves: databases, but in ourselves: messy data, metadata, and messy data, metadata, and

interoperabilityinteroperability

Adam RabinowitzAdam Rabinowitz

The University of Texas at The University of Texas at AustinAustin

arabinow@mail.utexas.eduarabinow@mail.utexas.edu

British Museum Collection CRM schema - work in progress version by Dominic Oldman is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.

Rabinowitz ● FAIMS 2012 ● 8.17.12

Rabinowitz ● FAIMS 2012 ● 8.17.12

Rabinowitz ● FAIMS 2012 ● 8.17.12

Go Go Fish!Fish!

Rabinowitz ● FAIMS 2012 ● 8.17.12

Does anyone have… a medieval Islamic Does anyone have… a medieval Islamic padlock?padlock?

Rabinowitz ● FAIMS 2012 ● 8.17.12

Rabinowitz ● FAIMS 2012 ● 8.17.12

Rabinowitz ● FAIMS 2012 ● 8.17.12

Rabinowitz ● FAIMS 2012 ● 8.17.12

Hey, Corinth: do you have… a ceramic krater?Hey, Corinth: do you have… a ceramic krater?

Rabinowitz ● FAIMS 2012 ● 8.17.12

Hey, Corinth: do you have… a ceramic krater?Hey, Corinth: do you have… a ceramic krater?

Rabinowitz ● FAIMS 2012 ● 8.17.12

Hey, Corinth: do you have… a ceramic krater?Hey, Corinth: do you have… a ceramic krater?

Rabinowitz ● FAIMS 2012 ● 8.17.12

Does anyone have… stuff from Athens?Does anyone have… stuff from Athens?

Pelagios graph explorerPelagios graph explorer

Rabinowitz ● FAIMS 2012 ● 8.17.12

Hey, South Italy, do you have… Geometric ceramics?Hey, South Italy, do you have… Geometric ceramics?

Rabinowitz ● FAIMS 2012 ● 8.17.12

Mockup of metadata description for digital object Mockup of metadata description for digital object sfi_CH05SR_3065_m.JPGsfi_CH05SR_3065_m.JPG

<dc > <title>sfi_CH05SR_3065_A9_m.JPG</title> <temporal>Hellenistic to Late Byzatine</temporal> <spatial>Chersonesos South</spatial> <creator>ICA</creator> <description>Lock</description> <date>2005:08:23</date> <format>.JPG</format> <identifier>sfi_CH05SR_3065_A9_m.JPG</identifier> <publisher>ICA</publisher> <language></language> <relation>image_documentation/special_finds/photographs/master</relation> <isInContext>CH05SR_434</isInContext> <isPartOf>sfi_CH05SR_3065</isPartOf> <rights>Specified in SIP</rights> <subject>Chersonesos South</subject> <subject>Archaeology</subject> <subject>photographs</subject> <source>Master</source><type>image_documentation</type></dc>

•StaticStatic

•Harvested From ARKHarvested From ARK

•Technical MetadataTechnical Metadata

•Harvested from Hierarchy Harvested from Hierarchy

•Harvested From File NameHarvested From File Name

Rabinowitz ● FAIMS 2012 ● 8.17.12

iRODS (integrated rule-oriented data system)iRODS (integrated rule-oriented data system)www.irods.orgwww.irods.org

Rabinowitz ● FAIMS 2012 ● 8.17.12

Descriptive metadata using METS and Dublin CoreDescriptive metadata using METS and Dublin Core

Rabinowitz ● FAIMS 2012 ● 8.17.12

Automated integration of post-season documentation with Automated integration of post-season documentation with database using unique ID in filename of iRODS objectdatabase using unique ID in filename of iRODS object

Rabinowitz ● FAIMS 2012 ● 8.17.12

Preservation metadata using METS and PREMISPreservation metadata using METS and PREMIS

Rabinowitz ● FAIMS 2012 ● 8.17.12

The next step: a freestanding desktop toolkit to The next step: a freestanding desktop toolkit to automatically extract metadata without iRODS?automatically extract metadata without iRODS?

Rabinowitz ● FAIMS 2012 ● 8.17.12

Go Fish, again! Some thoughts about interoperability and Go Fish, again! Some thoughts about interoperability and archaeologists off the gridarchaeologists off the grid

• The bar for metadata creation has to be lowered – it needs to be cheaper and easier to document your own material outside repositories

• This is especially true for documentation outside the database: individual files like images (or spatial data) that tend to proliferate even after a project is over, and that aren’t self-explanatory

• Data federation strategies should take into account the tendency of archaeologists to begin a search for information with a few key concepts, like

• Space• Time• Record type (site, building, object, excavation photo, etc)

• and within this, for objects, material

• URI services could really help here in the production of minimally interoperable metadata; Dublin Core probably isn’t a bad starting point as a schema

• In building services to use interoperable data, we should not neglect the highly visual nature of Go Fish searches, which get us past issues of terminology, classification, and misclassification (I can tell that’s not a hand-axe when I see it)