resource.title.sufficient

NCEAS / metadig-checks

MetaDIG suites and checks for data and metadata improvement and guidance.

Apache License 2.0

8 stars 9 forks source link

resource.title.sufficient #85

Open gothub opened 5 years ago

gothub commented 5 years ago

Description

Check if the dataset title has an appropriate length and contains needed information to understand the dataset content and context

Priority

Choose a priority for the FAIR suite (Required or Optional)

FAIR: Required

Issues

need to agree on and develop/obtain an algorithm to determine title what, where, when

Procedure

extract the title and perform a word count - if it is too long or too short, the check fails
- Minimum: 7 words
- Maximum: 20 words
analyze the title to check for what, where, when (what was measured and on what entity, where were the resources collected or created, when do these measurements apply)

JEDamerow commented 5 years ago

analysis of title should include topic (what), location (where), dates (when). The "who"/creator would not be included in the title.

JEDamerow commented 5 years ago

The current check in ESS-DIVE and Arctic Data Center is for between 7-20 words. Should include the exact number of words to look for here to be clear. Will it be the same here?

mbjones commented 5 years ago

Agreed @JEDamerow -- I added the word count to the description, and let's discuss here what the appropriate checks are. We've had some feedback that 20 words is too low for the max. In general, the title should be conveniently citable and recognizable as a title. More extended details should go in the abstract. A hard limit is of course arbitrary. We know that short titles are generally inadequate to describe the content and context of datasets. Maybe longer titles are acceptable? Would a 400 word title be acceptable? 200? 100? 50? 30? 20?

gothub commented 5 years ago

In order to check the title for "what", "when", "where", this check could extract other fields in the metadata and check that they are in the title. For example, in EML:

for when: extract <temporalCoverage>, check if any elements are in the tile
for what: extract <abstract>, remove prepositions ("of", "at", ...), articles ("the"), ... and check for matches in title
for where: extract <geographicDescription>, cleanup and check for matches

Use similar files for the other metadata dialects.

tedhabermann commented 5 years ago

The target for this check exists in the DataCite Dialect: data[*].attributes.titles[*].title. It is currently implemented as dataset.title.present.