what's the significance of saying that a (`Biosample`?) slot `is_a attribute`

[turbomam]

https://microbiomedata.github.io/nmdc-schema/attribute/ only has 16 subproperties, out of the ~ 500 Biosample slots.

1. zinc
2. manganese
3. ammonium_nitrogen
4. nitrate_nitrogen
5. nitrite_nitrogen
6. lbc_thirty
7. lbceq
8. env_package
9. gold_path_field
10. principal_investigator
11. execution_resource
12. url
13. display_order
14. git_url
15. file_size_bytes
16. md5_checksum

[turbomam]

name: attribute
description: 'A attribute of a biosample. Examples: depth, habitat, material. For
  NMDC, attributes SHOULD be mapped to terms within a MIxS template'
from_schema: https://w3id.org/nmdc/nmdc
aliases:
- field
- property
- template field
- key
- characteristic
rank: 1000
abstract: true
alias: attribute
range: string

[turbomam]

So we shouldn't be asserting that a slot is_a attribute unless

• it's domain is Biosample
  • I should logically assert that (and check if there are any violations already)
• it has a string range
  • check for violations there, too
  • @bmeluch we should not make bulk_elect_conductivity an attribute since you'rte asserting it's range as QuantityValue (which makes sense to me)

[bmeluch]

Thank you Mark! And also, how binding is "attributes SHOULD be mapped to terms within a MIxS template"?

[turbomam]

Good question. I started a thread in the NMDC Slack metadata channel and tagged @cmungall. We should copy any conclusions from there back in this issue.

[turbomam]

In the absence of any feedback, I propose that we eliminate the attribute parent slot.