Suggested revisions for ENVO-SWEET sssom file

[cmungall]

This SSSOM was introduced in

• #1397

Many thanks for this!

Some issues, syntactic and content:

• syntactic (not sure why the validator didn't pick up on these)
  • curie map is incorrect syntax, should be inlined as dict not list
  • standard CURIEs e.g. ENVO:nnnnn should be used for interoperability
  • you may want to use the registered sweet prefixes to avoid slashes in the local namespace (may be invalid in some rdf serializations need to check
  • the comments have unneccessary quotes, these actually become part of the content
  • use correct version IRIs or IDs in the header
  • use semver for mapping versions
  • Comments in header should not be capitalized
• predicates
  • I think more of these should be exact. In SSSOM having an exact match does not depend on having the same text definition or on making the same BFO commitments
    • e.g. I am pretty sure that sweet and ENVO's frazil ice is the same concept
• completeness
  • these should be an indication of completeness

[rduerr]

Much larger problem recently discovered. That is that the labels for SWEET have capitals and no spaces. That needs to be fixed! I'll try to do that later this week, if someone can tell me what to do about the syntactic things mentioned above (not an ontologist, more a data manager).

As for having more exact matches, we deliberately did not assume that two terms were the same when one of them (the SWEET term) had neither a definition nor any commitments at all (SWEET practically has nothing beyond a hierarchy which in general is very different from ENVO's). Saying that something that literally has no meaning is equivalent to something that has both a definition and some commitments seems totally silly to me; especially when you find out later that the term (say for example calf) in SWEET literally meant a baby cow and in ENVO it meant a chunk of ice that fell off a bigger chunk of ice. More over, the SWEET community has decided to make it a hub for multiple definitions from various glossaries, etc. so it isn't clear that the terms in SWEET are actually concepts at this point. In any case, I presume that as SWEET and ENVO both evolve, the mapping will have to evolve too...

In the example given, SWEET was updated to use ENVO's definition; but now this is suspect given the re-purposing of SWEET that has happened. This probably should be an issue in SWEET at this point (to undo the work that had been done during the harmonization process before the SWEET direction change was made).

I don't get the bit about completeness... What does that mean?

Lastly, the example sssom file I am following doesn't have versions, so I am not sure what to do about the following:

• use correct version IRIs or IDs in the header
• use semver for mapping versions

[rduerr]

To DO:

• [x] Fix object_labels
• [x] curie map is incorrect syntax, should be inlined as dict not list - followed format in https://raw.githubusercontent.com/mapping-commons/sssom/master/examples/embedded/mp-hp-exact-0.0.1.sssom.tsv
• [x] standard CURIEs e.g. ENVO:nnnnn should be used for interoperability
• [x] you may want to use the registered sweet prefixes to avoid slashes in the local namespace (may be invalid in some rdf serializations need to check
• [x] the comments have unneccessary quotes, these actually become part of the content
• [ ] use correct version IRIs or IDs in the header
• [ ] use semver for mapping versions
• [x] Comments in header should not be capitalized

[cmungall]

I am not sure the baby cow analogy holds, in this case the intended meaning is obvious from parentage.

In the example given, SWEET was updated to use ENVO's definition; but now this is suspect given the re-purposing of SWEET that has happened.

Do you have any more context on this? I haven't heard of any re-purposing but I haven't been following the tracker closely

[rduerr]

The person to talk to here is @brandonnodnarb (I think that is Brandon Whitehead's github id) who is the current chair of the ESIP Semantic Technology Committee; though I note that his term is up later this month and it is likely that Marshall Ma will be the new chair. So who knows if that will change anything.

The committee decided that given how lightweight SWEET is on definitions and axioms, that it would make the most sense to repurpose it to sort of connect a bunch of earth science vocabularies together by associating each term in SWEET with the definitions from each of those other vocabularies (of which there are likely hundred's) , so someone using a SWEET term could see all the other terms (including IRI's) and their definitions. The hope was that this would spur folks to take a look at all these various definitions and clean things up!!! Has that happened - well, some of it but not a lot... So I don't know where this is going to go... Without funding, probably nowhere.