Terms missing definitions #1159
Comments
|
Thanks @wdduncan! |
|
Thanks @wdduncan how did you derive this? For example in #976 I added defs to two of the above missing terms. I think if this is based on the output envo.owl file it'll be better to run it again after the next release, or to rerun this on the envo-edit.owl to get the more accurate list. I had though of using robot export to do this curious how you @wdduncan did it. |
|
I did check some of the terms against the current The terms were extracted using sparql against the current release (not the edit file). Details can be found in the query.Makefile on the nmdc-ontology repo. Once the new release is out, we can easily cross-check to see which definitions are present. |
I'll check all just in case |
|
Progress tracking... To doTerrestrial stuff
Aquatic stuffmarine stuff
Done
|
|
Follow up issues:
@wdduncan interestingly, Some terms missing defs were not spotted, e.g. |
|
In ebedb20, I've reordered landforms in a new sub-hiearchy of solid astronomical body part, which could help clean things up, but also competes with the elevated / depressed landform (formerly elevation / depression) hierarchies. Equivalence axioms sort of help, but also not that well. Possibly a topological ordering would be better, but this doesn't help with elements of topology. I think this will end up being structured by formation processes, which would make more sense on several fronts and be more useful ontologically. The other groupings would be by inference. I've also relabelled "elevation" and "depression" to prevent accidental axiomatisation with PATO qualities, fixing a few in the process. |
@pbuttigieg would it be possible to see a graphical summary of this. I would like to run it past my Australian soil-and-landscape vocabulary people. |
|
Hi @dr-shorthair sure, this is the rough sketch of the work in progress (there are of course more classes in each category, but this is the top level bit that's a struggle):
You can immediately see the issue - there's a competition between quality-based categorisation (elevated, depressed, can be many others) or location-based categorisation (surface, subsurface, submerged). There will also always be things that fall outside these categories. This is partially resolved by the equivalence axioms on things like Just like biological taxonomy, I think the processes of formation are the most "natural" way to sort this out. The rest can be populated by inference. I don't want to hijack this issue with this topic, so I'll say no more of it here (we can pick up in another [existing? #954] issue) - which I think would be a great topic for an ESIP Semantic Harmonization cluster activity (@rduerr) |
|
In ea10af9, I took the opportunity to add some more coherence to the spring hierarchy. Textual definitions are more regular and |
* addressed the a... terms of #1159 * added defs for soil terms in #1159 * addressing #1159 with some significant rearrangements to landforms * more definitions to address #1159, with some resultant relabellings and hierarchy shuffling * more defs and hierarchy refinement for #1159 * Further addressing #1159 and adding coherence to spring hierarchy * biomes defined for #1159 * some updates for 1159 Co-authored-by: Bill Duncan <[email protected]>
@cmungall @kaiiam @pbuttigieg @turbomam while l working on creating subsets of terms for NMDC, I found a number of terms missing definitions.
The text was updated successfully, but these errors were encountered: