ENVO cryosphere wrongly set as a subclass of glacial feature, etc.

[wiegandn]

@rduerr @pbuttigieg @kaiiam

1. When searching ENVO for cryosphere (and also seen searching for ice calf (and perhaps elsewhere too)), cryosphere is listed as a subclass of glacial feature, which is incorrect.
2. Under ice calf, fixing the above may also fix ice mass being both a subclass and sibling of cryosphere, which also would need to be fixed.
3. Further, under ice calf, ice mass should probably be a sibling of glacial feature and not a subclass. (maybe a separate issue)

[cmungall]

Hi @wiegandn - I agree this is a problem

I tracked this down to an incorrect logical definition. Here is the reasoner explanation (if this doesn't mean anything to you don't worry, the editors will take care of this):

This is due to an overly broad logical definition of glacial feature, as being equivalent to:

material entity and composed primarily of some (snow or water ice)

Note that under this definition all of the following things are glacial features:

• a popsicle
• a cryosphere
• a snowball
• a snowman
• an ice mass

[kaiiam]

Thank you very much @wiegandn for the comments and @cmungall for checking this.

The version currently published doesn't have all the changes we made/want to make, they are in my Pull Request (PR) #1172. I'll address the issues noted above one by one with a reasoned version of the latest Pull Request.

1. cryosphere is listed as a subclass of glacial feature, which is incorrect.

In the PR we obsoleted the old cryosphere and replaced it with new terms for cryosphere and planetary cryosphere (with a different equivalence axiom). Such changes (for now) fixed that problem of cryosphere being subclass to glacial feature

2. ice mass being both a subclass and sibling of cryosphere

This is fixed as well.

3. ice mass should probably be a sibling of glacial feature and not a subclass. (maybe a separate issue)

Given the current definitions the way we have it now (ice mass subclass to glacial feature) makes sense

where glacial feature is

A hydrographic feature characterized by the dominance of snow or ice.

and ice mass is

A mass of water ice.

@wiegandn if you have any suggestions for better definitions/hierarchical placement of those terms please let us know.

[The] overly broad logical definition of glacial feature

Yes @cmungall I think it might be good to review these and perhaps also other EQ axioms in ENVO.

In this pull request the relevant EQ axioms are on

cryosphere

'composed primarily of' some 
    ('environmental material'
     and ('has quality' some frozen))

and glacial feature

'material entity'
 and ('composed primarily of' some 
    ('water ice' or snow))

@cmungall let us know if you think the axiom on cryosphere is also more broad then intended. Thanks for illustrating the issue with the axiom on glacial feature.

[smrgeoinfo]

the EQ axioms are tricky. Are the ice cubes in my freezer composed of some 'environmental material'? I don't think they'd be considered part of the cryosphere. Seems like you need something like 'part of the Earth composed primarily of water ice or snow formed by natural processes'

[smrgeoinfo]

Are moraines, outwash plains, eskers glacial features? They are not composed of ice or snow. The Oxford Languages definition of 'glacial' is useful start "relating to, resulting from, or denoting the presence or agency of ice, especially in the form of glaciers." I don't think the term 'glacier' should be used in a definition of glacial-- its circular. Maybe something like 'presence or agency of thick accumulations of naturally occurring water ice'

[wiegandn]

I'm also getting confused about my Point 3 in the issue. There are lots of complications in ENVO's current hierarchy. I think many things need to change. I disagree with ice mass being a subclass of glacial feature. Original question with Kai's response: 3. ice mass should probably be a sibling of glacial feature and not a subclass. (maybe a separate issue) Given the current definitions the way we have it now (ice mass subclass to glacial feature) makes sense where |glacial feature| is A hydrographic feature characterized by the dominance of snow or ice. and |ice mass| is A mass of water ice.

…

On 2/21/2022 10:42 AM, Stephen Richard wrote: the EQ axioms are tricky. Are the ice cubes in my freezer composed of some 'environmental material'? I don't think they'd be considered part of the cryosphere. Seems like you need something like 'part of the Earth composed primarily of water ice or snow formed by natural processes' — Reply to this email directly, view it on GitHub <#1291 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AXWCY5TCXXWKXLFEQMP3D5DU4JTN7ANCNFSM5N6UBTIA>. Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[rduerr]

@wiegandn I agree that there are a lot of issues with the cryosphere structures in ENVO at this moment and even after this next release of ENVO. The question is should these all be fixed at once or is fixing a few at a time sufficient. My take is that fixes are going to have to be piecemeal since no one is funded to make a global set of fixes and even if they were, that might take years! Especially with so many cooks in the kitchen and no uniform understanding of ontological best practices (especially with how to deal with multiple definitions for a term and how to resolve that by making several terms that make the distinctions plain and computable).

I think a new issue should be written about the definition of "glacial feature", since there are clearly at least two different types of glacial features:

1. Features of a glacier (toe, ablation zone, etc.)
2. Landforms created by glaciers (eskers, moraines, etc.) which may no longer have any ice at all (i.e., the glaciers are no longer around)!

The current definition is clearly referring to the first of these and clearly predates establishing annotation conventions for ENVO, since it isn't even dated nor does it have any sort of author or creator. Even so the definition could be improved since it clearly doesn't exclude sea ice or cloud related ice terms (e.g., frazil ice and cirrocumulus clouds are not a glacier features).

However, this issue is about cryosphere being a subclass of glacial feature. Since that is fixed (at least in the #1172 pull request); I say that once that pull request is approved, then this issue should be closed. Specific new issues related to other problems should then be created. Yes, a lot of individual issues; but I don't believe that the whole tree can be addressed at once without real funding for someone!

[wiegandn]

To add to confusions of the what the term 'glacial feature' means, currently subclasses of 'glacial feature' in ENVO are: brine channel, frazil ice, fumarole-derived ice tower, ice field, ice fog, ice mass, sea ice floe, sea ice hummock, snow, snow mass, water ice crystal, and water ice layer (omitting cryosphere)

…

On 2/21/2022 6:54 PM, rduerr wrote: new issue should be written about the definition of "glacial feature", since there are clearly at least two different types of glacial features: 1. Features of a glacier (toe, ablation zone, etc.) 2. Landforms created by glaciers (eskers, moraines, etc.) which may no longer have any ice at all (i.e., the glaciers are no longer around)!

[rduerr]

@wiegandn yup - lot's of new issues to write, though maybe (unless you can look at the #1072 version of the ontology) you should wait to write those until after #1072 is released.

[kaiiam]

I agree with @rduerr, hopefully many of @wiegandn's concerns should be fixed when #1172 is released, but there will of course be more issues.

I agree we might want to sort out glacial feature if you @wiegandn feel it's really wrong. I think it's just a simple grouping class for anything having to do with an existing glacier, however as @cmungall mentioned the EQ axiom could use review. Although the current axiom is helping to infer a lot of subclasses to glacial feature, as Chris mentioned something like a popsicle would also get placed there.

If anyone has a suggestion on what to change perhaps the axiom @cmungall we can add it to #1172 otherwise I think we should review and merge that PR for now.

[cmungall]

Thanks, I agree with all this, and especially with keeping each individual GitHub issue scoped and actionable

…

On Mon, Feb 21, 2022 at 4:54 PM rduerr ***@***.***> wrote: @wiegandn <https://github.com/wiegandn> I agree that there are a lot of issues with the cryosphere structures in ENVO at this moment and even after this next release of ENVO. The question is should these all be fixed at once or is fixing a few at a time sufficient. My take is that fixes are going to have to be piecemeal since no one is funded to make a global set of fixes and even if they were, that might take years! Especially with so many cooks in the kitchen and no uniform understanding of ontological best practices (especially with how to deal with multiple definitions for a term and how to resolve that by making several terms that make the distinctions plain and computable). I think a new issue should be written about the definition of "glacial feature", since there are clearly at least two different types of glacial features: 1. Features of a glacier (toe, ablation zone, etc.) 2. Landforms created by glaciers (eskers, moraines, etc.) which may no longer have any ice at all (i.e., the glaciers are no longer around)! The current definition is clearly referring to the first of these and clearly predates establishing annotation conventions for ENVO, since it isn't even dated nor does it have any sort of author or creator. Even so the definition could be improved since it clearly doesn't exclude sea ice or cloud related ice terms (e.g., frazil ice and cirrocumulus clouds are not a glacier features). However, this issue is about cryosphere being a subclass of glacial feature. Since that is fixed (at least in the #1172 <#1172> pull request); I say that once that pull request is approved, then this issue should be closed. Specific new issues related to other problems should then be created. Yes, a lot of individual issues; but I don't believe that the whole tree can be addressed at once without real funding for someone! — Reply to this email directly, view it on GitHub <#1291 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAMMOPEMDKG4PT45IQZ7RLU4LNCVANCNFSM5N6UBTIA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[kaiiam]

Thanks @cmungall and @wiegandn for the pointing out the issues with the axiom on glacial feature, we've obsoleted it now in #1172.

[kaiiam]

Although there were some interesting points raised here I believe we have dealt with them and have fixed the core issues in a PR. I think this can now be closed. I've added the SWEET label to keep this issue reference-able with our ESIP Semantic Harmonization SWEET-ENVO cryosphere mapping work.