Various suggestions for improving definitions and labels

[mbrush]

Hi. Reporting a handful of minor issues i noted on reviewing the CRO - mostly concerning definitions and labels. Happy to split these out int separate tickets if you want - just say the word.

• [x] 1. As noted by others, a challenge in evaluating the ontology is that so many terms lack definitions - and many that exist are not sufficiently discriminating. This can make it hard to evaluate how some terms are different from each other, and if their classification makes sense. I would make adding defs a top priority - and making sure the defs are clear and precise to distinguish between classes. Also, be sure to include a definition for the root 'contributor role' class as well, and perhaps some clarifying comments- to provide a foundational understanding of what we mean by a 'contribution' in general. See https://github.com/data2health/contributor-role-ontology/issues/32

• [x] 2. 'Author role' needs better definition. Currently reads "Contributions to the published research object." Defs of most children start with and then extend the phrase "Preparation, creation and/or presentation of the published work" so makes sense to use this in some way as the definition of the 'author role' parent. Also, the def or a comment should clarify the types of artifacts that are "authored" - to help scope when author roles should be used to describe creation of an artifact. e.g. is this only for things like books, journal articles, and other publications? or can things like technical documentation, web sites, code, ontologies, or data model be authored?

• [ ] 3. Other pairs of terms that are not clearly distinguished (to add to the list started by Anne in #20):

  • [x] - 'writing original draft role' and 'original draft preparation role' - defs are nearly identical. See https://github.com/data2health/contributor-role-ontology/issues/52
  • [x] - 'conceptualization role' and 'research conceptualization role' - identical definitions (one has Credit prefix and the other CRO - not sure what the relationship is here)
    • [x] - 'educational training role' and 'educational instruction role' - lack definitions, but sound very much the same.
  • [x] - the definition for 'program administration role' sounds like it may be the same thing as 'research coordination role' - but the latter has no definition so I cannot be sure.
  • [x] - 'IT hardware systems design and implementation role' and 'hardware systems role' - no def for the latter, but sound similar by name.

• [x] 4. I'd like to see a definition for 'data role' in particular, to be able to evaluate what belongs in this hierarchy of terms.

  • [x] - e.g. things like data standards development and data modeling may not involve actually working with data - s depending on how data role is defined, these may belong elsewhere.
    • [x] - and 'formal analysis role' and 'visualization role' needs improved definition as well - as I am not convinced these belong here either.

• [x] 5. 'data role' > 'data visualization role' has same text definition as 'author role' > 'figure development role'. These should be clearly distinguished as they are different things. and also noticed that 'data role' > formal analysis role' also has this same definition, which is clearly a mistake.

• [x] 6. 'Infrastructure role' is not defined, but from its context, I'm not sure I like the name of this class. Seems too generic/vague.

• [x] 7. 'investigation role' as a label seems too generic - given that I think its intended meaning is to describe someone who actually executed experiments/computational analyses in the course of a research project. With such a generic name, I would expect to find other types of contributions to investigations here - e.g. funding acquisition role, conceptualization role, study design role, etc. NV: this comes from CREDIT, we would have to change the CREDIT ontology to address this. @kristiholmes or @marijane could you comment?

• [x] 8. Rename 'resource role' to 'resource provision role'?

  • [x] - 'resources role' is quite vague - the def seems to describe someone who provides resources used in a research project. NV: this comes from CREDIT, we would have to change the CREDIT ontology to address this. @kristiholmes or @marijane could you comment?

• [x] 9. Is there a good reason that software systems role is not classified under 'software role' NV: Could we give classes multiple parents? @mellybelly or @marijane could you comment?

• [x] 10. 'standards development role' sounds like it should subsume some other classes in the ontology. e.g. 'standard operating procedure development role', 'data standards developer role' NV: Could we give classes multiple parents? @mellybelly or @marijane could you comment?

• [x] 11. 'validation role' could use a more specific name, given its def. NV: this comes from CREDIT, we would have to change the CREDIT ontology to address this. @kristiholmes or @marijane could you comment?

[nicolevasilevsky]

Thanks @mbrush, this is incredibly helpful!

@marijane created a list of terms with missing defs, and I created a separate ticket (#32)

I'll take a look at all your comments asap and will work through them.

Thank you!

[nicolevasilevsky]

I think we've addresses points 1 and 2

for point 3 - I will obsolete CRO_0000088 'original draft preparation role' and move CREDIT_00000013 'writing original draft role' to a child of 'author role'

[nicolevasilevsky]

I will obsolete CRO_0000086 'research conceptualization role'

[nicolevasilevsky]

'educational training role' and 'educational instruction role' - lack definitions, but sound very much the same. --> I obsoleted educational training role on this PR: https://github.com/data2health/contributor-role-ontology/pull/45

[nicolevasilevsky]

'IT hardware systems design and implementation role' and 'hardware systems role' - I added defs, there is a pending PR, we should revisit this once the PR is merged

[nicolevasilevsky]

a suggested def for data role is here #42

the def of formal analysis role and 'visualization role' come from casrai. @kristiholmes do you think we should update these?

[nicolevasilevsky]

@mbrush point 5 - there isn't a class called 'data visualization role', do you mean CREDIT_00000012 'visualization role'? Never mind I see that is what you meant :)

[nicolevasilevsky]

point 5 - I will update the definition of 'figure development role'

[nicolevasilevsky]

point 8 - Rename 'resource role' to 'resource provision role'. I am having the same problem noted in #52

[nicolevasilevsky]

point 9 - @marijane @mellybelly, could we add multiple parents here?

[kristiholmes]

hi @nicolevasilevsky - resources role (from 8 above) seems really important to keep as-is, actually. I would especially see this as a contributor role that an organization might assert eg a core that provides resources (which I would argue could be expanded to include services or financial resources) used in the execution of a research activity. What do you think?

from CREDIT: Provision of study materials, reagents, materials, patients, laboratory samples, animals, instrumentation, computing resources, or other analysis tools

[kristiholmes]

hi @nicolevasilevsky and @marijane - regarding 7 above, I think it is important to keep investigation role (and every CREDIT concept) in their WT form. I would see this as something that could be a more general concept. Perhaps I am not understanding what you are asking.

To confirm, all of the CREDIT classes are still in this data model, correct? thanks for your work on this - let me know if it would be helpful to have a 10 min chat.

[kristiholmes]

hi @nicolevasilevsky and @marijane - regarding 11 above, same basic question/opinion from me as I mentioned on 7 and 8. :-)

[nicolevasilevsky]

Hi @kristiholmes sounds good to leave the CREDIT classes in their original form. Yes, nothing has been changed for the CREDIT classes, they are still in CRO and in their original form. Thanks!

[marijane]

my thoughts:

• i agree that we should not change names or definitions for roles that are in credit; i would consider that term hijacking. i would be okay with adding alternate or preferred labels/definitions in the CRO context as long as they are consistent with/in the spirit of the original label or definition in credit.
• this makes me realize that some of the duplicate-ish classes that have been obsoleted could have labels that could be used as alternate labels
• i am fine with giving classes multiple parents, in fact i think this may be the only available course of action for credit roles that fit under CRO roles, but it does also seem very appropriate for the software systems role.

[nicolevasilevsky]

good point - I added alt terms from obsoleted term on the 'replaced by' terms

[nicolevasilevsky]

I added an alt term to 'resources role'

[nicolevasilevsky]

I think everything on this ticket has been addresses now - pending PRs

[nicolevasilevsky]

The PRs have been merged, closing