search

mcwdsi / geographical-entity-ontology

An ontology and inventory of geopolitical entities (such as sovereign states and their administrative subdivisions) as well as various geographical regions (including but not limited to the specific ones over which the governments have jurisdiction)
Creative Commons Attribution 4.0 International
8 stars 4 forks source link

GEO prefix for Geographical Entity Ontology conflicts with Gene Expression Omnibus #19

Open cthoyt opened 3 years ago

cthoyt commented 3 years ago

I've been doing a lot of prefix mapping across various registries and been pleasently surprised to find few conflicts. Unfortunately, both the Geographical Entity Ontology and the Gene Expression Omnibus have used GEO as a prefix. Geographical Entity Ontology is listed in BioPortal, Prefix Commons, OLS, and the OBO Foundry, while Gene Expression Omnibus is listed in Identifiers.org.

However, I think Geographical Entity Ontology was only recently listed in some of these resources. Now it's causing a conflict. Would you consider using a different prefix for your ontology, perhaps geogeo?

hoganwr commented 3 years ago

I'm not sure I understand. OBO identifiers are not intended to be globally unique (although they are when you include the full IRI). How could they conflict with other non-OBO resources?

Or is Gene Expression Omnibus somehow now included in OBO scope?

Bill

On Mon, May 24, 2021 at 3:27 PM Charles Tapley Hoyt < @.***> wrote:

I've been doing a lot of prefix mapping across various registries and been pleasently surprised to find few conflicts. Unfortunately, both the Geographical Entity Ontology and the Gene Expression Omnibus have used GEO as a prefix. Geographical Entity Ontology is listed in BioPortal, Prefix Commons, OLS, and the OBO Foundry, while Gene Expression Omnibus is listed in Identifiers.org.

However, I think Geographical Entity Ontology was only recently listed in some of these resources. Now it's causing a conflict. Would you consider using a different prefix for your ontology, perhaps geogeo?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ufbmi/geographical-entity-ontology/issues/19, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAJR55QCL7UMBMAHDXWMQILTPKSB5ANCNFSM45NZFSXA .

cthoyt commented 3 years ago

It's definitely true that OBO Foundry/OLS aren't enforcing global uniqueness (that would be a pipe dream). It's just by a happy chance that conflicts happen very infrequently with the common prefixes used more generally in the bioinformatics community. Unfortunately, this is one of those cases.

I'll admit I know relatively little about your ontology. Based on its name, I might guess that you're less familiar with bioinformatics resources than semantic web resources.

I'm working on the Bioregistry (https://bioregistry.io/), which is a much larger registry that's combining/normalizing many different ones. Either way, I will have to internally solve the issue of the GEO clash. Since in bioinformatics, GEO always means Gene Expression Omnibus, please let me know under what prefix you would like your ontology listed there (right now, it's geoeo)

hoganwr commented 3 years ago

I'm not entirely sure what you mean by "prefix". I mostly understand it to mean xmlns prefixes.

In that case, we don't use geo as a prefix anywhere in GEO. All GEO IRIs have an obo prefix.

I don't fully understand all the consequences of this decision, but I think it's probably fine to list us under your "prefix" geogeo

Bill

On Tue, May 25, 2021 at 1:01 PM Charles Tapley Hoyt < @.***> wrote:

It's definitely true that OBO Foundry/OLS aren't enforcing global uniqueness (that would be a pipe dream). It's just by a happy chance that conflicts happen very infrequently with the common prefixes used more generally in the bioinformatics community. Unfortunately, this is one of those cases.

I'll admit I know relatively little about your ontology. Based on its name, I might guess that you're less familiar with bioinformatics resources than semantic web resources.

I'm working on the Bioregistry (https://bioregistry.io/), which is a much larger registry that's combining/normalizing many different ones. Either way, I will have to internally solve the issue of the GEO clash. Since in bioinformatics, GEO always means Gene Expression Omnibus, please let me know under what prefix you would like your ontology listed there (right now, it's geoeo)

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ufbmi/geographical-entity-ontology/issues/19#issuecomment-848051455, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAJR55QS7D55FDEEV2ZFYDDTPPJVNANCNFSM45NZFSXA .

cthoyt commented 3 years ago

Sorry to throw around "prefix" as some jargon. I mean the key that's used in OBO Foundry and OLS:

Identifiers.org, which also covers many of the ontologies in OBO Foundry/OLS uses GEO for Gene Expression Omnibus:

hoganwr commented 3 years ago

Thanks for the clarification. Good information to know.

So as I understand it, your action affects bioregistry but not identifiers.org, correct? In which case would it be useful to ensure we have the same key in both, and thus should try to register geogeo w/ identifiers.org?

On Tue, May 25, 2021 at 1:11 PM Charles Tapley Hoyt < @.***> wrote:

Sorry to throw around "prefix" as some jargon. I mean the key that's used in OBO Foundry and OLS:

Identifiers.org, which also covers many of the ontologies in OBO Foundry/OLS uses GEO for Gene Expression Omnibus:

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ufbmi/geographical-entity-ontology/issues/19#issuecomment-848058347, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAJR55X4DQUUV4BYKEXPYKLTPPK4JANCNFSM45NZFSXA .

cthoyt commented 3 years ago

Yeah, that would be a great way to increase visibility of your work! Unfortunately one of the reasons I've been working on the Bioregistry is to make a more community driven version of Identifiers.org. I can't vouch that they will be responsive to requests to add new content

hoganwr commented 3 years ago

Fair enough. Right now, I've got all green checkboxes on the request form, and I get an error when submitting that says to try again later....

Bill

On Tue, May 25, 2021 at 1:23 PM Charles Tapley Hoyt < @.***> wrote:

Yeah, that would be a great way to increase visibility of your work! Unfortunately one of the reasons I've been working on the Bioregistry is to make a more community driven version of Identifiers.org. I can't vouch that they will be responsive to requests to add new content

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ufbmi/geographical-entity-ontology/issues/19#issuecomment-848068090, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAJR55RNTZYJ2UK6LPKKOPDTPPMJFANCNFSM45NZFSXA .

cmungall commented 3 years ago

I would recommend having a more systematic way of allocating namespaces in the case of conflicts. Can we use one of:

In future, I do not think OBO should grant namespaces that are in use by major databases (https://github.com/OBOFoundry/OBOFoundry.github.io/issues/1522)

hoganwr commented 3 years ago

Already applied for and received geogeo

Maybe I could ask them to change

On Sun, May 30, 2021 at 10:47 PM Chris Mungall @.***> wrote:

I would recommend having a more systematic way of allocating namespaces in the case of conflicts. Can we use one of:

  • obogeo
  • geoont

In future, I do not think OBO should grant namespaces that are in use by major databases

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ufbmi/geographical-entity-ontology/issues/19#issuecomment-851129199, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAJR55TT2YQL5NGJE5V45ZLTQL2D7ANCNFSM45NZFSXA .

cmungall commented 3 years ago

'Them' being identifiers.org?

Yes, they can surely update it.

I think you would want to do this anyway. Whoever registered this made the same mistake that is often made when registering the OBO ontologies of including the prefix in the local part.

https://registry.identifiers.org/registry/geogeo

Your identifiers are "geogeo:GEO_000000021"

I have seen the doubling up before but this is the first time I have seen a triplification. I hope you agree this is undesirable!

I recommend:

hoganwr commented 3 years ago

I registered it as a trial run, and I learned a lot.

I used geogeo because it was suggested to me by Charles, who seems to have much expertise in this area (more than I have, so I certainly didn't see a reason to question it).

I'll file an update request soon, although iran't really registering an entirely new prefix?

Bill

On Tue, Jun 1, 2021 at 11:32 AM Chris Mungall @.***> wrote:

'Them' being identifiers.org?

Yes, they can surely update it.

I think you would want to do this anyway. Whoever registered this made the same mistake that is often made when registering the OBO ontologies of including the prefix in the local part.

https://registry.identifiers.org/registry/geogeo

Your identifiers are "geogeo:GEO_000000021"

I have seen the doubling up before but this is the first time I have seen a triplification. I hope you agree this is undesirable!

I recommend:

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ufbmi/geographical-entity-ontology/issues/19#issuecomment-852221654, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAJR55TNM5AWTBTHKLQADE3TQT4QZANCNFSM45NZFSXA .