DwC "picklist" extension to MIxS IRIs

older
DwC-MIxS TG meeting follow-up and...

Pier Luigi Buttigieg

22 Jun 2021 22 Jun '21

07:02

Dear TG, Dear Ramona (GSC CIG), Chris, Bill,

In our DwC-MIxS mapping activity, we noted a practice to support the IPT to generate "picklists" (for drop-down menus etc) by adding extensions to IRIs.

This was suggested as a means to link the values to keys in terminology-focused fields.

You can see the link to the "picklists" in the tracker: https://github.com/tdwg/gbwg/tree/main/dwc-mixs/dwc#vocabularies

An example is available in [1] and looks like: dc:URI='https://w3id.org/gensc/terms/MIXS:0000005/contigs'

I'm not quite sure about this and would like your input. To me, it seems like a very strange way to modify an identifier *issued by another standards group* in order to help a tool generate dropdown menu.

It essential creates a false impression that there's something standardised on the other end of that IRI that may or may not be the case.

Best,

Pier Luigi

[1] https://github.com/tdwg/gbwg/blob/main/dwc-mixs/dwc/vocabulary/contam_screen...

-- https://orcid.org/0000-0002-4366-3088

Show replies by date

Tim Robertson

22 Jun 22 Jun

08:18

Thanks, Pier

Just for a bit of background - you're assuming these are IRIs, but the DwC-A work was all designed at a time when URNs were common and expectations of resolvability weren't as high as today. The original design was really just to ensure a globally unique URN was used.

I think the concerns you raise are valid, and especially with the current intent to achieve resolvability for all the w3id.org/... terms. We do need a URN though, and I understand there isn't an intention to make resolvable terms for these concepts. The rs.gbif.org namespace has been used for this kind of thing in the past, which might be an alternative option to lower expectations.

Thanks, Tim

On 22/06/2021, 16.02, "dwc-mixs on behalf of Pier Luigi Buttigieg" <dwc-mixs-bounces@lists.tdwg.org on behalf of pier.buttigieg@awi.de> wrote:

Dear TG, Dear Ramona (GSC CIG), Chris, Bill,

In our DwC-MIxS mapping activity, we noted a practice to support the IPT to generate "picklists" (for drop-down menus etc) by adding extensions to IRIs.

This was suggested as a means to link the values to keys in terminology-focused fields.

You can see the link to the "picklists" in the tracker: https://github.com/tdwg/gbwg/tree/main/dwc-mixs/dwc#vocabularies

An example is available in [1] and looks like: dc:URI='https://w3id.org/gensc/terms/MIXS:0000005/contigs'

It essential creates a false impression that there's something standardised on the other end of that IRI that may or may not be the case.

Best,

Pier Luigi

[1] https://github.com/tdwg/gbwg/blob/main/dwc-mixs/dwc/vocabulary/contam_screen...

-- https://orcid.org/0000-0002-4366-3088

_______________________________________________ dwc-mixs mailing list dwc-mixs@lists.tdwg.org http://lists.tdwg.org/mailman/listinfo/dwc-mixs

Pier Luigi Buttigieg

08:41

Thanks Tim, this clarifies a great deal.

Measures to make it clear that the URNs (wherever they appear, e.g. on documentation pages, lists...) used by legacy systems are not to be understood as resolvable or under the control/custodianship of the GSC would help.

While technically valid, prefixing URNs with HTTP does indeed set a false expectation - anything that can mitigate that would be appreciated.

On the GSC and TDWG side however, it would be very interesting to find out how to do this more sustainably by, e.g., sharing vocabulary files with all the metadata needed to, e.g., map to ontologies, qualify terms with comments, include provenance data, etc

On 22/06/2021 17:18, Tim Robertson wrote:

...

Thanks, Pier

Just for a bit of background - you're assuming these are IRIs, but the DwC-A work was all designed at a time when URNs were common and expectations of resolvability weren't as high as today. The original design was really just to ensure a globally unique URN was used.

I think the concerns you raise are valid, and especially with the current intent to achieve resolvability for all the w3id.org/... terms. We do need a URN though, and I understand there isn't an intention to make resolvable terms for these concepts. The rs.gbif.org namespace has been used for this kind of thing in the past, which might be an alternative option to lower expectations.

Thanks, Tim

On 22/06/2021, 16.02, "dwc-mixs on behalf of Pier Luigi Buttigieg" <dwc-mixs-bounces@lists.tdwg.org on behalf of pier.buttigieg@awi.de> wrote:
 Dear TG, Dear Ramona (GSC CIG), Chris, Bill,

 In our DwC-MIxS mapping activity, we noted a practice to support the IPT
 to generate "picklists" (for drop-down menus etc) by adding extensions
 to IRIs.

 This was suggested as a means to link the values to keys in
 terminology-focused fields.

 You can see the link to the "picklists" in the tracker:
 https://github.com/tdwg/gbwg/tree/main/dwc-mixs/dwc#vocabularies

 An example is available in [1] and looks like:
 dc:URI='https://w3id.org/gensc/terms/MIXS:0000005/contigs'

 I'm not quite sure about this and would like your input. To me, it seems
 like a very strange way to modify an identifier *issued by another
 standards group* in order to help a tool generate dropdown menu.

 It essential creates a false impression that there's something
 standardised on the other end of that IRI that may or may not be the case.

 Best,

 Pier Luigi

 [1]
 https://github.com/tdwg/gbwg/blob/main/dwc-mixs/dwc/vocabulary/contam_screen_input.xml


 --
 https://orcid.org/0000-0002-4366-3088

 _______________________________________________
 dwc-mixs mailing list
 dwc-mixs@lists.tdwg.org
 http://lists.tdwg.org/mailman/listinfo/dwc-mixs

-- https://orcid.org/0000-0002-4366-3088

Tim Robertson

12:20

Thanks, Pier

We'll clearly state expectations in the description, and this will be kept as internal as it can be.

Using the example of the vocabulary of controlled terms for lib_layout[1] currently we have: https://w3id.org/gensc/terms/MIXS:0000041/vector

This could become something like the following e.g. https://rs.gbif.org/vocab/dna/lib_layout/vector

Would this be preferable to the GSC folk to avoid confusion, please? Note that the source from where it is derived will still be cited.

Thanks, Tim

[1] https://rs.gbif.org/sandbox/vocabulary/mixs/lib_layout.xml

On 22/06/2021, 17.41, "Pier Luigi Buttigieg" pier.buttigieg@awi.de wrote:

Thanks Tim, this clarifies a great deal.

While technically valid, prefixing URNs with HTTP does indeed set a false expectation - anything that can mitigate that would be appreciated.

On 22/06/2021 17:18, Tim Robertson wrote: > Thanks, Pier > > Just for a bit of background - you're assuming these are IRIs, but the DwC-A work was all designed at a time when URNs were common and expectations of resolvability weren't as high as today. > The original design was really just to ensure a globally unique URN was used. > > I think the concerns you raise are valid, and especially with the current intent to achieve resolvability for all the w3id.org/... terms. > We do need a URN though, and I understand there isn't an intention to make resolvable terms for these concepts. The rs.gbif.org namespace has been used for this kind of thing in the past, which might be an alternative option to lower expectations. > > Thanks, > Tim > > > > On 22/06/2021, 16.02, "dwc-mixs on behalf of Pier Luigi Buttigieg" <dwc-mixs-bounces@lists.tdwg.org on behalf of pier.buttigieg@awi.de> wrote: > > Dear TG, Dear Ramona (GSC CIG), Chris, Bill, > > In our DwC-MIxS mapping activity, we noted a practice to support the IPT > to generate "picklists" (for drop-down menus etc) by adding extensions > to IRIs. > > This was suggested as a means to link the values to keys in > terminology-focused fields. > > You can see the link to the "picklists" in the tracker: > https://github.com/tdwg/gbwg/tree/main/dwc-mixs/dwc#vocabularies > > An example is available in [1] and looks like: > dc:URI='https://w3id.org/gensc/terms/MIXS:0000005/contigs' > > I'm not quite sure about this and would like your input. To me, it seems > like a very strange way to modify an identifier *issued by another > standards group* in order to help a tool generate dropdown menu. > > It essential creates a false impression that there's something > standardised on the other end of that IRI that may or may not be the case. > > Best, > > Pier Luigi > > [1] > https://github.com/tdwg/gbwg/blob/main/dwc-mixs/dwc/vocabulary/contam_screen... > > > -- > https://orcid.org/0000-0002-4366-3088 > > _______________________________________________ > dwc-mixs mailing list > dwc-mixs@lists.tdwg.org > http://lists.tdwg.org/mailman/listinfo/dwc-mixs > -- https://orcid.org/0000-0002-4366-3088

Ramona Walls

24 Jun 24 Jun

09:17

Thank you both for working this out and explaining the details.

I do think it is less confusing to use the GBIF namespace for URNs, rather than append the GSC URIs. Citing the MIxS URI then allows people to find the original term.

Tim, would it help if we assigned URIs to the values in the CVs? That is our eventual plan.

Ramona

------------------------------------------------------ Ramona L. Walls, Ph.D. Data Scientist, Ontologies, Semantics, and Metadata, Critical Path Institute Assistant Research Professor, Bio5 Institute, University of Arizona Part-time artist (https://www.instagram.com/omnivoresgratitutude/)

On Tue, Jun 22, 2021 at 12:21 PM Tim Robertson trobertson@gbif.org wrote:

...

Thanks, Pier

We'll clearly state expectations in the description, and this will be kept as internal as it can be.

Using the example of the vocabulary of controlled terms for lib_layout[1] currently we have: https://w3id.org/gensc/terms/MIXS:0000041/vector

This could become something like the following e.g. https://rs.gbif.org/vocab/dna/lib_layout/vector

Would this be preferable to the GSC folk to avoid confusion, please? Note that the source from where it is derived will still be cited.

Thanks, Tim

[1] https://rs.gbif.org/sandbox/vocabulary/mixs/lib_layout.xml

On 22/06/2021, 17.41, "Pier Luigi Buttigieg" pier.buttigieg@awi.de wrote:
Thanks Tim, this clarifies a great deal.

Measures to make it clear that the URNs (wherever they appear, e.g. on
documentation pages, lists...) used by legacy systems are not to be
understood as resolvable or under the control/custodianship of the GSC
would help.

While technically valid, prefixing URNs with HTTP does indeed set a
false expectation - anything that can mitigate that would be
appreciated.
On the GSC and TDWG side however, it would be very interesting to find
out how to do this more sustainably by, e.g., sharing vocabulary files
with all the metadata needed to, e.g., map to ontologies, qualify
terms with comments, include provenance data, etc
On 22/06/2021 17:18, Tim Robertson wrote:
> Thanks, Pier
>
> Just for a bit of background - you're assuming these are IRIs, but
the DwC-A work was all designed at a time when URNs were common and expectations of resolvability weren't as high as today. > The original design was really just to ensure a globally unique URN was used. > > I think the concerns you raise are valid, and especially with the current intent to achieve resolvability for all the w3id.org/... terms. > We do need a URN though, and I understand there isn't an intention to make resolvable terms for these concepts. The rs.gbif.org namespace has been used for this kind of thing in the past, which might be an alternative option to lower expectations. > > Thanks, > Tim > > > > On 22/06/2021, 16.02, "dwc-mixs on behalf of Pier Luigi Buttigieg" < dwc-mixs-bounces@lists.tdwg.org on behalf of pier.buttigieg@awi.de> wrote: > > Dear TG, Dear Ramona (GSC CIG), Chris, Bill, > > In our DwC-MIxS mapping activity, we noted a practice to support the IPT > to generate "picklists" (for drop-down menus etc) by adding extensions > to IRIs. > > This was suggested as a means to link the values to keys in > terminology-focused fields. > > You can see the link to the "picklists" in the tracker: > https://github.com/tdwg/gbwg/tree/main/dwc-mixs/dwc#vocabularies > > An example is available in [1] and looks like: > dc:URI='https://w3id.org/gensc/terms/MIXS:0000005/contigs' > > I'm not quite sure about this and would like your input. To me, it seems > like a very strange way to modify an identifier *issued by another > standards group* in order to help a tool generate dropdown menu. > > It essential creates a false impression that there's something > standardised on the other end of that IRI that may or may not be the case. > > Best, > > Pier Luigi > > [1] > https://github.com/tdwg/gbwg/blob/main/dwc-mixs/dwc/vocabulary/contam_screen... > > > -- > https://orcid.org/0000-0002-4366-3088 > > _______________________________________________ > dwc-mixs mailing list > dwc-mixs@lists.tdwg.org > http://lists.tdwg.org/mailman/listinfo/dwc-mixs > -- https://orcid.org/0000-0002-4366-3088

Tim Robertson

09:24

Thanks Ramona

...

would it help if we assigned URIs to the values in the CVs? That is our eventual plan

Yes, longer term this makes sense. I anticipate we’ll probably arrive at SKOS or similar for these kind of vocabularies.

...

…it is less confusing to use the GBIF namespace for URNs

Thanks. We’ll change to use GBIF namespaces unless someone brings new information and will update the TG report accordingly.

Best wishes, Tim

From: Ramona Walls rlwalls2008@gmail.com Date: Thursday, 24 June 2021 at 18.18 To: Tim Robertson trobertson@gbif.org Cc: Pier Luigi Buttigieg pier.buttigieg@awi.de, "dwc-mixs@lists.tdwg.org" dwc-mixs@lists.tdwg.org, Chris Mungall cjmungall@lbl.gov, Bill Duncan wdduncan@gmail.com Subject: Re: [dwc-mixs] DwC "picklist" extension to MIxS IRIs

Thank you both for working this out and explaining the details.

I do think it is less confusing to use the GBIF namespace for URNs, rather than append the GSC URIs. Citing the MIxS URI then allows people to find the original term.

Tim, would it help if we assigned URIs to the values in the CVs? That is our eventual plan.

Ramona

On Tue, Jun 22, 2021 at 12:21 PM Tim Robertson <trobertson@gbif.orgmailto:trobertson@gbif.org> wrote: Thanks, Pier

We'll clearly state expectations in the description, and this will be kept as internal as it can be.

Using the example of the vocabulary of controlled terms for lib_layout[1] currently we have: https://w3id.org/gensc/terms/MIXS:0000041/vector

This could become something like the following e.g. https://rs.gbif.org/vocab/dna/lib_layout/vector

Would this be preferable to the GSC folk to avoid confusion, please? Note that the source from where it is derived will still be cited.

Thanks, Tim

[1] https://rs.gbif.org/sandbox/vocabulary/mixs/lib_layout.xml

On 22/06/2021, 17.41, "Pier Luigi Buttigieg" <pier.buttigieg@awi.demailto:pier.buttigieg@awi.de> wrote:

Thanks Tim, this clarifies a great deal.

While technically valid, prefixing URNs with HTTP does indeed set a false expectation - anything that can mitigate that would be appreciated.

On 22/06/2021 17:18, Tim Robertson wrote: > Thanks, Pier > > Just for a bit of background - you're assuming these are IRIs, but the DwC-A work was all designed at a time when URNs were common and expectations of resolvability weren't as high as today. > The original design was really just to ensure a globally unique URN was used. > > I think the concerns you raise are valid, and especially with the current intent to achieve resolvability for all the w3id.org/.http://w3id.org/.. terms. > We do need a URN though, and I understand there isn't an intention to make resolvable terms for these concepts. The rs.gbif.orghttp://rs.gbif.org namespace has been used for this kind of thing in the past, which might be an alternative option to lower expectations. > > Thanks, > Tim > > > > On 22/06/2021, 16.02, "dwc-mixs on behalf of Pier Luigi Buttigieg" <dwc-mixs-bounces@lists.tdwg.orgmailto:dwc-mixs-bounces@lists.tdwg.org on behalf of pier.buttigieg@awi.demailto:pier.buttigieg@awi.de> wrote: > > Dear TG, Dear Ramona (GSC CIG), Chris, Bill, > > In our DwC-MIxS mapping activity, we noted a practice to support the IPT > to generate "picklists" (for drop-down menus etc) by adding extensions > to IRIs. > > This was suggested as a means to link the values to keys in > terminology-focused fields. > > You can see the link to the "picklists" in the tracker: > https://github.com/tdwg/gbwg/tree/main/dwc-mixs/dwc#vocabularies > > An example is available in [1] and looks like: > dc:URI='https://w3id.org/gensc/terms/MIXS:0000005/contigs' > > I'm not quite sure about this and would like your input. To me, it seems > like a very strange way to modify an identifier *issued by another > standards group* in order to help a tool generate dropdown menu. > > It essential creates a false impression that there's something > standardised on the other end of that IRI that may or may not be the case. > > Best, > > Pier Luigi > > [1] > https://github.com/tdwg/gbwg/blob/main/dwc-mixs/dwc/vocabulary/contam_screen... > > > -- > https://orcid.org/0000-0002-4366-3088 > > _______________________________________________ > dwc-mixs mailing list > dwc-mixs@lists.tdwg.orgmailto:dwc-mixs@lists.tdwg.org > http://lists.tdwg.org/mailman/listinfo/dwc-mixs > -- https://orcid.org/0000-0002-4366-3088

1172

Age (days ago)

1174

Last active (days ago)

List overview

Download

5 comments

3 participants

tags (0)

participants (3)

Pier Luigi Buttigieg
Ramona Walls
Tim Robertson