draft a GenreFacetType ontology collecting most of PRONOM, NFO, UDFR #2

barmintor · 2015-04-10T20:47:56Z

No description provided.

awoods · 2015-04-10T21:04:59Z

We should establish a set of initial "committers" for this PCDM Git-repository. Otherwise, what is the vetting approach for updates/additions such as this?

acoburn · 2015-04-12T01:02:07Z

@awoods regarding committers for this, I'm inclined to cast a pretty wide net, at least initially.

As for this particular contribution, I find it to be remarkably similar to an ontology that I maintain for our local repository https://github.com/AmherstCollege/acdc-ontology/blob/master/rdf/objectTypes.rdf (we use skos:mappingRelation instead of owl:sameAs, and we're more focused on Getty AAT, LOC, schema.org and opengraph). Either way, I'd be supportive of something like this, especially since I'd much rather rely on a community supported mapping than maintain our own.

/cc @escowles

barmintor · 2015-04-12T14:11:33Z

@acoburn interesting to see some little differences that I am guessing emerge from arriving from a PRONOM/digipres vector vs Getty/cataloguing one. Do you agree that there's value in preserving the narrower scope of this document? If skos relations are a better description of the lineage here, do you want to add a commit to that branch with the change? I'd like the commit history to give appropriate credt for authorship!

acoburn · 2015-04-13T13:04:49Z

@barmintor I prefer skos, because it doesn't imply the strong assertions of identity that accompany owl:sameAs.

awoods · 2015-04-13T14:06:18Z

On a side note, what do you think about a "call for committers" message to the lists?
@escowles, @acoburn, @barmintor ??

Maybe taking nominations for 5 people?

acoburn · 2015-04-13T14:09:45Z

That sounds like a great idea!

barmintor · 2015-04-14T15:17:32Z

This draft should also include:

mappings to LoC and Getty concepts where relevant
guidance on usage of the terms

escowles · 2015-04-17T11:39:31Z

file-format-types.rdf

+    <rdfs:subClassOf rdf:resource="Document" />
+    <skos:exactMatch rdf:resource="http://reference.data.gov.uk/technical-registry/Dataset" />
+    <skos:exactMatch rdf:resource="http://www.udfr.org/onto#DatasetGenre" />
+    <skos:exactMatch rdf:resource="http://purl.org/dc/elements/1.1/Dataset" />


Should be http://purl.org/dc/dcmitype/Dataset, right? Same for other DCMI Type terms.

Correct -- those should be DCMI Types

azaroth42 · 2015-04-17T19:41:13Z

Can we keep contributions separate from the core in a different directory structure somehow?

acoburn · 2015-04-17T19:47:13Z

such as a ./pcdm-core directory and a ./pcdm-ext directory?

azaroth42 · 2015-04-17T19:49:46Z

That sounds good to me. My concern is confusing adopters with what counts as PCDM (which at the moment is not very much) and what is suggested extensions that could be adopted or ignored without peril.

escowles · 2015-04-17T20:28:40Z

👍 to a pcdm-ext directory to hold non-core stuff. I agree it makes sense to make it clear what's core PCDM and what's more peripheral. The file use vocab could potentially go in the ext directory, too.

mjgiarlo · 2015-04-17T20:33:35Z

👍 and I've also seen contrib used similarly but w/e.

awoods · 2015-04-17T20:34:31Z

👍 ./pcdm-ext for the extras. Should the core be at the top-level or in its own directory?

acoburn · 2015-04-17T20:58:03Z

Putting models.rdf into a ./pcdm-core directory would give (possibly) a stronger indication to someone browsing github that the core models are located there. Keeping it at the root level does so implicitly, and I'm fine with that if you think that's sufficiently clear.

awoods · 2015-04-17T22:41:53Z

There is an interesting analogy with what is being discussed on the Fedora side (ontology), questioning where the core ontology goes and where the extra ontologies go. So far, the discussion has been leaning towards a separate Git repository for the extras. Maybe it is simpler, however, to just split them into directories.
In any case, consistency across these efforts would be a treat.

changed remaining owl:sameAs to skos:*Match

added loc and getty concepts

changed dc to dcmi for formatTypes

ruebot · 2015-06-04T12:28:48Z

Should we come back to this one, and vote/merge it?

escowles · 2015-06-04T13:33:04Z

Yes, I think we should resolve this -- I've created a PR to add a new OfficeDocument supertype of Presentation, Spreadsheet and a newly-added WordProcessingDocument: barmintor#6

awoods · 2015-06-05T14:36:42Z

Can someone champion this PR to bring it to resolution?

scossu · 2015-06-05T15:44:58Z

file-format-types.rdf

+  </udfrs:GenreFacetType>
+  <udfrs:GenreFacetType rdf:ID="Spreadsheet">
+    <rdfs:isDefinedBy rdf:resource="http://pcdm.org/file-format-types#"/>
+    <rdfs:subClassOf rdf:resource="Document" />


Would it be more precise for this to be a subclass of Dataset or Database?

I think Spreadsheet could be a subclass of Dataset, as well as a subclass of an OfficeDocument type I introduced in barmintor#6

I'm still a little suspicious of the OfficeDocument superclass: It seems
like it reifies a marketing strategy, doesn't it?

On Fri, Jun 5, 2015 at 11:56 AM, Esmé Cowles [email protected]
wrote:

In file-format-types.rdf
#2 (comment):

<udfrs:GenreFacetType rdf:ID="HTML">

<rdfs:isDefinedBy rdf:resource="http://pcdm.org/file-format-types#"/>

<rdfs:subClassOf rdf:resource="Markup" />

<skos:exactMatch rdf:resource="http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#HTMLDocument" />

<skos:exactMatch rdf:resource="http://vocab.getty.edu/aat/300266021" />

/udfrs:GenreFacetType

<udfrs:GenreFacetType rdf:ID="Presentation">

<rdfs:isDefinedBy rdf:resource="http://pcdm.org/file-format-types#"/>

<rdfs:subClassOf rdf:resource="Document" />

<skos:exactMatch rdf:resource="http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#Presentation" />

<skos:exactMatch rdf:resource="http://reference.data.gov.uk/technical-registry/Presentation" />

<skos:exactMatch rdf:resource="http://www.udfr.org/onto#PresentationGenre" />

/udfrs:GenreFacetType

<udfrs:GenreFacetType rdf:ID="Spreadsheet">

<rdfs:isDefinedBy rdf:resource="http://pcdm.org/file-format-types#"/>

<rdfs:subClassOf rdf:resource="Document" />

I think Spreadsheet could be a subclass of Dataset, as well as a subclass
of an OfficeDocument type I introduced in barmintor#6
barmintor#6

—
Reply to this email directly or view it on GitHub
https://github.com/duraspace/pcdm/pull/2/files#r31825169.

@escowles According to your comment: https://github.com/duraspace/pcdm/pull/2/files#r31825042 Database seems a more appropriate superclass, no?

I don't think OfficeDocument is just a marketing strategy -- it's also a functional category of documents: the kind of documents that typical users are likely to produce. But I'm fine with deferring that for now, since I haven't heard anyone besides me support that.

Though I do think that, separate from the OfficeDocument issue, the WordProcessingDocument is a real missing type: https://github.com/barmintor/pcdm/pull/6/files#diff-c8440ea0bf1d35562d43e1323c36f30dR157

I think it is a mistake to drift from the goal of making a
union/reconciliation set of these vocabularies to trying to sort out an
ontological hierarchy. We are going to spend a lot of time arguing about,
eg, how a MS-Word document is really different from a PDF, and in the end
it is going to chafe against cataloging concerns.

On Fri, Jun 5, 2015 at 12:06 PM, Stefano Cossu [email protected]
wrote:

In file-format-types.rdf
#2 (comment):

<udfrs:GenreFacetType rdf:ID="HTML">

<rdfs:isDefinedBy rdf:resource="http://pcdm.org/file-format-types#"/>

<rdfs:subClassOf rdf:resource="Markup" />

<skos:exactMatch rdf:resource="http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#HTMLDocument" />

<skos:exactMatch rdf:resource="http://vocab.getty.edu/aat/300266021" />

/udfrs:GenreFacetType

<udfrs:GenreFacetType rdf:ID="Presentation">

<rdfs:isDefinedBy rdf:resource="http://pcdm.org/file-format-types#"/>

<rdfs:subClassOf rdf:resource="Document" />

<skos:exactMatch rdf:resource="http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#Presentation" />

<skos:exactMatch rdf:resource="http://reference.data.gov.uk/technical-registry/Presentation" />

<skos:exactMatch rdf:resource="http://www.udfr.org/onto#PresentationGenre" />

/udfrs:GenreFacetType

<udfrs:GenreFacetType rdf:ID="Spreadsheet">

<rdfs:isDefinedBy rdf:resource="http://pcdm.org/file-format-types#"/>

<rdfs:subClassOf rdf:resource="Document" />

@escowles https://github.com/escowles According to your comment:
https://github.com/duraspace/pcdm/pull/2/files#r31825042 Database seems a
more appropriate superclass, no?

—
Reply to this email directly or view it on GitHub
https://github.com/duraspace/pcdm/pull/2/files#r31826088.

I think the intent of PageDescription's source classes is to encompass
texts produced by layout/word-processing programs. I wasn't able to
satisfactorily defend a distinction to our metadata librarian, anyway.

On Fri, Jun 5, 2015 at 12:12 PM, Esmé Cowles [email protected]
wrote:

In file-format-types.rdf
#2 (comment):

<udfrs:GenreFacetType rdf:ID="HTML">

<rdfs:isDefinedBy rdf:resource="http://pcdm.org/file-format-types#"/>

<rdfs:subClassOf rdf:resource="Markup" />

<skos:exactMatch rdf:resource="http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#HTMLDocument" />

<skos:exactMatch rdf:resource="http://vocab.getty.edu/aat/300266021" />

/udfrs:GenreFacetType

<udfrs:GenreFacetType rdf:ID="Presentation">

<rdfs:isDefinedBy rdf:resource="http://pcdm.org/file-format-types#"/>

<rdfs:subClassOf rdf:resource="Document" />

<skos:exactMatch rdf:resource="http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#Presentation" />

<skos:exactMatch rdf:resource="http://reference.data.gov.uk/technical-registry/Presentation" />

<skos:exactMatch rdf:resource="http://www.udfr.org/onto#PresentationGenre" />

/udfrs:GenreFacetType

<udfrs:GenreFacetType rdf:ID="Spreadsheet">

<rdfs:isDefinedBy rdf:resource="http://pcdm.org/file-format-types#"/>

<rdfs:subClassOf rdf:resource="Document" />

I don't think OfficeDocument is just a marketing strategy -- it's also a
functional category of documents: the kind of documents that typical users
are likely to produce. But I'm fine with deferring that for now, since I
haven't heard anyone besides me support that.

Though I do think that, separate from the OfficeDocument issue, the
WordProcessingDocument is a real missing type:
https://github.com/barmintor/pcdm/pull/6/files#diff-c8440ea0bf1d35562d43e1323c36f30dR157

—
Reply to this email directly or view it on GitHub
https://github.com/duraspace/pcdm/pull/2/files#r31826593.

I think that's definitely true of nfo:PaginatedTextDocument -- but I don't agree about pronom:PageDescription or udfr:PageDescriptionGenre. Pronom has a separate WordprocessedText class, which seems to pretty clearly indicate that it's separate. There's not a lot to go on for UDFR since there really aren't any descriptions, examples or relations defined, but my reading of udfr:DocumentGenre is that it's more like pronom:WordprocessedText, too (distinct from the other genres, not a superclass of them).

ruebot · 2015-07-17T15:37:52Z

So, we just need to resolve this before we can merge this? Am I understanding that correctly?

escowles · 2015-07-24T12:21:49Z

This has dragged on too long, so I've closed barmintor#6 and I'm 👍 on merging this as-is. Moving the file to pcdm-ext/file-format-types.rdf would be a good improvement, though.

barmintor · 2015-07-24T14:04:25Z

Esme how would you feel about breaking the file refactoring and office type
into 1 pr and the inheritance and word processing type changes into another
(or two)? I feel like we're throwing out two uncontested improvements with
the bathwater that wants more discussion.
On Jul 24, 2015 8:21 AM, "Esmé Cowles" [email protected] wrote:

This has dragged on too long, so I've closed barmintor#6
barmintor#6 and I'm [image: 👍] on
merging this as-is. Moving the file to pcdm-ext/file-format-types.rdf would
be a good improvement, though.

—
Reply to this email directly or view it on GitHub
#2 (comment).

escowles · 2015-07-24T14:19:19Z

Ben, I've opened barmintor#7 for just the move, since that's the thing that complicates discussing other changes. Once that's merged, I can open a couple of other PRs for the other changes.

Moving file-format-types.rdf to pcdm-ext directory

acoburn · 2015-09-09T15:59:40Z

What is the status of this PR?

ruebot · 2015-09-09T16:06:04Z

I think we're good to vote on it now since we have @escowles PR merged into @barmintor's PR.

awoods · 2015-09-09T16:06:35Z

I think a thumbs-up from @escowles and @barmintor would seal the deal.

azaroth42 · 2015-09-09T16:08:29Z

Agree, I think all other objections were resolved with pcdm-ext.

ruebot · 2015-09-09T16:09:39Z

...having yet to vote on this. 👍

escowles · 2015-09-09T16:22:24Z

👍

barmintor · 2015-09-09T17:34:21Z

+1

awoods · 2015-09-09T17:36:12Z

This is ready to go. I will squash and commit it unless someone else is already on it.

ruebot · 2015-09-09T17:36:37Z

@awoods all you 😄

awoods · 2015-09-09T17:38:23Z

Resolved with: c6b4cd7

ruebot · 2015-09-09T17:40:23Z

@awoods shall we update #24?

awoods · 2015-09-09T17:52:11Z

Sounds like a great idea, @ruebot. If you create the xsl, I will push it to pcdm.org... along with file-format-types.rdf.

ruebot · 2015-09-09T17:54:03Z

@awoods cool. i'll see if i can get that done by end of day.

ruebot · 2015-09-09T19:10:20Z

...and I just noticed something as I was working on the stylesheet. This owl, all the other ones are rdfs. Should we make this rdfs?

draft a GenreFacetType ontology collecting most of PRONOM, NFO, UDFR

f1c3a9b

changed vocab concept mappings from owl: to skos:

d8f08f9

changed remaining owl:sameAs to skos:*Match

31fbd01

escowles reviewed Apr 17, 2015
View reviewed changes

barmintor and others added 5 commits April 22, 2015 13:13

Merge pull request #2 from barmintor/more-skos-mappings

ed2d350

changed remaining owl:sameAs to skos:*Match

added loc and getty concepts

646949b

Merge pull request #3 from barmintor/loc-getty-genres

3338328

added loc and getty concepts

changed dc to dcmi for formatTypes

b5d205f

Merge pull request #4 from barmintor/dcmi-types

a77787c

changed dc to dcmi for formatTypes

scossu reviewed Jun 5, 2015
View reviewed changes

Moving file-format-types.rdf to pcdm-ext directory

c0eb4f6

Merge pull request #7 from escowles/format-genres-ext

6458d87

Moving file-format-types.rdf to pcdm-ext directory

awoods closed this Sep 9, 2015

ruebot mentioned this pull request Sep 9, 2015

file-format-types.rdf is OWL, everything else is RDFS. file-format-types.rdf should be RDFS #26

Closed

awoods mentioned this pull request Apr 20, 2016

Release process #51

Closed

scossu mentioned this pull request Aug 30, 2016

Clarify a FileSet's membership #57

Closed

draft a GenreFacetType ontology collecting most of PRONOM, NFO, UDFR #2

draft a GenreFacetType ontology collecting most of PRONOM, NFO, UDFR #2

Conversation

barmintor commented Apr 10, 2015

awoods commented Apr 10, 2015

acoburn commented Apr 12, 2015

barmintor commented Apr 12, 2015

acoburn commented Apr 13, 2015

awoods commented Apr 13, 2015

acoburn commented Apr 13, 2015

barmintor commented Apr 14, 2015

Choose a reason for hiding this comment

Choose a reason for hiding this comment

azaroth42 commented Apr 17, 2015

acoburn commented Apr 17, 2015

azaroth42 commented Apr 17, 2015

escowles commented Apr 17, 2015

mjgiarlo commented Apr 17, 2015

awoods commented Apr 17, 2015

acoburn commented Apr 17, 2015

awoods commented Apr 17, 2015

ruebot commented Jun 4, 2015

escowles commented Jun 4, 2015

awoods commented Jun 5, 2015

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ruebot commented Jul 17, 2015

escowles commented Jul 24, 2015

barmintor commented Jul 24, 2015

escowles commented Jul 24, 2015

acoburn commented Sep 9, 2015

ruebot commented Sep 9, 2015

awoods commented Sep 9, 2015

azaroth42 commented Sep 9, 2015

ruebot commented Sep 9, 2015

escowles commented Sep 9, 2015

barmintor commented Sep 9, 2015

awoods commented Sep 9, 2015

ruebot commented Sep 9, 2015

awoods commented Sep 9, 2015

ruebot commented Sep 9, 2015

awoods commented Sep 9, 2015

ruebot commented Sep 9, 2015

ruebot commented Sep 9, 2015