RE: How to describe things in an archive collection?

Hello, 

So far I have seen a lot the term "collection" used in the context of Archives. I have worked in the past with archivists on the distinction between collection and hierarchies [1] . Collections are qualified by a membership relation between the collection and the items it includes when hierarchies are qualified by a part-whole relationships that are more granular than the membership relation. On the Web you would probably need to know that a particular item can't be understood on its own because it structurally or semantically relates to another item.  I don't know if this distinction would matter for the work on the extension of schema.org. 

Best, 
Valentine 
 
[1] http://pro.europeana.eu/files/Europeana_Professional/EuropeanaTech/EuropeanaTech_taskforces/Hierarchical_objects//TF%20report%20V1.0%20PDF.pdf 


________________________________________
From: Giovanni Michetti [michetti@mail.ubc.ca]
Sent: 07 August 2015 03:43
To: Young,Jeff (OR)
Cc: Richard Wallis; public-architypes
Subject: Re: How to describe things in an archive collection?

Jeff,

I guess I agree with you. As far as I see, we are in the first stage of
the process you decribe, i.e., looking for language-tagged *labels* for
"???". Please note that I wasn't pushing for a certain label, I just
presented some critical considerations and wished we could talk about it
later. Now it's more important to model the class than focus on the
label. So 'Foo' is perfectly fine to me at this stage. 'Foo' may even be
fine in the future, if the definition is posed properly.

However, if 'Archive' is used in place of 'Archives', or
'ArchivalCollection' seems to me a better label than
'ArchiveCollection', I will use all my arguments to convince you, and I
expect you to do the same. I guess this is how these processes work.

Giovanni



On 2015-08-07 3:10 AM, Young,Jeff (OR) wrote:
> Giovanni,
>
> The thing I like about schema.org is that schema:??? is capable of utility that only starts with our assumptions about the meaning of "???".
>
> Note: http://martinfowler.com/bliki/TwoHardThings.html
>
> When our use cases stumble on the name "???", we can (or at least should) look for language-tagged *labels* for "???". This will be a one-to-many relationship. Likewise, when that fails to clarify things we can look to language-tagged *descriptions* of "???". When that fail, we can look to domain and range assignments to "???". When that fails, we can look at examples. Etc.
>
> Schema.org's success seems to be based on the 80/20 rule of "common sense" at all these levels. Somehow magic seems to happen when applications do that.
>
> Jeff
>
>> On Aug 6, 2015, at 7:54 PM, Giovanni Michetti <michetti@mail.ubc.ca> wrote:
>>
>> Hi Richard,
>>
>> I understand this is the very early stage of the process, when the main areas of work are identified, so I didn't highlight some other issues that we may discuss later when we go back to each thread.
>>
>> Since you asked, I don't think that ArchiveCollection is a good label, not because of the singular/plural issue, but because "Archive Collection" means a collection of archives, just like a book collection is a collection of books. I don't think this is what we want to represent here. Archives are not the members of the collection. Archives are the collection. We just want to say that the collection has an archival nature.
>>
>> In short, in my opinion 'ArchivalCollection' is the proper term.
>>
>> Also, please be aware that subsuming an archival collection under CreativeWork is a problem I would like to discuss later. Unfortunately the definition of CreativeWork is a tautology ("The most generic kind of creative work"), so technically anything can be put in there. However, I couldn't find under CreativeWork any class that suits a typical administrative records, i.e. the most common archival object in public archives.
>>
>> I think a new class is needed. If this is agreed, we may think whether it has to be developed under CreativeWorks, and this may have consequences on the Archiv*Collection too.
>>
>> However, I guess we'll discuss these issues when we focus on the specific threads.
>>
>> Giovanni
>>
>>
>>
>>> On 2015-08-06 1:44 PM, Richard Wallis wrote:
>>> [...]
>>> Thanks for the clarification about 'archives'.  Would this mean that the
>>> suggested new Type ArchiveCollection should be changed to
>>> ArchivesCollection?
>>>
>>> ~Richard.
>>>
>>>
>>> Richard Wallis
>>> Founder, Data Liberate
>>> http://dataliberate.com
>>> Linkedin: http://www.linkedin.com/in/richardwallis
>>> Twitter: @rjw
>>>
>>> On 6 August 2015 at 03:28, Giovanni Michetti <michetti@mail.ubc.ca
>>> <mailto:michetti@mail.ubc.ca>> wrote:
>>>
>>>     Hello Richard,
>>>
>>>     I think that the most common standards for archival description
>>>     offer quite a long list of the descriptive properties that would be
>>>     of use in describing an item's place and role in an archives. In
>>>     particular, if we take a look at the most famous one, ISAD(G), and
>>>     avoid focusing on the specific information elements, we can
>>>     recognize some aspects that are fundamental to any archival description:
>>>
>>>     - identification (identifier, title, location, ...)
>>>     - physical characteristics (extent, dimensions, medium, ...)
>>>     - creation (info on creator, dates and circumstances of creation)
>>>     - custody (info on custodial history, including curation actions)
>>>     - content (scope, content, keywords...)
>>>     - access & use (info on conditions governing access & use)
>>>     - related materials (info on relationships with any relevant objects
>>>     within or outside the collection, including the collection itself)
>>>     - description control (info on the description process, including
>>>     authors, dates, methods ...)
>>>
>>>     I would like to point out that:
>>>
>>>     1) such a rough list comes from a quick review of ISAD only;
>>>
>>>     2) these broad categories result in a set of properties whose
>>>     granularity has to be defined and shared by the group. I think it
>>>     may be useful though to proceed in a top-down approach, identifying
>>>     the areas of interest, and then specifying the properties;
>>>
>>>     3) information on creation and custody, as well as on the
>>>     description control, is fundamental to trace the Provenance of an
>>>     archival object, which is a core issue in the research agenda of
>>>     many scientific initiatives around the world. I would put a lot of
>>>     attention in these areas;
>>>
>>>     4) with reference to your tentative list of properties, they all fit
>>>     the above list.
>>>
>>>
>>>     On a separate note, I would like to add a quick comment on archives
>>>     vs archive: 'archives' is not plural, it is a technical term used in
>>>     the archival domain to designate a specific object--an archives.
>>>     Just like a fonds is a fonds--this is not about plural. I guess no
>>>     one would ever create a class called "ArchivalFond", just like no
>>>     one would create a class called "New" to refer to news. In short, I
>>>     don't think the argument on respecting the Schema.org convention
>>>     (i.e., use singular for the names of classes) holds here. Also,
>>>     please note that the ICA Glossary does not have any "archive"--only
>>>     archives. Anyway, I'm sure there will be time for further discussion
>>>     later.
>>>
>>>
>>>     Giovanni
>>>
>>>     Giovannni Michetti
>>>     InterPARES Trust
>>>
>>>
>>>
>>>     On 2015-08-05 12:03 PM, Richard Wallis wrote:
>>>
>>>         In other threads we have been discussing how to describe an
>>>         Archive as
>>>         an Organization/LocalBusiness
>>>         <http://lists.w3.org/Archives/Public/public-architypes/2015Jul/0002.html>
>>>         when
>>>         appropriate, and how to describe an ArchiveCollection
>>>         <http://lists.w3.org/Archives/Public/public-architypes/2015Jul/0008.html>.
>>>         Now I think it is time to add one more area to our attention -
>>>         how to
>>>         describe the physical/digital things that we find within an archive
>>>         collection.
>>>
>>>         In archives we find all types of things from creative works such as
>>>         books, letters, artworks, videos, web pages etc., to furniture,
>>>         personal
>>>         items, vehicles, fossils, rocks and of course the favourite box of
>>>         things yet to be identified.  From what I understand there are
>>>         certain
>>>         common categories of things such as physical creative works, digital
>>>         creative works, physical containers of things identified or not,
>>>         but it
>>>         would be far too limiting to build our recommendations around these.
>>>         The result is that we need to be able to describe anything that
>>>         could be
>>>         found in an archive which means /anything!/.
>>>
>>>         Fortunately in our world all these things have one aspect in
>>>         common -
>>>         they are in an archive.
>>>
>>>         If we can establish a set of descriptive properties that would
>>>         be of use
>>>         in describing an item's place and role in an archive, we can
>>>         then look
>>>         to some, schema.org <http://schema.org> <http://schema.org>,
>>>         techniques to apply them
>>>         alongside other properties that are already available in the Schema
>>>         vocabulary.
>>>
>>>         Properties that come to mind include:
>>>
>>>              isPartOf - a reference to the collection a thing is in
>>>              condition - state of preservation of an item
>>>              containedIn - the box or digital file containing the item
>>>              curatedBy
>>>              curationDate
>>>              CurationEvent - possibly a better way to describe a
>>>         curation event -
>>>              linking where when and by who
>>>              location - of item, not necessarily the collection location
>>>
>>>
>>>         We could look to already existent standards, CIDOC-CRM for
>>>         example, as a
>>>         source of inspiration.
>>>
>>>         So, over to you for suggestions.  Once we have assemble a few by
>>>         email
>>>         discussion, we can create a page in the Wiki to capture them and
>>>         become
>>>         the basis for the core of our proposals.
>>>
>>>         ~Richard.
>>>
>>>
>>>         Richard Wallis
>>>         Founder, Data Liberate
>>>         http://dataliberate.com
>>>         Linkedin: http://www.linkedin.com/in/richardwallis
>>>         Twitter: @rjw
>>>
>>>
>>
>>
>
>



Received on Friday, 7 August 2015 15:03:19 UTC