Jamie,
I think one question we can usefully ask is whether our current
collections management systems are actually the best starting point for
a move in a Linked Data direction. The orthodox approach has always
been that 'publishing a collection' means 'putting the collections
database online', but this hasn't always worked that well even for
old-school HTML pages. We've all seen those collections search pages
containing lists of minimal tombstone records ...
We should maybe take a step back and ask "what information do we want to
share on the Web?". I'm sure the answer wouldn't be "all of our
collections, and nothing else". Nor, I must say, would it be "lots of
Linked Data". I see Linked Data URLs as a means to end: a common frame
of reference in a shared information space, full of stories. We need to
work out what concepts we share an interest in, and create Linked Data
resources for them. They will be people, places, historical events,
classes of object, /not /the individual objects in our collections.
Richard
On 2016-02-03 12:46 PM, Unwin Jamie wrote:
> I think part of the issue is legacy data and the lack of first mover advantage.
>
> Starting afresh it’s (fairly) easy to ensure that every term, category, person or attribute is tied to a linked data URI (although which sets of URI’s and how similar sets are interlinked is still an open question).
>
> Hopefully collection managment systems will start to incorporate URIs and Linked Data seamlessly into their products/interfaces, so it should start to happen by default (ie. if I pick a term from a Thesaurus my collection system stores both the URI and a String and most orgs will go with the default sources unless they have good reason to go elsewhere). Likewise an interface to a person authority should ideally / automatically check the public space for common URIs before allowing me to create my own record (even if I’m just augment the base URI with my own biography/data).
>
> Where it becomes tricky is when you have hundreds of thousands of existing record that may take years to tidy up by hand, which is the situation most large organisation find themselves in.
>
> While I expect extremely difficult to perfect, I can’t help but thinking that a public API/service which could weight a match between a free text field and a URI based on context (the other fields in the records, the relationships or the legacy thesaurus a term was possibly taken from) and return a linked data URI for each attribute / relationship / person with a degree of certainty for each (ie. a percentage weighting), would be a god send (it’s probaly god like it it’s complexity too!)
>
> Obviously these URI’s couldn’t be blindly accepted, but anything with a weighting of say over 80% could be more quickly checked by hand than by going back over your record and manual finding/adding a URL for each term.
>
> Not sure if there are exiting projects out there that are already working on this sort of thing, I expect there are and I am just not aware of them.
>
> Jamie
>
> From: Museums Computer Group on behalf of Richard Light
> Reply-To: Museums Computer Group
> Date: Wednesday, 3 February 2016 at 11:42
> To: "[log in to unmask]<mailto:[log in to unmask]>"
> Subject: Open cultural data in 2016
>
> Denis,
>
> I think the rate of progress towards open cultural data is fairly
> glacial here in the U.K. See my provocation to UKMW15 [1] for a summary
> of some relevant issues.
>
> As you imply, one major issue is centralization vs. silos. I support
> the idea that individual institutions should take responsibility for
> publishing their own resources, but in a way which is interoperable with
> the rest of the community.
>
> A central initiative which is worthy of note is the Research and
> Education Space [2]. This offers a platform for cultural heritage
> Linked Data resources, which a community of developers is now eager to
> get to work on, developing educational applications. All it needs is
> lots of cultural heritage Linked Data ...
>
> One problem is: what does 'interoperable' look like? I have been
> advocating the use of the CIDOC CRM [3] to express statements about
> collections objects in an interoperable manner, but there is no
> immediate answer to the problem that all actual collections data
> currently held by MLAs is expressed as string values, not URLs. Most of
> the software systems MLAs are using offer them no help to move towards a
> Linked Data publication strategy. And (with a few exceptions, such as
> the Getty vocabularies, Geonames, ...) there are no common frameworks we
> can turn to as a source of such URLs.
>
> I would be delighted to be told I'm wrong, but my impression is that
> most UK institutions don't really 'get' the open/linked data idea, and
> where they do, they have no idea how to go about supporting or
> implementing it.
>
> Best wishes,
>
> Richard
>
> [1]
> http://www.slideshare.net/RichardLight/museums-home-of-unlinked-data-57824586
> [2] https://bbcarchdev.github.io/res/
> [3] http://cidoc-crm.org/
> P.S. I thought MCG were putting all these talks up after the meeting.
> Clearly not ...
>
> On 2016-02-03 12:34 AM, Denis Nazarov wrote:
>> Hello MCG,
>>
>> I am one of the engineers working on Mediachain.
>>
>> Mediachain
>> <https://medium.com/mine-labs/introducing-mediachain-a696f8fd2035#.414x0afc4>
>> is an open source, blockchain based metadata registry that enables
>> institutions and developers to easily publish, use, link and extend media
>> datasets, while maintaining attribution for contributors and creative works.
>>
>> Instead of releasing metadata as static data dumps
>> <https://github.com/NYPL-publicdomain/data-and-utilities> or requiring
>> developers to use disparate siloed APIs
>> <http://museum-api.pbworks.com/w/page/21933420/Museum%C2%A0APIs>,
>> Mediachain functions as a shared data layer where institutions can
>> verifiably publish data. Once ingested into the datastore, developers can
>> easily reuse and extend it without obscuring its source.
>>
>> The goal of Mediachain is to lower the barrier to publishing and consuming
>> open data, while preserving data attribution. Datasets are no longer
>> siloed, but integrated with one another so each new entrant has an
>> opportunity to securely enrich the existing corpus.
>>
>> I was inspired by Mia Ridge's awesome article from 2013
>> <http://www.museum-id.com/idea-detail.asp?id=387> to reach out to this
>> group to ask:
>>
>> - What is the state of open cultural data three years later in 2016?
>>
>> - For technologists with experience at GLAMs, how can current open data
>> strategies and practices be improved to enable richer collaboration between
>> institutions and a digital audience?
>>
>> We are looking to pilot Mediachain with a few institutions and their open
>> data and I'd love to talk to anyone interested. Here is more detail on the
>> Mediachain protocol
>> <https://medium.com/mine-labs/mediachain-483f49cbe37a#.vdksi4onb>.
>>
>> -Denis
>>
>> ****************************************************************
>> website: http://museumscomputergroup.org.uk/
>> Twitter: http://www.twitter.com/ukmcg
>> Facebook: http://www.facebook.com/museumscomputergroup
>> [un]subscribe: http://museumscomputergroup.org.uk/email-list/
>> ****************************************************************
>>
> --
> *Richard Light*
>
> ****************************************************************
> website: http://museumscomputergroup.org.uk/
> Twitter: http://www.twitter.com/ukmcg
> Facebook: http://www.facebook.com/museumscomputergroup
> [un]subscribe: http://museumscomputergroup.org.uk/email-list/
> ****************************************************************
>
>
> ________________________________
>
> This email has been scanned for email related threats and delivered safely by Mimecast.
>
> ________________________________
> --------------------------------------------------------------------------------------------
> This e-mail and attachments are intended for the named addressee only and are confidential.
>
> If you have received this e-mail in error please notify the sender immediately, delete the
> message from your computer system and destroy any copies.
>
> Any views expressed in this message are those of the individual sender and may not reflect
> the views of the Science Museum Group
>
> This email has been scanned for email related threats and delivered safely by Mimecast.
> --------------------------------------------------------------------------------------------
>
> ****************************************************************
> website: http://museumscomputergroup.org.uk/
> Twitter: http://www.twitter.com/ukmcg
> Facebook: http://www.facebook.com/museumscomputergroup
> [un]subscribe: http://museumscomputergroup.org.uk/email-list/
> ****************************************************************
--
*Richard Light*
****************************************************************
website: http://museumscomputergroup.org.uk/
Twitter: http://www.twitter.com/ukmcg
Facebook: http://www.facebook.com/museumscomputergroup
[un]subscribe: http://museumscomputergroup.org.uk/email-list/
****************************************************************
