you're reading...

Collections and Connections

Create Once, Publish Everywhere – Reusing museum collection content

Creating and managing information relating to museum collections was historically a mostly internal process. Museums are now finding new opportunities to share information about their collections and need to find ways to encourage and streamline the reuse of this content. This post is the result of my presentation at the Museums Australia conference in May. You can view the shorter slide deck on reusing collections information online on Slideshare.

C.O.P.E. – Create Once, Publish Everywhere

The concept of “Create Once, Publish Everywhere” was devised as part of National Public Radio’s content management strategy. They wanted to change the way they recorded and shared information so that their news content could be used as much and as easily as possible. Aggregation of collection content The key concept is separating the creation and storage of digital content from how it will be delivered or presented. There are lessons here for the museum sector, as we manage similarly rich and complex data and are increasingly expected to present that information in multiple locations and formats.

C.O.P.E. relies on identifying small nuggets of data that can be captured in a consistent format. For museums, one important piece is the museum catalogue record which holds key information about each object in their collection. It’s information that museums often already have in great detail and for which there is demand from the outside world.

The C.O.P.E. concept is only workable if content is stored in a structured way. A collections management system can help museums store cataloguing data as individual pieces of information which can be exported or published to other systems for a range of uses. Other museum digital content might be stored in additional systems, such a web content management system for blog posts or calendar events.

A range of museum data standards, include the UK’s SPECTRUM standard, help guide how cataloguing information is recorded. A typical example of museum collection content being reused is collection record aggregation sites. The  collection records are initially published to the museum’s website. The museum then partners with an aggregation site (like Culture Grid in the UK or Trove in Australia). These sites represent many organisations, and the original collection records are copied and presented in a different standard format.

Multiple channels and devices

Responsive DesignMuseums need ways to create content that can be published to multiple locations without excessive manual re-work. Access to web content is from an increasingly wide range of devices. Responsive design websites dynamically change their layout according to the device that is accessing the page. Here we see examples of the Australian Museum’s topic page about Spiders in different screen sizes. This is one example of how the same content might be presented in different ways.

One organisation looking at how they reuse content is NZ’s Ministry of Culture and Heritage (read Matthew Oliver’s post on multi-channel publishing). They recognise that each channel they publish to may have a different focus and a different audience, but they are working to create nuggets of data that can be reused. For example, audio clips that have been created are used in their MyTours app and are separately available through iTunes. They have also taken the biography web pages on a selection of famous New Zealanders and have republished these as eBooks.

Sharing basic object details

The most basic information that can be shared about an object is a short name or title for what it is, a description, a media file, and a web address for where the authoritative record of the object can be accessed. This already opens up options for publishing basic details of the object to multiple destinations, possibly with links back to a more detailed public record. Curatorial Poetry

One example is slightly tongue-in-cheek “Curatorial Poetry” Tumblr based on Cooper-Hewitt National Design Museum’s collection. An automated program posts the descriptions of random objects every two hours, highlighting the often poetic style of the original curatorial descriptions. Each object links back to the record in Cooper-Hewitt’s online collection website. Having the object records published to the web on permanent pages provides other opportunities for reusing this content. The National Library in New Zealand came up with the idea of a Twitter #tbreaktweets hashtag. Each day they post a link through to an item from the collection, providing a interesting break in someone else’s day. Social media is a great way to share links to topical collection records.

The original concept of C.O.P.E. was to store structured data so that it could be output to multiple presentation layers via programming interfaces. However, the C.O.P.E. concept is useful for museums not just for automated reuse of content. Whenever content is created is it worth considering whether the style and structure allows it to be reused in the future, both automatically and manually. For museums, one of the key standards that allows sharing of data is the Dublin Core standard. This includes the basic fields needed to describe resources including physical object in museum collection. Standards like this make aggregation possible – the copying of data from many contributor websites into a single site.

Problems with C.O.P.E.

The basic description of the object is easier to share. The more detailed catalogue record is where things get difficult. A lot of museum cataloguing data was created for use by staff, not for different audiences on the Internet.  The content required for different devices and audiences may not be the same and museums often don’t have the resources to create variations of collection content for different audiences.

One challenge with sharing collections data is that the information is both complex and diverse. Natural history catalogue records for example don’t fit well into the Dublin Core standard, so records can lose important data when copied to sites relying on this standard. The more contributors there are, the more likely that the ‘squishing’ of the data into a common standard will result in a loss of some of the quality or detail of the original data.

There are many overlapping standard terminology lists, plus museum-specific lists, for describing objects. This limits options for cross-collection searching. Object tagging, mapping terms between thesaurii, and building search interfaces that understand synonyms for terms all help, but there is still a lot of work to do.

Museums also need to consider the rights to the content. Published data should be accompanied by a rights statement so the users know what they are allowed to do with the content but the rights statement can be lost as the data gets multiple steps away from the source.

Sharing contextual descriptions

Managing detailed descriptionsOne area where C.O.P.E. has particular merit for museums is the managing of contextual descriptions. For many museums rich higher level descriptions including stories about the collection, exhibition wall labels and collection level descriptions, are not stored and managed to allow them to be accessed easily in the future. Huge effort goes into creating these – we need to be managing these so that there are more opportunities to reuse and share these descriptions. These descriptions are usually written with an external audience in mind, so are better suited to general publishing than tombstone collection records. Collections management systems are evolving to capture and share this data as well as traditional catalogue records.

Improving how data is shared

More detailed alternatives to the Dublin Core standard are now available. CIDOC’s Component Reference Model (CRM) describes the concepts and relationships used in cultural heritage documentation, not just the basic object information. The British Museum has made their data available using the CRM standard.

The Lightweight Information Describing Objects standard builds upon several existing museum standards including CIDOC’s CRM. Europeana is one of the biggest projects to get behind the LIDO standard. This is a drive to pull in more detailed, well-structured records into the Europeana database. It covers the full range of common collection types, including art, social history and natural history. Both CRM and LIDO are more complex standards than Dublin Core and will take time to gain widespread adoption. Users need to understand their importance, systems need to add these as output options, and aggregation sites need to be able to import data in the new format.

Linked data is another approach for extending the reach of collections data. Linked data is a set of best practices for publishing and connecting structured data on the web. Data can still live on your own website but the data includes links between individual pieces of information, both within your own records and to other sites. While this may fall outside of C.O.P.E, it is worthwhile looking at how content museums create can be linked up within the cultural sector and to larger projects like Wikipedia. For example, the American Art Museum is working to reconcile their artist records with Wikipedia artist pages. It involves manual work to verify possible matches but creates a link to Wikipedia’s often more detailed artist biographical information.

Aggregation sites are also working to reconcile disparate data. DigitalNZ maps a huge variety of format, object type and usage terms from contributors into a simple set of options for the end users. For example, various categories of art are all flagged so that they all appear under the Artwork filter option.

Points to consider when applying the C.O.P.E concept

You should consider specific cases where you want your museum records to be reused. e.g. In Google results or in an aggregation site. Deciding on the optimal way for the content to appear in another place takes time and may require compromises. This might involve looking at the data a service can import (like Culture Grid) and reviewing how you have used particular fields in your database in the past. It might involve changing your own museum website to embed the metadata in a specific way on collection record pages so that certain channels better present links back to this content. e.g. Using Twitter metadata cards.

However, the demand for content to be in many places won’t go away.We have better tools now for studying what content is getting the most interest. Use services like Google Analytics to review demand for existing content. You can also use these statistics to help you choose content that you might republish because it has new relevance to events happening right now.

Creating new content takes a lot of time and care. We need to find ways to get the most value out of that effort. Communities help fund their museums. Providing easy access for these communities to find, use and share collections information is now an essential service and the C.O.P.E. concept provides some strategies for improving this service.


6 Responses to “Create Once, Publish Everywhere – Reusing museum collection content”

  1. Great article Paul. I’ve been doing some similar research/work in this area – specifically in regards to developing content in a higher education setting. My experience with COPE is that it is a wonderful model, but quite difficult to apply to complex institutions. NPR were lucky that this was part of their overall content strategy so were able to move forward as an organisation. My experience is that having an actual content strategy even considered would a luxury! And working somewhere like a university or museum there is a need (and benefit) in supporting diversity rather than trying to make everything conform to single system/standard.

    My work has led me to develop a framework that provides the COPE functionality – but is an attempt to create something far more adaptable so that it has a much wider application. I’ve detailed the concepts and ideas on my blog http://timklapdor.wordpress.com/2013/06/04/adaptive-digital-publishing-current-work-ideas/

    What I’ve tried to do is compensate for the fact that standard formated text is quite easily adapted – what sucks is the more complex items, the media. To overcome this I’ve conceptualised the idea of the Adaptive Media Element (AME) which acts as a hook to which logic can be applied. I could see how this could apply to your work – in that an AME could be developed to query individual data sources, rather than trying to pull them all into a single source. I go into more detail on the blog, though not specifically for your application – but I’d love to hear some feedback!

    Posted by Tim Klapdor | June 8, 2013, 5:56 pm
    • Thanks for taking the time to respond Tim. My conclusion was similar to yours – there’s a need to streamline the publishing content in more than one place or format, but content in the museum and education sector often doesn’t lend itself to this.

      Even text has its challenges – the same narrative may need a different length or tone for different devices or audiences. There are some easier wins, like sharing museum collection records with aggregation sites. In this case the process for copying the content between the platforms might not be perfect, but the benefit in reaching a larger audience makes it well worth considering. We make use of XSLT to transform structured museum collection data into standard formats (e.g. Dublin Core).

      On the software side I think the solution will be one of evolution rather than revolution. Two examples of software packages that our company uses are WordPress and Author IT. The way that WordPress handles media embedded in posts has come a long way. Media files exist as a complex entries in the media library, each with its own caption / alt text / formatted description / links / embed codes. Perhaps most importantly, WordPress is taking care of the creation of multiple derivatives, so there are a range of media file sizes automatically available for different circumstances. Author IT is the package we use to write documentation and it has some concepts similar to your Adaptive Media Element idea. Nuggets of documentation, including separate text and media elements, are created in a presentation agnostic form. Author IT then provides templates for various output formats. e.g. PDF / Word / Windows Help files / Web pages. Multiple template styles can be created and altered to publish subsets of the documentation for use on different devices.

      Some of the content problems are harder to solve. Jason Grigsby gives a good example of images which need to be swapped out with a custom cropped version for a particular style of presentation (‘art direction’). http://blog.cloudfour.com/8-guidelines-and-1-rule-for-responsive-images/

      In many cases individual presentation formats really do require custom content, particularly when the audience’s needs are different. As much as anything, COPE is a reminder to content creators that we need better practices for archiving content. In the museum world, carefully crafted content is often lost over time. Even if the content can not be used as is / where is, we need to be able to bring it back as a reference point during future work.

      Posted by Paul Rowe | June 9, 2013, 5:00 pm
      • Your post has given me another perspective which is great! On the software front I think WordPress does a lot of things right. I think the idea of developing content *for* a database, rather than just going in one is an interesting concept. The AME is a vision of that, a flexible tool that can be developed and customised depending on the purpose or requirements. To me it is about creating a concept that is applicable to many context, which is something that COPE unfortunately won’t do. I’ve been following Jason’s work too, and it’s helped shape my ideas, and I think what he’s looking for in images could be done via an AME. You can insert all the different images as seperate objects within the AME and then dynamically pull out the right one depending on the logic you apply – for mobile, app, post location etc. I came across some other interesting articles over the weekend – working on a blog post at the moment – but it sounds like a shared problem across many industries!

        Posted by Tim Klapdor | June 11, 2013, 1:04 am


  1. […] also coincides with a referral to a post via @CathStyles by Paul Rowe on the Create Once Publish Everywhere (COPE) concept. Paul gives a really great overview of COPE in a really fair and balanced view. He also offered me […]

  2. […] problems…. — Trusted advice on "How-to…..The Increasing Momentum of Content StrategyCreate Once, Publish Everywhere – Reusing museum collection content /*Is […]

Post a Comment

Recent Tweets (@armchair_caver)