DepositMO extensions for SWORDv2 clients

Previous posts have described two tools developed in DepositMO – the Word Add-in, and the Watch Folder – for improved repository deposit of in-progress works. These tools are freely downloadable for use now, but if you were to try and use them to add content to your local repository it is likely a key element will be missing. Your target repository has to support SWORDv2. Most support SWORD, but not yet the latest version (v2). They could do, however, because SWORDv2 is available with current versions of EPrints (3.3) and DSpace (1.8). In addition, DepositMO has extended SWORDv2 to add significant features required by the deposit tools produced by this project and which we believe will support further applications, and these extensions are also supported in current repository releases. Forthcoming posts here will look in more detail at how EPrints and DSpace each implement SWORDv2 with DepositMO extensions. First, in this post we look at the reasons for these extensions.

Content Management Systems (CMSs) should allow a number of basic functions to be carried out by a user, and digital repositories are no exception. Historically, such repositories have provided pretty web-based, human-focused interfaces through which users can manually perform the tasks of creating and publishing resources. With the development of SWORD, the creation and management of resources became something which, in part, could be carried out using an automated, machine-based client.

In SWORD version 1, this process was focused on making the operation of publishing a completed resource simpler. Further interaction (including retrieval of the same object) was not defined as a requirement. Thus this first specification, while useful, fell short of being a complete CMS system.

It was the intention of SWORDv2 and the DepositMO project to address these shortcomings. In addition, a DepositMO implementation of SWORDv2, the DepositMO profile, extends the functionality of SWORDv2 using a number of the as yet unused parts of the protocols SWORDv2 is based on – AtomPub and HTTP.

This post describes the features SWORDv2 provides to turn repositories into full CMSs, and details two extensions implemented in the DepositMO profile. While the DepositMO profile is a technical document, here we show how the extensions support two particular use cases, and how the DepositMO-developed deposit clients exploit SWORDv2 and the extensions.

Extending repositories as full content management systems

Figure 1 shows some of the basic operations that need to be supported by a digital repository-based CMS system.

Discover: The ability to provide simple description services which allow the client to choose the best method for deposit. Additionally, a client should be able to discover existing resources in the repository, thus skipping Create directly to Update.

Create: Provides a way for a client to create a resource  – a publication object – in the repository. It is this object that is then populated with relevant content.

Update: By making the repository a workspace, content should also be changeable.

Publish: Provide users with a way of describing an object as being suitable for publishing, making it public.

Retrieve: A critical aspect of any CMS system, a user must be able to retrieve their own content.

Delete: Similar to update, a user may be able to delete any resources they own. This is not always the case in digital repositories due to the ties of such systems with the classic publication model (a physical printed copy is hard to ‘delete’).

Use cases for DepositMO extensions to SWORDv2

As well as playing a key role in developing SWORDv2, the DepositMO specification details two key extensions to the SWORDv2 specification:

1. A user/client must be able to retrieve a list of their content regardless of the status of that content (edit, review, published, etc.).
2. A client must be able to obtain basic information about individual objects such that distributed copies can be kept synchronised and up to date.

To demonstrate the importance of these aspects consider the following two use cases:

• A user creates a formal paper which on completion is deposited and published in a repository for a conference. At the conference the author gives a presentation to accompany this submission. Logically this presentation should also accompany the publication already in the repository. Using the DepositMO discovery specification, the presentation software can list all the resources already in the repository and the user can “attach” the presentation as a complementary resource.
• A user is working on a publication for a conference, but wants to work both at home and at work without having to carry around a laptop or a USB memory stick, which will get lost just like half  a pair of socks. By using the repository as a workspace, two clients can maintain synchronisation of the same resource. To save download and comparing the document, which might be of substantial size, a DepositMO extension can be used to check the state of the object without having to download it.

As part of the DepositMO project two clients were produced – a Microsoft Word Add-in client, and a Watch Folder drag-and-drop deposit tool that works with file managers on standard PCs – to demonstrate the principles of these use cases. Although neither the most complete or prettiest clients, they do demonstrate the power of the two combined DepositMO extensions to the SWORDv2 specifications.

DropBox-style repository resource control – Watch Folder client

DropBox is easy. You drag stuff in, it is shared, simple. The barrier to entry for such services is low, meaning that people use them for convenience and uptake is high.

During the DepositMO project, questions were asked about how digital repositories, a trusted authority, could keep up with such technologies. This led to the development within DepositMO of a DropBox-style client for a digital repository.

Called “Watch Folder”, this client enables a user to create a folder, representing their resource, in a standard file manager on a personal computer, and then to add objects that are part of this resource to this folder.

As an example, the repository could contain a collection of images (the objects) relating to a single art exhibition (the resource, folder). As each object is copied into the folder it is simultaneously uploaded to the repository. View and edit links also appear as files on the desktop, connecting the user directly with the corresponding object in the digital repository.

As with DropBox, if an object is deleted or updated via either the repository or desktop interface then this change is instantly reflected in both locations. This functionality is not possible without the DepositMO extensions to the SWORDv2 specification.

The Watch Folder client demonstrates the ideal simplicity of a SWORDv2/DepositMO client to the user.

Windows application integration – Word add-in client

The second client produced as part of the DepositMO project integrates Microsoft Word with a digital repository. Although not a new concept – a similar client was produced by Microsoft to conform with the SWORDv1 specification – the DepositMO version integrates the complete CMS functionality as detailed by the SWORDv2 and DepositMO specifications.

From the start the Windows development separated the low-level driver from application integration. In this way the driver connects the application to the SWORDv2/DepositMO server via an easy-to-use API. This has the beneficial side-effect that the same driver can be used by many applications, saving substantial amounts of time when developing further clients.

The Microsoft Word add-in adds a new tab to the menu “ribbon” in Office 2010. Using this tab a user can set up their credentials with the repository, which can then be used to create and edit resources. Again, the user remains connected to their content whether on the local machine or in the repository.

Future of the DepositMO profile of SWORDv2

The DepositMO profile complements the SWORDv2 specification in order to provide more advanced functionality to deposit clients. For now the DepositMO specification remains an extension to SWORDv2, albeit one that is already integrated into the core of the latest EPrints and DSpace softwares, as will be revealed in the next two posts here. It would not be a huge leap to include the two simple DepositMO extensions in a future SWORDv2 specification, as without them many simple operations are just not possible.