DSpace is a service of the MIT Libraries, built using a software platform jointly developed by Hewlett-Packard Company and MIT. The MIT Libraries use DSpace as a long-term storage system for digital documents; it captures, preserves, and provides access to the research output of MIT researchers, centers, and labs.
This document gives a brief overview of how DSpace works. It is oriented to help anyone get started submitting to DSpace quickly. Because DSpace is not your run-of-the-mill content management system, we hope you'll review the concepts presented here, and that doing so will help you prepare and submit your files more smoothly.
On to the overview! First, two meta-points:
You will need to be logged in to DSpace before you can submit; MIT users must use a web certificate to authenticate. Most collections will also require specific authorization for you before you can submit to it. Contact the organizers of the collection to which you wish to submit if you are unsure whether you have permission.
After starting a submission, you will be led through a seven-step process: these include some basic questions about your materials first, then several screens where you can describe your materials (see metadata below), then file uploads, a verification screen (where you can even compare file checksums if you wish), and a license granting screen. Each of these steps has more detailed documentation available as links on each screen.
After you finish, the folks administering the collection to which you are submitting may review, edit, or approve your submission according to the collection's policies. In other words, your materials might _not_ go directly into the archive. Check with the organizers of your collection for their policy if you want to know how it will work.
DSpace content is organized into communities, with logical groups of materials organized into collections within individual communities. The atomic unit of content in DSpace is an "item", a.k.a. one "thing", whether that's a single file article, or a single dataset made of several files, or several file format versions of the same image. Take a look at the communities overview, and browse some DSpace items to get a feel for how it works. If you have questions about whether your materials should be one or several items, or even a collection, get in touch with us and we'll be happy to consult with you.
We librarians are big on metadata. Hopefully after thinking about it a bit you will be too. In collections as big as libraries, of which DSpace is just a part, being able to find a single item or a set of relevant items for a given user in some context of information need is very dependent on the quality and consistency of descriptive information available for querying. Some description can be done mechanically: file sizes, checksums, and full-text indexing, for instance, are critical elements, so we've built those in. But we also need a higher order of human-made description, to determine things like titles, authors, unique identifiers, and abstracts. If DSpace handled just one genre of content (say, just articles), we could standardize formats and templates to make this easier. DSpace handles many genres, and many formats, however, so we need your help in providing this description.
DSpace uses a qualified version of Dublin Core as a descriptive lingua franca across all content. Some communities or collections may also have tailored metadata available (such as MARC records for book collections, or FGDC records for geographic datasets). But even where that's available for some items, we crosswalk more detailed metadata records into our Dublin Core vocabulary to ensure a common layer of descriptive specificity for browsing and searching across everything. The "describe" pages in the submission interface map the values you enter to our vocabulary: what you enter will be what others use to find your content in the future. So please take a moment to read the explanations of each field, check the help screens, and read our vocabulary documentation if you're interested. And don't hesitate to contact us with questions, comments, or concerns!
Over time, items stored in DSpace will be preserved using a combination of time-honored techniques for data management and best practices for digital preservation. As for specific file formats, however, obsolescence and proprietary formats make it impossible to guarantee the exact same level of preservation support services for every file.
We have written a format support policy that we believe balances the needs of most researchers with the long-term reality and costs of supporting changing technologies over time. Put simply, our policy for file formats is:
Complete documentation of this policy, including a list of currently known and supported formats, is available in DSpace Supported Formats. That document explains exactly what we mean by "support", and explains what to do if you have files with formats that aren't already listed.
DSpace implements access controls so items requiring usage restrictions will be properly limited. Some collections are limited to MIT network access only, for instance; other materials might be made freely available to anyone in the world. Please ask your collection's administrators what the collection's default access policies are. If you have an item that requires a different level of control than the default, please let the administrators know, and we can work together to set your items up properly.
If you are at all unsure about whether you have permission to submit a particular item to DSpace, don't do it! Get in touch with us and we'll help figure it out what will be best.
The fastest ways to contact DSpace support staff are via our support line, 452-3485, or via our help list, dspace-help(@)mit.edu. You can quickly send email to our help list from anywhere in DSpace; just click "feedback" at the bottom of any page. Be sure to give us a valid email address so we can respond.
You can also come visit us! Our office is 10-423, one floor down from the main floor of the Barker Engineering Library.
We intend to publish a broad range of DSpace documentation within our own DSpace community; look there for more details about:
The open-to-the-public DSpace site is dspace.org, where general overviews, staff information, and a FAQ are kept.