>> Repository Comparison: CONTENTdm, DSpace, Fedora and WikiD.
'Repository' means different things to different people, even within the library community, but there are some common themes. Last year Ralph LeVan put together a framework to help us think about and compare some repositories we were involved with at the time. Since then Jeff Young has done a lot of work on WikiD which grew out of our work with OAI-PMH repositories. Jeff and I occasionally have discussions about how to explain WikiD to people and it occurred to me that it would be interesting to add WikiD to the three repository systems that Ralph looked at and see how it compares.
The three other systems are CONTENTdm, DSpace, and Fedora. As you are probably aware these are very different systems--so different that it may never have occurred to you to compare them on a feature-by-feature basis, but that's what Ralph did. Here are three tables, each focusing on a set of features: Data Support, User Support, and Miscellaneous Infrastructure. A solid bullet means good support, an open bullet partial support, and a blank implies the system doesn't have anything worth mentioning for that feature.
|Arbitrarily Complex Objects||•||•||•||•|
|Local Metadata Elements||•||•||.||•|
|Rich Metadata Searching||•||◊||◊||◊|
WikiD can accept arbitrary XML, but doesn't handle binary objects yet. It's batch imput (the ability to read in a file of objects) is limited to MARC-21 or something in an OAI-PMH repository. Ralph is working on giving Pears the ability to automatically generate descriptions of arbitrary XML databases, so that will extend WikiD's searching capabilities (right now something like full text search requires manual editing of Pears configuration files).
|User Roles with Privileges||◊||•||.||.|
|Arbitrary Bitstream Retrieval||◊||◊||◊||.|
|Arbitrary Object Retrieval||.||.||•||.|
|Content Easily Integrated into Web Pages||•||.||◊||•|
Object marshalling is the ability to compose a complex object (e.g. an album consisting of multiple mp3 files plus other information about the album) for submittal, and possibly manage the relationships between the component objects. Other than the Web interface that comes naturally with WikiD, it doesn't do so well in this chart.
Miscellaneous infrastructure is where WikiD really shines. In fact that's what it is built on!
I realize that not too many people can critique how WikiD is presented in the charts, but there are many people that have more experience with some of the others than we do. Anyone take issue with the marks? Maybe our biases are showing, or our information is just dated. What other sort of features should have been included?
I suspect that most repository selection decisions are made with only a passing interest in the capabilities reflected in these charts, but as soon as you want your repository to do something new, these characteristics can become very important.
If there is enough interest, I'll consolidate the comments (both here and email) and do some new charts.