Arthur,

I understand what you are saying, and I agree with much of it.  While it's only one (and probably incomplete) method to do so, we are hoping that by gathering various scenarios from multiple institutions, we are gathering some sense of what type of functionality people require.  I understand this factor may, in fact, end up not reflecting the choices that *people* make, but I hope there's some connection between what people state they need, and what software they choose accordingly.  Then again, I've always believed that economists are the only ones who really believe that people make rational choices :-)  It may be the case that people are simply making choices because they feel compelled to do so, and don't have a road map or guideline for making choices.  Without some sort of objective and rigorous analysis, people most probably can not make informed or even consistent decisions.  What we are attempting to do is not easy, and requires a fairly unique set of expertise.  In other cases, people may not have the resources to conduct a more thorough review.

I would hope that by identifying and delineating relative strengths and weaknesses of various systems, we might encourage developers to refrain from building the entire set of functionality within their particular software, and rather interface with other software that offers complementary functionality.

Regarding the point about researchers as "searcher" as opposed to "author" I would welcome further feedback from you.  Would you be willing to develop scenarios that reflect these distinct roles as you see them?  I would also be eager to hear what you may have heard from others in response to your original message.  Thanks.

Sayeed

On Dec 10, 2005, at 5:50 PM, Arthur Sale wrote:

Sayeed

 

Thank you for this email. It was interesting to read and look at the wiki, but I think what I am asking is completely orthogonal to what you are doing. I am not looking to choose software, nor to evaluate their technical features. May I say however, that you do not distinguish between the roles of researcher as “searcher” and as “author” which are two completely disjoint roils and which demand separate attention.

 

Rather, I am trying to find out why *people* choose the software they do. Many choices are puzzling, and the stated reasons are often simply absurd. When implementation costs for an IR range from $3k to $1M, something is going on that is unexplained. Partly that is explained by different purposes for IRs (for example for some IRs digital preservation is a non-issue; for others it is high on the priority list even if long-term solutions are not available).

 

Until we resolve why people choose what they do, we are destined to walk in the dark, especially the developers in adding features to IR software, which already show some signs of software bloat.

 

Arthur

 


From: Sayeed Choudhury [mailto:sayeed@jhu.edu]
Sent: Sunday, 11 December 2005 06:44
To: Stevan Harnad
Cc: dspace-tech@MIT.EDU; dspace-general@MIT.EDU; eprints-tech@ecs.soton.ac.uk; Fedora-users-list; ahjs@ozemail.com.au; diglib@infoserv.inist.fr
Subject: Re: [Dspace-general] Comparison of EPrints, DSpace and Fedora (fwd)

 

** Apologies for receipt of duplicate postings **

 

Arthur,

 

I have noted your message on more than one list, so I've taken the liberty of copying each of them to inform individuals who follow those lists. Regarding a comparison of repository software, our group at Johns Hopkins is conducting a technology-based analysis of repositories and applications with funding from the Mellon Foundation. Our choices include DSpace, Fedora, ePrints, Digital Commons (ProQuest's offering based on BePress), and applications include Sakai, and various e-publishing systems such as Open Journal Systems (OJS), and DiVA. We are working with the DPubS team to include their software as well, and we're also considering if we can include Moodle and LionShare as well. Finally, we have a strong emphasis on digital preservation capabilities for the repositories. The main purpose of our analysis is to examine each of these systems with a transparent, clearly defined methodology beginning with stories or scenarios that are mapped into use cases and what we are calling key events, from which we are defining functional requirements for repositories to support various types of content and uses. In addition to the repositories and applications, we are examining whether JSR-170, OKI DR OSIDs and perhaps an implementation of IMS DRI can support integration of repositories and applications through a generalized interface layer that spans across different services without specific out of band agreements.

 

While we haven't included costs information, we have tracked installation issues, features, and other technology matters. Ultimately, we hope to develop a methodology that might allow the community to address the types of question you describe with an objective, rigorous approach, and a taxonomy of repositories that will allow us to identify gaps in functionality.

 

 

You can also find further information at http://ldp.library.jhu.edu/projects/repository

 

From this web page, you can find our original proposal to the Mellon Foundation and two presentations at previous CNI and DLF conferences under the "Documents" tab.

 

I would be happy to answer any questions you (or others) might have in this regard.

 

Sayeed

 

 

Sayeed Choudhury

Associate Director for Library Digital Programs

Hodson Director of the Digital Knowledge Center

Sheridan Libraries

Johns Hopkins University

 

 

On Nov 19, 2005, at 6:53 PM, Stevan Harnad wrote:



 

*Forwarding: apologies for cross-posting **

 

---------- Forwarded message ----------

Date: Sun, 20 Nov 2005 08:27:46 +1100

From: Arthur Sale <ahjs@ozemail.com.au>

Subject: Comparison of EPrints, DSpace and Fedora

 

I am doing a comparison of functionality and adopter experiences for the two

most widely used institutional archive-creating software packages for

repositories: EPrints and DSpace, and also Fedora (a minor player globally

but possibly important in Australia). I am seeking your help in collecting

information. Information about other packages would also be welcome.

 

(1) If you have used or compared any of this software, could you

please take the time to let me know what you consider the respective

advantages/disadvantages of each to be, and for what purposes? I am also

interested in features that you think are equivalent or readily achieved

in each.

 

(2) The two major software packages explain their orientation as

follows: EPrints puts a particular emphasis on OA content (preprints

and postprints of institutional research output, plus theses), DSpace

on digital curation in general. Fedora describes itself as repository

storage layer software requiring custom front-ends for any purpose. If

you have any specific comments on these overall orientations and

whether they are appropriate, they would be very helpful too.

 

(3) While all these packages are free and open source, I would also

be interested in any cost estimates in implementing the one you chose,

how many hours or dollars you spent on setup, how much maintenance you

have to expend, and how reliable the software is (crashes, downtime,

etc). Would you recommend it to someone else?

 

I will post a summary of the results (and maybe an interim report) on AmSci

OA Forum, and may get back to you if I reed a bit more detail. Thank you in

anticipation of a prompt response and a flood of emails. Email direct to me

at Arthur.Sale@utas.edu.au if you want.

 

Arthur Sale

Professor of Computing (Research)

University of Tasmania

 

_______________________________________________

Dspace-general mailing list