DESIRE: Peer Review Report: D3.3b-DESIRE Toolkit (Final Release): Eric Miller

Project Number:

RE 4004 (RE)

Project Title:

DESIRE II - Development of a European Service for Information on Research and Education II

Deliverable Number:

D3.3b (final)

Deliverable Title:

DESIRE Software Toolkit

Deliverable Type:

PU

Deliverable Kind:

PR

Principal Reviewer:

Name

Eric Miller



Address

Online Computer Library Center, Inc. Office of Research and Special Projects 6565 Frantz Road, Dublin, Ohio, USA>



E-Mail

emiller@oclc.org

Telephone

(usa) 614-764-6109



Fax

(usa) 614-764-2344



Credentials

Chair, W3C RDF Model and Syntax Working Group 1997-8; Member of Dublin Core Metadata Initiative Directorate.


Other Reviewers:


(if relevant)


Summary:

Relevant

4 (1 = poor, 5 = excellent)



State-of-Art

4



Meets Objectives


4



Clarity

4



Value to Users


5


Specific Criticisms


1

Original comment: It is hard to review this as anything other than a 'work in progress'; if the deliverable continues according the plans set out in the interim deliverable report, this toolset and the infrastructure being put in place to support it should be of very significant value to a number of sections of the metadata and resource discovery community.

Final comments, May 2000:

The presentation, packaging and implementation of the toolkit components have improved noticably since the Interim Report on this work reviewed previously. In particular, the Toolkit Web pages and CVS repository have improved and the installation of the Java RDF/recommender tools has been made much easier.

These tools constitute an excellent basis for further refinement and integration in future projects. The original scores assigned during the interim review remain unchanged: the work appears to have continued according to plan and my previous evaluation was broadly on target.




2

The Microsoft Word version of the Toolkit overview will be more useful to many in the community when made available online in HTML (ideally with XML or other machine-readable descriptions of the component descriptions).
Final comments, May 2000:

The files are now available in HTML.



3

The scope and aims of the DESIRE tool-set (and project) should be made more explicit in the toolkit documentation.
Final comments, May 2000:

The revised Toolkit Scenarios section and the improved documentation goes a long way towards addressing this concern. It remains hard to precisely describe the scope of a broad initiative such as DESIRE, and therefore difficult to precisely scope the aims of a toolkit such as DESIRE's. The new documentation does a good job of providing some detailed information on the capabilities of the various components.



4

Developer Response:


1

We have addressed Eric Miller's concerns as raised in the Interim Review, and expect to make further efforts in post-DESIRE refinements of several of the components.



2

3

4


General Comments

There is a demonstrable need for "Open Source", standards-based, interoperable tools for creating, processing and aggregating metadata across indexing, cataloguing and library-like Web applications. The DESIRE work in this area, as described in the toolkit pre-release and other DESIRE deliverables, goes some significant way towards showing the community the benefits of investing in open solutions in these areas. While it remains to be seen whether the final deliverable lives up to this early promise, the DESIRE Toolkit work as outlined in this document seems to be an excellent basis for a modular and extensible tool set. In particular, the emphasis on the supporting environment (ie. CVS for collaborative development, IRC for inter-developer communications, and the concern for post-project continuation of support/development) are to be commended. The software components created and refined within DESIRE appear to make a number of innovations, but will benefit from further work on packaging and "shrink wrapping" over the remaining months of the project.

Final comments, May 2000:

My original comments still stand. The packaging and shrinkwrapping work of the last few months has improved the accessibility of the Toolkit for external developers. This work (the DESIRE toolkit) has come to an end at a time when the standard infrastructure in this area is itself just reaching maturity. For example, we recently announced the Dublin Core Metadata Initiative's approval of an initial set of qualifiers for DC metadata. While this wasn't available during the development of the DESIRE tools, DESIRE appears to have put in place a framework that will allow future projects to build upon, rather than re-invent, some basic building blocks for metadata interoperability. I expect these tools to find widespread use amongst Dublin Core, XML and RDF implementors, and through that use to benefit from improved documentation, debugging and 3rd party contributions.

Eric Miller