 |
 |
|
 |
|
 |
|
 |
|
     
|
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
The IMesh Toolkit
[ Work In Hand > Components > Reading Lists]
IMesh Toolkit Reading Lists Module
This module generates an RSS file which contains a list of reading list materials. Given a list of URLs each of which retrieves a record which complies to the OAI_dc schema for GetRecord responses, the module will output an RSS-formatted list of items together with a channel description.
How can I use this module?
This module can be used with any repository that returns records using
the OAI PMH, and supports the oai_dc format.
You will need to provide a web form (supported by a script)
that enables the user to select the URLs,
and allows the user to enter a channel title,
description etc (full list of additional data that can be sent).
The data from the form is sent to the script which
then calls the IMesh RSS module with the channel
data and the URL list. The module returns a string containing the
XML that makes up the RSS feed.
An example of a crude form for entering a list of URLs
If you are integrating the module within a subject gateway or OAI search
service, it is expected that the subject gateway will provide the interface
for searching resources and selecting the ones to be included in the reading
list. The subject gateway should then create the list of suitable URLs to be submitted to the module. A description of the reading list (and other details) can also be sent.
Documentation and Download
Diagram showing the module used from a subject gateway
How is the RSS file used?
After the IMesh module has generated the RSS file, the file can be edited using an RSS editor, such as RSSxpress to make changes (Note that the file needs to be available at an http address). The RSS file can then be presented to users through a presentation system that understands RSS i.e. one that can display newsfeeds. Alternatively, newsfeeds can very easily be displayed in webpages, e.g. by using RSSxpress-lite.
Examples
To be addedIssues with OAI repositories
The IMesh module is built to work with resource description that are
available in OAI repositories. This should guarantee a base line level
of interoperability, and in particular the module is based on the
specific format returned to GetRecord requests,
and more specifically the unqualified Dublin Core format indicated by the metadataPrefix 'oai_dc' in the request.
In this regard, the open Archives Intiative Protocol for Metadata Harvesting provides an XML schema and states "For purposes of interoperability, repositories must disseminate Dublin Core, without any qualification . Therefore, the protocol reserves the metadataPrefix `oai_dc', and the URL of a metadata schema for unqualified Dublin Core, which is http://www.openarchives.org/OAI/2.0/oai_dc.xsd . The corresponding XML namespace URI is http://www.openarchives.org/OAI/2.0/oai_dc/ . "
Some issues that arise when trying to use OAI records (in the oai_dc format) as a basis on which to build reading lists are noted here:
Whilst the schema accommodates the three fields of description that can be reused in the RSS item description (dc:title, dc:identifier, dc:description, equivalent to the RSS item title, item link and item description respectively) it does not mandate any of the fields. Therefore there are no guarantees that the fields will be present in all the records returned.
Furthermore, with respect to the resource identifier, the OAI-PMH states "the nature of a resource identifier is outside the scope of the OAI-PMH. To facilitate access to the resource associated with harvested metadata, repositories should use an element in metadata records to establish a linkage between the record (and the identifier of its item) and the identifier (URL, URN, DOI, etc.) of the associated resource. The mandatory Dublin Core format provides the identifier element that should be used for this purpose."
The flexibility in the record content returned when requesting oai_dc format presented the following problems:
- No dc:description field is provided
- Use of dc:identifier
- Multiple IDs are given in the record, ie. the dc:identifier field is repeated
- The dc:identifier is not a URL but a link to the OAI repository identifier
Where no descriptions of resources are available in the record, this makes for arather limited RSS feed. The user may, of course, add a description after generating the RSS, by using an additional tool, for example an RSS Editor.
For the purposes of reading lists, the linkage between the record and the identifier of the associated resource is rather important, since the aim of the end-user of the reading list is to access that associated resource. It is clearly a problem if this link cannot be established. Two examples of difficulties in defining the URL of the resource that the record describes are:
RSS items only have one URL link per item. The RSS link is generated from the dc:identifier field in the metadata portion of the OAI record. Where two or more IDs are present in the OAI record, if a URL which contains http in the address is found, this is chosen as the ID to be used in the RSS feed. In the case of two identifiers with http URLs, the choice is currently arbtrary (the first one listed gets picked).
Although some of the above instances do not break the OAI-PMH, they highlight some specific requirements for using OAI records to produce reading lists, namely a meaningful description of the resource is preferable and a unique resource identifier which gives access to the resource from an RSS feed is required. The examples of repositories which presented problems are linked below for reference. They were picked (using no specific criterion) from the list made available through the repository explorer, although some repositories looked like a better match for the kind of resource that would be likely to be useful in a reading list (e.g. it is not clear that an image would be the kind of resource that would be included).
- No decriptions
The Celebrating Women Writers site links to texts of publications by women. The returned records give lots of metadata about the text linked to, but does not provide a short description.
E.g. http://digital.library.upenn.edu/webbin/OAI-celebration?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Acelebration%3Aalcott%2Fgirl - Multiple dc:identifier tags
E.g. http://dlese.org/oai/provider?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai:dlese.org:DLESE-000-000-001-982&rt=text
In this case, screening for the presence of http picks out a link to the resource associated with the record. - dc:identifier is not a URL to the resource
In the Digital Library of Information Science and technology the identifier is a URL that uses http, but in this case the identifier retrieves a web version of the OAI record, from which a link to the associated resource can be followed.
E.g. http://dlist.sir.arizona.edu/perl/oai2?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3ADLIST.OAI2%3A46
This record gives the dc:identifier as http://dlist.sir.arizona.edu/archive/00000046/ which is a link to the record itself describing the pdf resource (which is the described resource).
However in this second example, the http://www.teachersdomain.org/ website, the identifiers given in the OAI record are identifiers for the record in the repository, which would not be usable as a link in the RSS feed in the form that it appears in the record.
E.g. http://www.teachersdomain.org/cgi-bin/oai.cgi?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Ateachersdomain.org%3Asci.life.cell.lactic