Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Are you attending/planning to attend the conference? What would you like to discuss during the week?

Meeting notes

OAI-PMH harvesting

Friedrich Summann joined the meeting to present some of the most common pitfalls he noticed in harvesting numerous Open Access repositories to populate the BASE-SEARCH initiative. Although some 1000 DSpace's are properly configured for harvesting and indexing. For about 200 DSpace repositories base did encounter some issues. Mr. Summann lists the following problems:

  • DSpace is often not correctly configured. This problem mostly occurs in African countries, India, China, Colombia and Ecuador. The problem lies in the incorrect configuration of the handle system, which causes OAI-PMH to be unable to harvest the repository. The exact configuration errors do vary. In some cases a default handle URL is shown, in other cases the end user UI and PMH interface show different links, or display the same erroneous link. Around 170 DSpace repositories suffer from this.
  • Registered handle numbers are not correct or not working: It is not certain why this issue occurs. It might be caused by repository administrators who configure handles on their own, without registering with handle.net. Around 40 DSpace repositories have this problem.
  • Someone tries to repair the situation: When Friedrich notices problems with a certain repository, he contacts the DSpace administrator. This often results in an attempt to repair the situation. In reality this sometimes results in a repository containing a mixture of old, still incorrect data and new, correct data. It is possible this is caused by incorrectly running the update script. Facilitating the correct use of this script through enhanced documentation could be a solution.
  • The administrator is unreachable: Related to the previous problem, the repository administrator is often difficult to contact. In many cases the administrator email address is not configured. Often there is also no contact email address to be found in the end user UI. In case there is a contact form provided, submitting an entry to this frequently results into no response.
  • The harvesting process crashes: This issue was found with newer versions of DSpace. During the OAI-PMH harvesting process an internal server error causes the harvesting problem to stop.
  • OAI-PMH webapp is not deployed: Sometimes there is a working end-user interface, but no OAI-PMH webapp deployed.
  • There are no UTF-8 characters in the responses but question marks: This is likely not a DSpace issue, but a Tomcat misconfiguration issue.
  • Problem with the LYNCODE interface: This specific issue delivers a "No matches for the query" for ListRecords. It is possible this is caused by cronjobs which are not running (frequently).

OAI harvesting related questions from the community

OAISTER/Worldcat only supports HTTP, it does not support HTTPS. This problem was overcome on DSpaceDirect by configuring OAI to go through HTTP, while leading all other traffic through HTTPS.

The question rises if it would be possible to hide a collection for OAI-PMH. In case a collection is already access restricted in your own repository the contents will also not be harvested by OAI-PMH. It is possible the collection's name will be harvested, and thus be visible in the harvesting repository or platform. The contents of that collection on the other hand, will not be visible.

Update: Future of the DSpace User Interface

A prototyping challenge was held to prototype new UI technologies, as there already was a consensus the future of the DSpace UI lays neither with JSPUI nor XMLUI. In January all the candidate technologies were demonstrated. Since last week there is a group digging deeper in the different technologies, looking mainly for similarities.

However, the final decision for a technology is yet to be made. This will be discussed during next week's DuraSpace summit. One of the main points of disagreement is the question whether to stick to a server-side approach, or move on to a client-side approach using for example a javascript framework.

The DSpace UI working group has already agreed there is a need for User Experience (UX) expert who can help improve the end user experience. In case you have such a person in-house, who would be able to contribute to the new UI's user experience, feel free to get in touch with the working group.

DCAT meeting on Open Repositories in Dublin

As the conference program is already tight scheduled, we will try to have a short meeting over lunch.

DCAT members are free to suggest topics of interest. Currently an interest in discussing Statistics & Analytics is the only one addressed.

Call Attendees