Page tree
Skip to end of metadata
Go to start of metadata


Agenda/ Notes

  • DSpace-CRIS hands-on (locally installed, review/discuss personal experiments)

    • The default data model of DSpace-CRIS is defined in this excel file: configuration-tool-demo.xls loaded in the system at the end of the installation process, see Installation#TheDSpace-CRISsampleconfiguration
    • Some documentation about the use of the excel file to change/create the data model is available here: configure the data model using an Excel file (please note that you can import/export the data model at any time combining changes done via the UI with changes done on the excel file. The excel file only add new configuration, to remove entities and attributes you need to use the UI and migrate by hand existent data.


  • Andrea Bollini made a simple overview of DSpace-CRIS documentation in DSpace wiki. He started the setup process using an installation of DSpace version 4.
  • PG: Is it possible to migrate between different versions?
    • Andrea: The migration process is the same as DSpace. It's possible to migrate between DS3 and DS5, for instance. The tricky part is to migrate older versions, or systems that had several migrations before.
  • Andrea Bollini showed the migration process between DS4 and DSpace-CRIS 6. In the migration process it's required to run several steps manually, including executing SQL queries. It was created manually specific folders or directories (for files for researchers profiles, organizations, etc..)
  • Andrea Bollini followed the DSpace-CRIS setup instructions located at this page:
    • Tim suggested that Bollini could explore FlywayDB to simplify and automatize the migration process. One of the FlywayDB features, the callbacks could be used to execute the registry-loaders, and the schema migrations.
  • Pascal: reported that he tried to install DSpace-CRIS using Vagrant but the system was in 100% of CPU usage.
    • PG: suggested Pascal to verify the folders permissions (solr, logs, etc). Tomcat system user need to have access to some folders in the deployed DSpace-CRIS.
  • Pascal: also asked the why of the SKIP_IDENTIFIER_CHECK setting.
    • Andrea reported that it's a restrition of newer versions of Tomcat and it has to be this way because it's a workaround solution in the JSPUI, regarding the used Expression Language (EL).
  • After following the migration process, Andrea Bollini executed the following tasks in DSpace-CRIS:
    • Creation of a single User Profile
    • Creation of a new field in User Profile - Google Scholar Profile - an URL field
    • Creation of a new field in the User Profile - Blogs, via DSpace-CRIS Excel configuration file
    • Creation of a new Entity - Funding Program
  • Pascal: managed to deploy DSpace-CRIS, using Vagrant. He noticed that it's required more resources (memory) to be able to deploy all DSpace apps, but it worked fine just with JSPUI and Solr
  • One of the comments was that DSpace-CRIS seem too powerfull, yet to complex to manage.
  • This interface will be shifted to Angular, so it isn't to be considered as it is.
  • Tim: reenforced the powerfullness of the solution versus the complexity to support it, that caracteristic is a correlation in complex applications
  • RCAAP: made a DSpace-CRIS pilot and state that was not easy to use this interface and Repository Managers had to deal with it. He also exposed a concern about exposing stored entities using DSpace-CRIS (OAI, REST)
    • Andrea Bollini: the majority of Admin users doesn't change inputs_forms.xml, or item submission process for instance. It's required tecnhical skills to do that and typically those persons don't have it.
    • Lieven: aggree that the importance of the user interface should be directed to the common user. He states that we should consider the power vs complex. Flexibility versus Complexity, where to we draw the line?
      • How flexible do we want DSpace to be and the level of configuration we want in DSpace
      • Basic entities by default (author, projects, funding by option)
      • how do we create them?
    • Andrea Bolline: agrees, there should be an basic configuration and we should draw a line in what we think
  • Tim: Is DSpace-CRIS too complex? We should discuss this in a next meeting
  • The next meeting, to be scheduled, will have the following approach:
    • Trade-off between the powerfull and complex
    • Is DSpace-CRIS too complex,
    • What's the simple DSpace Data Model proposal and analyse DSpace-CRIS based on that, and how can we tweak it to reach it.
  • No labels

1 Comment

  1. Video recorded session: