Hello Tim,
Thank you very much for this detailed presentation ! And thanks to
Christian for his advice.
In our case, we have considered Islandora because we have a lot of projects
under Drupal. But something bothers me: Islandora runs with rather old
versions of Drupal (7.x) and Fedora (3.x), and the new major version
("Claw") is under development.
I think at first we will be able to meet our customization requirements
with DSpace.
Thanks again ! J.Dornbusch
Le mercredi 11 octobre 2017 16:32:05 UTC+2, Tim Donohue a écrit :
>
> Hello J,
>
> It's hard to make broad comparisons between DSpace and Fedora as the
> systems are so very different in nature. And which system you use should
> likely be heavily based on your local needs, expertise, and goals.
>
> Fedora is a highly flexible data storage architecture/platform, with
> native linked data support. It natively speaks RDF, and the storage of all
> objects within Fedora can be highly customized based on local needs.
> However, few institutions use Fedora directly (as it requires a lot of
> resources to plan out your local Fedora data model, develop interfaces
> against it, etc). Instead, most institutions that use Fedora actually
> choose to use either Samvera (used to be called Hydra) or Islandora, which
> are full repository solutions (with their own pre-built data models and
> user interfaces). Samvera provides applications based on Ruby on Rails,
> while Islandora is essentially a flavor of Drupal (drupal.org) that is
> built on a Fedora backend.
> * https://samvera.org/
> * https://islandora.ca/
>
> DSpace is an out-of-the-box repository application, with a mind towards
> allowing institutions to more quickly manage their digital content and
> disseminate it on the web (i.e. optimized for SEO). It has a built in data
> model which is similar in nature to Samvera or Islandora, but is *not*
> based on Fedora. To make it easier to work with and install, its metadata
> is based on Dublin Core. The metadata is customizable, but it is not
> hierarchical (like RDF), and instead is limited to the format:
> [schema].[element].[qualifier] (you can create as many custom schemas,
> elements or qualifiers as you wish however). DSpace is generally
> considered an easier application to install/maintain (than those mentioned
> above), which is why it is the most widely used repository application
> worldwide (with over 2,000 known institutions using DSpace). DSpace is a
> Java web application (with a database backend), and is scalable to support
> large numbers of objects. (Fedora is similarly scalable)
>
> As for which is more tailored for a data repository, it also likely
> depends heavily on your needs/goals. As noted, Fedora is more flexible,
> but that flexibility requires a lot of management/staff to maintain and
> support. DSpace is less flexible in terms of data model or metadata, but is
> generally easier to maintain with less staff. There are some larger scale
> data repositories built on both platforms (see below). You could consider
> contacting one (or more) of them to talk about their experiences with
> either platform.
>
> DSpace-based Data Repositories:
>
> - DataDryad: https://datadryad.org/ (This system is a custom, open
> source product based on DSpace that is tailored more towards data
> repository needs)
> - Edinburgh Data Share: http://datashare.is.ed.ac.uk
> - University of Minnesota:
> https://conservancy.umn.edu/handle/11299/166578
>
>
> Fedora-based Data Repositories:
>
> - University of Alberta, Education & Research Archive:
> https://era.library.ualberta.ca/
> - University of Michigan, Deep Blue Data:
> https://deepblue.lib.umich.edu/data/
> - ICPSR: https://www.datalumos.org/ and
> https://www.openicpsr.org/openicpsr/
>
>
> I hope that gives you a decent overview from which to continue your
> research.
>
> - Tim
>
>
> On Wed, Oct 11, 2017 at 5:11 AM J.Dornbusch <[email protected]
> <javascript:>> wrote:
>
>> Hello
>> I am trying to determine what would be the right research data management
>> solution for my institution. There are comparison grids on the internet
>> but they are often outdated.
>> The data will be in tens of Gos of all types, including office or pdf
>> files as well as low and high resolution images, videos, quantitative
>> data, shapefiles, XML-TEI...
>> We need to add specific metadata standards and controlled vocabularies,
>> to assign DOIs, to fine-tune access rights, to interface with other
>> applications following semantic web standards. We would like to have
>> powerful search capabilities (full text search in data and metadata,
>> facetting, stemming...), perhaps on-the-fly OCR processing. Authentication
>> would be with Shibboleth but not only. We will use Omeka-S as a front
>> end for the general public, knowing that Omeka S already has modules to
>> interface with Dspace and Fedora.
>> After considering Dataverse, we finally found it too little extensible,
>> although
>> we liked some features like the "explore" button that projects an item
>> into a third-party data visualization interface.
>> We now hesitate between Dspace and Fedora 4 + the new Islandora "Claw".
>> An advantage of Dspace is that we would quickly deploy the turnkey
>> interfaces. I have the feeling that Dspace 6 offers a level of
>> functionality comparable to that of Fedora, for example regarding
>> versioning, scalability...
>> Does Dspace have some limitations compared to Fedora ? Is Dspace suitable
>> for a large-scale data repository, with lots of custom feature requests and
>> integration to other tools ?
>>
>> I am interested in any advice or resource that would help us in this
>> choice.
>>
>> Regards,
>> J.Dornbusch
>> https://www.ehess.fr/en
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "DSpace Community" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected] <javascript:>.
>> To post to this group, send email to [email protected]
>> <javascript:>.
>> Visit this group at https://groups.google.com/group/dspace-community.
>> For more options, visit https://groups.google.com/d/optout.
>>
> --
>
> Tim Donohue
> Technical Lead for DSpace & DSpaceDirect
> DuraSpace.org | DSpace.org | DSpaceDirect.org
>
>
--
You received this message because you are subscribed to the Google Groups
"DSpace Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/dspace-community.
For more options, visit https://groups.google.com/d/optout.
