3
edits
Changes
→Islandora: added notes from breakout session
==Islandora==
Institutions that are live w/Examples
*Colorado Alliance of Research Libraries
**Hosts Islandora 6 repositories on Drupal 6; migrating to Drupal 7
*Florida State University ([http://fsu.digital.flvc.org])
Question about new installation: additional resources aside from documentation
Additional resources:
*Islandora Google Group
* series of cookbooks with Chef for one-click installation of Islandora
* Could use Vagrant and virtual box
* New release of Islandora about to come out -
* Lyrasis is asked to review the documentation and will do sanity checks on the documentation
* Should see improvements to documentation soon
Hosted solutions available to purchase:
* Lyrasis
* DiscoveryGarden
What's the learning curve with Drupal - easier/harder than Hydra?
* Drupal 7 is easier to learn than Drupal 6
* Recommend the Using Drupal / O'Reilly Book
** Useful to understand theming / design
* Drupal may be more widely-held skill set than Ruby on Rails (in which Hydra is based)
* Though at least one school's Islandora developer started with Islandora without a PHP/Drupal background
* DrupalCon - first week in June, Austin TX in 2014
* Islandora Camp ([http://islandora.ca/camps])
Setup difficulty?
* Documentation for setting up Solr / Gsearch - some confusion between v. 6 & 7
* Would be better for things to be more grouped together
Using Islandora for Media Streaming
*Video streaming: streams in HTML5 / used to use Flash in Islandora 6
*Every video gets an Mp4; audio files use Mp3
*How are storing the original videos? In Fedora. Note: objects do not have to be stored in Fedora
*Can still think of Fedora as the repository - include all master files and metadata. if need to migrate out of Fedora in the future, could simply retain master files and metdata.
*Media Streaming installed on Drupal server
**FlowPlayer; jwplayer available for Islandora 7
*Grinnel contracted with Discovery Garden to create generic content solution pack that allowed the creation of an object that held multiple types of files (pdf images, audio, video). Traded the ability to stream for the ability to create a compound object. Now need to do additional work to standardize for Islandora 7.
*Binary object created by Discovery Garden - for download.
*Islandora has several solution packs for multiple file types (video, audio, compound, large images, small images). Colorado use case involved oral history (video, audio, text documents together) which should be possible to use with the compound solution packs.
Versioning problems encountered:
*For each object - new folder is created? Can turn off versioning based upon file type; do version the metadata, but not the actual master files themselves.
*Could also use Fedora as a registry - not store the objects themselves because of versioning and performance/scaling issues.
How to handle updates to metadata?
*Checkbox to version or not.
*METS-like XML schema as a container (FOXML). When version -it adds it in to FOXML; Islandora only displays the most recent version
Note about custom modules:
*Custom modules can be difficult to get community support. out-of-house development makes it difficult to join back with the larger community and core code.
Long-term preservation:
*MetaArchive - locked server, dark archive. May be module available for checksum generator and checksum checker - runs checksum on ingest and periodically check them.
*DuraCloud - application on top of Amazon storage run by DuraSpace
*Premise - preservation metadata system that helps record whenever records are touched or changed for audit trail.
What happened to Hylandora?
* Came down to the content model: Islandora and Hydra were too divergent.
Reasons to go with Islandora over Hydra:
* Just depends on the skill set available
* Both Hydra and Islandora have active communities
Issues:
* Upgrades can be painful / difficult in both Hydra and Islandora
What kinds of content is appropriate for Islandora?
* Good for special collections and ETDs, images from Athletic departments and ContentDM
* Crossover with Content Management systems, emphasize a policy on what should be included for long-term preservation
* Grinnell - small liberal arts college; emphasize promoting a place for faculty and students to store their research (undergraduate papers, faculty white papers/publication, data sets, digital humanities, videos of performances, audio of music performances, preserving websites (difficult with Islandora 6 vs. 7, which has a WARCIVE module)
** Don't want to necessarily limit to a library-only thing; include instructional designers, academic affairs, art gallery, history,
* Florida: modules: large images, small image, pdf and books. Books module - good for yearbooks. Want to use it for ETDs and research.
** Workflow module - has been useful. Similar to DSpace workflow for ETDs. When we ingest objects and they're not ready to be published - can work with metadata, wait to publish and not be visible to the public until they are ready to publish.
Access Controls/Embargoes, Harvesting, and Statistics
* Access controls for data: Fedora use XacMole (sp)? to set permissions. can set embargoes for ETDs. IP embargo module. Could extend to the bitstream level. Pretty easy to do without breaking anything.
* OAI:PMH module - might not like how the Dublin Core looks out of the box - can use MODS out of the box to do the MODS to DC transform. Needs additional work to pass the 2.0 specification. Module that includes a REST API that will return data on various levels if requesting just specific items.
* How to handle statistics? e.g., what are the counts for objects per creation date.
** Perl script that will lookup the size of the collection, how many objects - on Github
** Google Analytics module in 7
==Relevance Search & Ranking==