2016-11-21 Radio Haiti "Item-ness" and A/V

Date

Attendees

Goals

  • Review / set data model for Radio Haiti
  • Discuss A/V for Radio Haiti as time allows

Discussion items

TimeItemWhoNotes
5 minutesReview Meeting Goals



Radio Haiti Data Model Discussion
Project info: https://docs.google.com/document/d/1mWj37zS84jPMnfrAutcPLGayy0ilbDaQ44y7loKHvzU/edit

A/V for Radio Haiti
  • Ingest into DDR
  • Display needs
  • Discuss YouTube-like streaming for access from Haiti

Itemness:

→ what about metadata on the "tapes" tab?  do we need to preserve this information in anyway. 

files are items

items have one component that have 1 audio item.  

Differnt types of items - are listed on the project info. 

Structural metadata could help solve some of these issues – but a descriptive metadata approach may be better for RH. 


Questions:

  • Are there overlap between the types of items? 
  • Are things that are sequential also copies of something else.  


Suggested approach:

  • Use the same solution we used for Duke Chapel recordings.  
    • Instead of using the "series" field, use "category" to note relationships between items.  This is an unordered relationships. 
      • We can potentially force the items to sort in order based on file ID. 
      • This is the simplest approach. 
    • Using "relation" on the collection level to handle the lib guide links. 
  • We can have different kinds of related items, but not in a sequence. 
  • They way we have created sequence order has been w/ METS in Tripod2. 


alternate approach:

  • Single item w/ multiple components and structured metadata
  • DDR does support structured metadata but only w/in an item.  Changing this to support multiple items could be overly complicated.


Mock ups:

  • ask Craig and Laura to point us to at least one example of each type of relationship. 
  • We will create some mock ups from there. 


A/V in General

  • We can fall back on the Chapel Recordings solution for this collection if need be. 

Plan is going to use RH as the driving force to getting a solution together.  

Avalon is out. 

Warpwire is the current exploration focus. 

  • Warpwire would accommodate the high and low resolution copies elegantly. 


Time synced metadata

  • We could possibly do something where the time codes display as links and could use those to click to that time of the file. 
  • Sub-description would be indexed for discovery. 
  • Could bookmark a timecode and use that as a link. 
  • Also a possibility to use chapter information.  

General plan:

  • find the long term A/V solution first
  • Captions / time-coded metadata features are part of our requirements for the tool. 


Access from Haiti:

  • Back up:  dump a copy of all the A/V on YouTube w/ Haitian metadata
  • How light-weight is DDR?
    • Its the actual media that slows everything down?
    • If we are delivering low-fi files, will that be sufficient?
    • Warp-wire checks bandwidth of the end users and delivers file appropriately. 
  • YouTube:  advantage is the content delivery network. 
  • Questions:  how heavy is the DDR-Public now?
    • Can we use this project as a way do some testing on DDR-Public?
  • Maggie:  panel at DLF on distributing content to low-bandwidth locations.  


Launching in general:

  • If the files are not all ready, how should we handle the launch?
    • Metadata:  if all the metadata isn't standard then there could be issues w/ later batches that we don't find at the beginning.  
    • ingest/launch:  we would like fewer big chunks of material than infinite small chunks.  
    • Lets try and just ingest what we can launch at the time so we do not have to separate things that will be launched and not launched. 


Next Steps:

Molly to ask Craig / Laura questions

Cory to spearhead more mock ups



Action items

  •