Tuesday, October 11, 2011

Metadata Use Cases

Are there situations where you would like to change your metadata, but lack the tools or technical expertise to make those changes happen? 


Please tell us what are you want to accomplish. If you can, include what kind of environment (organizationally, technically) you'd be working in, along with anything else that you think would help other understand what you need.


Thank you for sharing your thoughts!

13 comments:

  1. My data is locked in a system I do not control. The ILS is run by the IT people of my organization and seem to have it under "lock and key". If I want to share it - the only way I can is to have people use the z39.50 standard - which is better than nothing but very limiting. There is so much talk about moving beyond MARC but I don't even have access to my MARC to begin to move beyond!

    ReplyDelete
  2. Here's one: I can get records in, for example, MODS from my ILS for reuse in exhibitware or virtual collection tools which are OAI harvestable. I'd like to be able to pass a batch o'records to an intermediate tool where I can specify data elements to look up against specific data sources, such as but not restricted to names in VIAF, subjects in id.loc.gov, geographics in Getty TGN. The goal is to find a match and populate the authorityURI and valueURI attributes in MODS. Give me an interface where I can preview a result list of potential matches, with the option to expand a line to see the MODS record and the authority record(s)(even though record is a dirty word now) side by side. Let me mark matches yea or nay from the result list or the fuller view. In a perfect world, I could set a threshhold for matches I don't even need to see. Spit out a revised set of MODS records with URIs added to attributes for the controlled values, which I can then publish as I see fit. Hope that makes sense!

    ReplyDelete
  3. I would like to work in a department that values metadata and includes metadata librarians in discussions on metadata. Unfortunately, my unit is too often associated with the "catalog" and MARC data. We try again and again to promote our skills with mapping, transformations, and the like. More often than not, the answer we get is that descriptive metadata isn't necessary or doesn't add value to the record. It is sufficient to just get the object out there on the web with a couple of keywords and a title. It's frustrated, especially when you see that the data is inconsistent or when people don't even know how it's being mapped from one system to the next or don't care. So how do metadata librarians get past this attitude of not caring about descriptive metadata?

    ReplyDelete
  4. I would like tools that allow visual representation/manipulation of metadata and items resulting from metadata searches. For example: visual representation and navigation of related items/collections, like http://www.muckety.com/Rick-Perry/5139.muckety .

    ReplyDelete
  5. Working with a digital repository and trying to find a clean, streamlined solution for gathering descriptive metadata for serials and multi-volume monograph series. Ideally we want one resource record for serial/series in the repository with links to its vols/issues/parts. Preferably, we want a way to automatically pull the vol info from our ILS, most likely from item records, and then to create the vol links in the resource page with that data. I've looked at the BHL Wonderfetch workflow and devised a semi-automated workflow by putting the vol info into a spreadsheet, sort of based on the Wonderfetch model, then turning that into xml and merging it into our metadata files. However, the spreadsheet is manually created. We can't fully automate our descriptive metadata creation processes if we have to manually create these spreadsheets.
    I don't work closely on our ILS (Voyager) but it's my understanding that we cannot export item or holdings records.

    ReplyDelete
  6. I agree that there seems to be significant gulf between the needs of metadata librarians and those folks developing the systems designed to manage digital resources and metadata. Even the most basic functionality such as control of names seems to be absent in those platforms that would benefit from it the most - institutional repositories.

    ReplyDelete
  7. "I don't work closely on our ILS (Voyager) but it's my understanding that we cannot export item or holdings records."

    Theoretically there are no limits with Voyager, as you can query the database directly. So it is possible to write a program that pulls whatever information you want from the catalog (including item records). Of course there could be practical barriers in terms of the system being locked down, or not having the local expertise etc.

    Additionally, Voyager provides mechanisms for bulk export of bib and holding MARC records. There are also APIs for getting at bib to item data, depending on which version of Voyager you're running.

    ReplyDelete
  8. Great! Let's Say: I found SPARQL dataset with nice sparkley data ready to link. I want to link to it in my RDF document (that, let's say I have already online) - I need a tool that helps me put the links in the right places. Right? Automatically somehow?

    ReplyDelete
  9. The Subject is Subjects. I'd like to have more options for providing subject access than "LCSH" - particularly for metadata for digital projects outside the ILS. We've used it in the past because it's the only comprehensive maintained vocabulary in English, but there are issues... I'd like to be able to link user supplied subject keywords (or perhaps, a bit more controlled vocabulary for specific projects) to LCSH via a semi-automated process (don't trust fully automated for something like this, but a tool where I can see the pairings and parts of the item description and make changes). Then if I need to port metadata to something that wants or likes LCSH (such as our "discovery interface" Primo) in order to have things collocate with other resources, I can swap in LCSH for the user vocab - without having to make the metadata creators (untrained) have to grapple with LCSH or even FAST... just use words they are familiar with (or better yet have an automated process suggest terms they can pick from or add to).

    With or without RDF/linked data (preferably with)...

    Also, I agree with Anonymous above... about everything...

    ReplyDelete
  10. Eduardo F. Cidade, Sr.October 24, 2011 at 9:51 AM

    Metadata is as Metadata does. If a system allows users to embed metadata, then the architecture has been well thought out. If it doesn't? It's time to address the weakness within the system to the architects. Case in point: We use an interactive library whereby we append just a little metadata to PDFs interlinked through a search engine. It was considered passe' because of the associated manpower costs to do this. Once demonstrated WITH metadata, the eyes opened and now it's a matter of course: Over 1,000 PDFs joined together in a searchable index providing over 5,000 users more tools than was initially thought of......Go back to the architects and have them fix your system. It may seem to be a simple fix, but then again, shouldn't this be an issue of simplicity over complexity?

    Have an interesting day.....

    ReplyDelete
  11. Although we have a library catalogue I am ever more conscious that there are ways in which I would like to be able to alter the way records are presented so as to increase their value. There is also a desire to re-purpose records in different scenarios, again to increase access. Based on our current Millennium system it is unclear what routes can be taken to improve presentation and access that are not too tech-heavy or require another major system.

    ReplyDelete
  12. That was my post above about "Subject is subject". I'm also looking ahead to wanting to move parts of description between records at different "levels" - for example, description at file level, "item" level (say, 4 digital files that represent a 2 page letter), "set" level (a group of correspondence between two people that contains the letter), etc. Some data would be the same - the digital publisher, the date (s) created, the digital collection this is a part of. Possibly subjects, genre, and other fields. How do we technically make use of data at different levels for display, metadata output, etc. without repeating metadata in all those records, in a way that will make sense?
    System is a Fedora repository with interface/infrastructure "in process"... but problem is applicable to other systems as well.

    ReplyDelete
  13. Assuming that digital objects deposited to a digital repository get assigned a persistent URN, could each object please be embedded (upon repository ingest) with a URN *linked to the object's descriptive metadata record*?

    As library metadata and digital objects become available more and more via APIs (happily freed from our library interfaces), I would like to embed files with a homing beacon that leads back the item's authoritative record.

    ReplyDelete