Q: Why did Harvard build LibraryCloud?
A: Initially, the idea was to create a multi-library platform to enable broader use of metadata that libraries possess but have generally not chosen to make public. Over time, it became a server that would help those maintaining the complex information flows libraries deal with, as well as a platform for innovation.
Q: What is its current status?
A: Development of the pilot came to an end in November, 2014. It is a robust, functioning engine with working APIs, but as a pilot is not being developed further at this time. The code is available at GitHub for anyone who would like to use it. The development team is hopeful that additional funding will let us take it to the next level.
Q: What is the relationship of the two LibraryClouds?
A: Ah, so you have heard of the prior version. Excellent! The first version was created by the Harvard Library Innovation Lab. The new version, by Harvard Library Technology Services, builds on that work, and adds a workflow engine to automate its processes. The new version is designed to be more scalable and maintainable. The original version, now renamed lilCloud, serves as a prototyping engine for the new version.
Q: What is the connection of LibraryCloud and the Digital Public Library of America?
A: None, except that both are platforms that provide access to library metadata. LibraryCloud could eventually become a way Harvard shares data with the DPLA. And we are fans of the DPLA. There are some informal connections, however. The project lead for LibraryCloud (Jeffrey Licht of Pod Consulting) was also the leader of the DPLA's platform development team. Further, †he group that created the first version of LibraryCloud also created the first version of the DPLA’s platform.
Q: How about the connection to the Linked Data for Libraries (LD4L) project?
A: LD4L is a Mellon-funded project initiated by Cornell. The Harvard Library Innovation Lab and Stanford Library are both partners in that project. LD4L is a Linked Data project. LibraryCloud is not. We hope and expect that eventually there will be some level of integration between the two projects.
Q: How about the connection to StackLife?
A: StackLife is a visual library browser developed by the Harvard Library Innovation Lab. It uses the original LibraryCloud API. At some point, we might switch it to the new LibraryCloud API.
Q: Can I use LibraryCloud for my institution's library?
A: Yes please! The code is available on GitHub. Of course you’ll likely have to write your own connectors so that you can ingest your data, but please do and share them!
Q: Can I get the bibliographic data as a download or only through the API?
A: The bib data is already available for bulk download, thanks to Harvard’s Open Metadata policy that put all but a tiny percentage of those records into the public domain.
Q: How do you manage the privacy of the users whose the usage data you’re making available through the API?
A: We compute what we call a “stackscore” that accords each item a number 1-100 based on how relevant the item is to the Harvard community. To compute the stackscore, we combine checkouts aggregated over the past ten years, the number of times an item has been put on reserve, the number of copies in Harvard’s 73 different libraries, and more. We assign different weights to these factors. We do not believe there is any significant risk of identifying individuals who have consulted these works.
Q: What schema do you use, and why?
A: We use MODS because it’s a well-developed standard that captures a vast amount of detailed data relevant to library items. Our API returns JSON modeled on MODS or in the Dublin Core format. Dublin Core is simpler, and is used more widely across the Web, but MODS preserves more of the library-based information.
Q: What is the status and future of LibraryCloud?
A: LibraryCloud is a pilot project. We hope that it will be implemented and further developed by Harvard Library. In any case, lilCloud will remain operational for the foreseeable future, and the Harvard LibraryCloud code will remain available at GitHub.