Page MenuHomePhabricator

Evaluate WorldCat Search API v2
Open, Needs TriagePublic

Description

This task involves the work of evaluating the WorldCat Search API v2 [i] through the lens of the benefits it could offer to volunteers editing Wikimedia projects.

This evaluation is one input [of many] into more broadly deciding whether we will prioritize migrating to the WorldCat Search API v2 within the next 6 months.

Decision(s) to be made

  1. Will the Editing Team prioritize migrating to the WorldCat Search API v2 [i] in the next 6 months?
NOTE: this decision will be made with input from other sources.

Open questions

This section contains the open questions we think answering will equip us with the information we need to make the decision(s) named above...

  • 1. What functionality has migrating to Zotero to fulfill ISBN citation requests (T336727, T336297) left volunteers without or with a more difficult time accessing?
    • We have no full text or title citation for the vast majority of books, only for academic papers.
    • Poor coverage of books published in Asia and Africa for ISBN search.
  • 2. How does the impact(s) "1." asks us to name vary among different groups of people? E.g. might volunteers from Sub-Saharan Africa be impacted differently than volunteers editing from within North America? If so, how?
    • Medium sized wikis for which no national translator exists are most affected. Small wikis often don't have citoid enabled.
  • 3. What work would be required of the Editing Team to migrate to the WorldCat Search API v2?
    • Dependant on implementation requirements.
  • 4. What – if any – work would be required of volunteers to migrate to the WorldCat Search API v2?
    • Likely N/A
  • 5. How often have people been using/attempting to use the functionality the WorldCat Search API v2 delivers? Ideally, this data can be split by platform, wiki, and request origin (e.g. third-party vs. on-wiki usage) and viewed over time (e.g. past two years).
    • This grafana chart offers some insight (thank you, @Mvolz). Tho, answering this question fully might require further analysis and/or the implementation of new instrumentation cc @MNeisler.
    • TODO: Check logs for how many of the requested ISBNs return no data, compared to total # of requests.

Background/References

Done

  • Answers to all "Open questions" are documented on this task

i. https://developer.api.oclc.org/wcv2#/Bibliographic%20Resources/search-brief-bibs

Event Timeline

Mvolz updated the task description. (Show Details)
Mvolz updated the task description. (Show Details)

Icelandic ISBN numbers are also a problem. They are managed by the National library (https://landsbokasafn.is), which does not have an search engine for them. They are added to "Global register of publishers" (https://grp.isbn-international.org), but that only shows publishers, not books, and is thus inadequate for Wikipedia use.
Prior to WorldCat's removal from Citoid, they showed up just fine.