Skip to content

Instantly share code, notes, and snippets.

@dmolesUC

dmolesUC/c4l16-day1-notes.md

Last active March 11, 2016 17:26
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save dmolesUC/d4fa7881884e7f23f389 to your computer and use it in GitHub Desktop.
Save dmolesUC/d4fa7881884e7f23f389 to your computer and use it in GitHub Desktop.
Download ZIP
Raw
c4l16-day1-notes.md

8 March 2016

Keynote: "Code for Liberation", Kate Krauss, Tor project

  • interviewing activists to find out what they need from libraries
    • BLM
    • transgender

BLM

  • personal experience
    • 19-year-old college student / sandwich delivery guy shot in his car by Philadelphia police near her house
    • endemic problem in Philadelphia & many other cities
    • BLM first to surface this in an organized way
  • BLM now under social media surveillance by FBI
    • cf. Occupy: Homeland Security, Joint Terrorism Task Force, NCIS, IMSI-catchers ("cell site simulators" used by police to intercept mobile phone traffic)
    • cf. COINTELPRO & civil rights
    • our anti-terrorism agencies have massive budgets to find terrorists, but there just aren't many terrorists to find, so they're finding terrorists where there aren't any

transgender community

  • discrimination -> control of personal information critical to job security, family relationships; can be life or death
  • there are political, social prices to pay when we collect data that we don't need
    • e.g. Aeon special collection system asks for considerable amounts of personal data, which is then accessible to library staff for generations
    • "when you keep that data, you are siding with your institution; you are not necessarily siding with the researcher ... that information can be subpoenaed"
    • "what to collect, what to retain, what to distribute, who can see what... these are political decisions, moral decisions"
    • ALA position: "the most minimal amount of data for the shortest possible time" (see ALA privacy policy guidelines)

who's a target?

what can libraries do?

  • teach workshops, e.g. using Signal for iPhone / Android
  • examine your access rules
    • "if you don't have it, you won't have to worry about it"
    • "don't collect what you can't protect"
  • "these are moral choices"
  • "if you have privacy at home, but you don't have privacy on line, you have no privacy"
  • DC public library had a privacy month: dclibrary.org/1984
    • livestreamed a 1984 readathon
    • privacy workshop for teams
    • film series
    • etc.
  • Tor exit nodes in libraries:
  • Host crypto parties
  • Host unconferences to learn activist or advocacy community needs
    • how to address a community that you don't know
    • "there's a whole bunch of software developers I know that want to make software that solves a problem that no one needs"

Q&A

  • Q: Google analytics?

  • A: "It's free, it works so beautifully, why can't we use it?" -> Tor doesn't collect any analytics, so they have to find ways to solve problems without analytics (e.g. whole-day A/B testing)

  • Q: Opt-in seems like a solution. [Ed. not a question]

  • A: People are so used to opting in, is it really informed consent? Opting in should be baseline.

'Can't wait for Perfect: Implementing "Good Enough" Digital Preservation': Shira Peltzman & Alice Prael

Bedrock principles

  • Bit preservation
  • Content management
    • ensuring files can be found, delivered, opened, read / played back
  • Ongoing management
    • preservation is an active, continuous process
    • requires ongoing funding & engagement
    • standards: OAIS, TRAC (aka ISO 16363 / TDR)

"Good-enough digital preservation"

  • "the most you can do with your current resources"
  • "probably a little more than you are"
  • institution-dependent
  • moving target
  • Don't go it alone
    • look at existing policies etc.
  • Inventory
    • for advocacy
    • for prioritization & budgeting
    • high priority items need more copies, geographically distributed, more fixity checks & monitoring
    • break down into sub-tasks, look for low-hanging fruit (naming conventions, basic metadata...), incremental daily progress
  • be an effective advocate for the material you try to preserve
    • you're going to have to educate people about / sell them on the whole idea of digital preservation
    • you'll need to communicate differently to different audiences
  • get digital preservation into your institution's mission statement
  • NDSA preservation levels are a good benchmark... but don't have any access-related guidelines.

"Enabling Access to Old Wu-Tang Clan Fan Sites: Facilitating Interdisciplinary Web Archive Collaboration": Nick Ruest & Ian Milligan

Every day of my life, I wished these archives were bigger ... my biggest problem is now abundance. I spend almost every day ... wishing we had less information.

  • Wayback machine requires you to know the URL
    • everybody knows that's not how people want to work
    • everyone's working on discovery
    • but discovery can't be a black box of ranking algorithms ("that black box is writing my book")
  • webarchives.ca: Warcbase + Shine

"Digital Preservation 101, or, How to Keep Bits for Centuries": Julie Swierczek, Harvard Art Museums

  • Slides

  • Handout

  • "when I say digital preservation, people think I mean digital storage"

  • "when I say digital archive, people think I'm talking about a backup drive"

  • OAIS is rocket science; it came out of NASA

  • Principles:

    • Provenance
    • Original order

  • Ingest

    • Special forensic floppy controllers, e.g. KryoFlux

  • Formats

If anybody tells you you should get a one-time grant for digital preservation, you have my permission to get mad.

"Guerrilla Usability Testing & Communicating Value", Eka Grguric

  • schedule regular usability testing sessions as part of your development schedule
  • guerrilla vs. standard
    • guerrilla is cheaper
    • minimal equipment, no labs, can be done by amateurs
    • 3-6 participants, broad strokes, most bugs
    • answers over statistical validity
  • figure out what you're testing
    • specific features: color scheme, can people logout
  • figure out stakeholders
    • subsets of user group: undergrads vs. grads vs. faculty
  • personas
    • you don't need robust, detailed personas
    • just get at needs & goals
  • goals
    • focus on the user, not on what you need to get out of testing the design
    • refine goals into concrete tasks
      • e.g. "look up grades" -> "look up your grades on the midterm exam"
      • bad tasks lead users ("log in, go to x, tell me what you think you would click on")
  • 2 facilitators: talker + notetaker (or record everything & then take notes)
    • bad prompt: (to frustrated user) "try logging out"
    • good prompt: "can you describe to me how you're feeling / what you're trying to do"

"Get Your Recon", Christina Harlow

  • Slides & notes

  • data is really messy

  • we don't want entities incorrectly not linked

  • we don't want entities incorrectly linked

  • solution: "get a student" to clean up the data, manually search authorities

Limits of linked data

  • limits of linked data services (OpenRefine, LODRefine)
    • mismatches, incomplete data, lots of data munging
    • different LOC services use incompatible APIs, aren't set up for bulk queries
    • some MARC fields just become opaque blank nodes when converted to RDF
    • Wikidata services don't provide fuzzy matching
    • theoretically standard identifiers turn up in different formats on different serves
  • "To fully realize the benefits of LD, a huge amount of entity matching / data remediation work needs to occur"
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment