Lightning talks on Day 3 of Code4Lib.
Bplgeo: A Geocoding Gem – Steven Anderson
Why?
* geographic data is neat, but unstructured and lacks standardization
* names are great for faceting
* great for plugins with coordinates
Steps
* just clone
* mapquest not so great
* bundle install
* just added
Code
Boiling a Frog: A Responsive Update – Charlie Morris @cdmo & Angie Fullington @afullington
- Starting with a good site, but was fixed width and needed more flat, responsive, brand/style standards
- Turn up the heat up slowly, won’t jump out
- doing a series of incremental changes e.g. increased font size, remove borders, remove drop-shadows, replacing icons
- Magic Escalator of Acquired Knowledge – need to fill knowledge gap between current knowledge to target knowledge
- wanted to avoid redesign because might fall back on escalator
- Sneak Peek – will launch end of semester
- Code
Dev-Ops at Penn State University – Justin Patterson
- bridge between sys admin and developers
- incorporated Capistrano (remove multi-server automation tool) – routine actions across multiple servers
- has many recipes e.g. Passenger upgrade
- server automation using Chef (server automation framework for deploying servers and apps)
- Gemnasium checks for vulnerabilities on all dependencies
- Repo Security use GPG encryption, have everything in public but secure
- Github
How to be a part of an Open Source Community – Jennie Rose Halperin
- we’re all using open source tools
- lots of benefits to contributing, particularly interacting with the community itself
- can contribute to lots of open source projects in lots of different ways, but don’t have to be a “coder”
- outreach for women program
- http://www.whaticandoformozilla.org
- Teach with webmaker
- bring open source to your campus
- open badges
Putting Things in Finding Aids – Tim Shearer
- discovery of finding aids using HTML version, had minimal metadata, Contentdm, support variable rates of ingest, flexible for volume, work independently
- did have consistent metadata, tight workflow and policies, easy to use for ingest, unique IDs for collections and containers
- Solution: XML -> XSLT -> .html / .xls (for Contentdm)
- Loose systems: Tight data (finding aid digitized content)
- Web API on top of Contentdm
- robust and extensible, because lets Contentdm handle content types
- sustainable: workflow well defined, integrates, simple to train
- problem: it was slow
- solution: caching shim, updated nightly
- UX updated
Schema.org and Google Custom Search for Finding Aids – Sean Aery @seanaery
- embed schema.org data into our HTML (using RDFa Lite)
- power our search with Google’s index
- powered by custom search, but built own rich snippets
- digital collections was a little more complicated (vs. finding aids) but using Google XML API instead
- good: DIY snippets (possible in free CSE), relevance, speed, non-roman characters
- gotchas: results/query limits, unstable, imprecise results count, quirky structured data omissions via APIs
- See also the blog article about schema & CSE
Chicago Collections Consortium
About
* 12 Chicago area institutions
* looking to build portal to integrated access to Chicago focused collections across members
* a lot of similar stuff across collections
* include digital images and fully searchable finding aids
Problems
* everyone has different metadata “standards”, differences even within
* no two EAD implementations are the same
* “descriptive standards” are very different
Technology Choice
* XTF from California Libraries
* flexible metadata processing
* no editing of the original metadata needed to tame subject madness
* lexicon in separate file
* will be making code and documentation available online
Contrary Technologies: Stuff you’re not using and probably won’t – Ian Walls
- CMS: Open Source, PHP, MVC = SilverStripe, already using it, capable, fewer developers and modules, but have local skillset
- responsive design framework = open source, gridbase = Zurb’s Foundation, easier to customize, great documentation and training, lower barrier to contribution
- search engine: open source, Lucene = elasticsearch, incredibly scaleable, JSON instead of XML, mix and match schemas, nested documents, baked in geospatial support, pluggable scoring scripts, rivers (slurp things in e.g. from Twitter)
Browse-Everything for Rails – Michael Klein
- Rails Gem
- developed for hydra, but not hydra specific
- everything: server-side directory, dropbox, skydrive, google drive, box.com, extendable = just add drivers