MIT

MIT's Internet Policy Research Initiative and CSAIL run their public research programs on WordPress, and they hired me to take four of those sites, SCRAM, IPRI, Future of Data and the ACM Symposium on Computer Science and Law, from unversioned production edits to versioned, branded, deployable software.

Key Takeaways

  • Gave four MIT research properties the same engineering baseline: a local environment, version control and a zero-downtime deploy path each.
  • Wrote two functionality plugins that separate site behavior into team profiles, security, performance and asset modules, so research content stopped living inside page-builder layouts.
  • Rebuilt the research record itself: the SCRAM research page restored around the full publication list, with every paper linked to its canonical source.
  • Made team pages a content type, so a staff member adds a person and picks a group instead of a developer editing layouts every term.
  • Hardened authentication and the public surface: email-only login with uniform messaging, XML-RPC and discovery output disabled, and development artifacts cleared from the web root.
  • Redesigned the ACM Symposium site around its archive, where the menu is the record, one branch per symposium year.

Organization

MIT's Internet Policy Research Initiative collaborates with policymakers and technologists to improve the trustworthiness of interconnected digital systems. SCRAM, Secure Cyber Risk Aggregation and Measurement, lets industry participants measure cyber risk without disclosing their own losses. Alongside them sat the Future of Data program and computersciencelaw.org, home of the ACM Symposium on Computer Science and Law.

Challenge

Research programs publish through their websites, and these four had grown through direct production edits: no local environments, no version control, and people lists and publication records maintained by hand inside page builders. For programs whose output is published research, the research pages needed to reflect the full record, and the platforms needed to become safe to change.

Solution

  • Local environments for every property, followed by git deployment straight to each host.
  • The mit-scram and mit-future-of-data plugins, each organizing behavior into four modules with compiled, cache-busted assets.
  • A team profile post type with group taxonomy, ordered by the lab's own ordering, rendered as a responsive card grid in MIT branding.
  • Research and technology pages rebuilt around complete, linked publication records.
  • A security module covering authentication, login messaging, XML-RPC and scanner-facing discovery output, plus a cleanup of the public web root.
  • The symposium site redesigned with its archive intact and a reusable generator for session tables that every later program year could use without engineering.

Results

  • Four properties moved to versioned, deployable software with zero-downtime deploys.
  • People-page changes cut from developer work to a 90%-lighter staff task.
  • 11 research entries restored and linked on the SCRAM research page; 18 team profiles live.
  • 6 to 8 hours saved per symposium program through generated session tables.
  • Login account-existence disclosure eliminated.