News organizations should care more about link rot

One of my favorite — if not my favorite — ironies I've experienced working in journalism is when a media company invests as little as it can in a content management system. No matter that this piece of software literally sits at the heart of how we make our money, let's pour our cash into the cheapest option!

There's a lot of things to get mad about in a bad CMS — limiting templates, poor custom Open Graph support, core options buried in sub menus, confusing domain-specific language, terrible print-digital workflows, impossible to find old stories. But what grinds my gears the most is how common it is for the CMS to not give a rip about the content itself.

My first experience with this phenomenon occurred after Northwestern University decided to upgrade the Medill Reports site. It went from some CMS I don't remember to WordPress (I think?). The switch meant all old articles were gone. Every story I wrote was trashed. The administrators were kind enough to let us know so we could archive what we wanted.

After a certain date, I would never again be able to read the mediocre-to-decent news stories I wrote in school 😔. That was fine, though, because I didn't need them. Where this was much more painful was a few years later after I left The Virginian-Pilot.

Early on in my time at the paper, we went from a homegrown Drupal fork to signing a contract with Town News. That changing of the guard torpedoed almost everything previously published. The few custom pages I had built just sat in a broken state — javascript disappearing or referencing missing assets only God knew the whereabouts of. Even more painful was a decade of stories stripped of any metadata now floating in a search API abyss.

There are plenty of excuses, and an enterprise CMS isn't built to talk to an in-house iteration of an open source platform. Fine, I guess.

I was at The Pilot for 2-2.5 more years. In my final months, the paper was bought by honky Tronc with rumors of switching to The Washington Post's Arc Publishing system.

Initially I had cautiously optimistic thoughts since moving to a CMS run by another newspaper could go somewhat smoothly. By this time, though, I had built dozens of custom elements and pages. My entire resume relied on the Town News ecosystem (specific CSS classes and rules, width of article templates, custom HTML placement, etc). Switching could immediately render all of those projects obsolete.

Spoiler alert: It did.

I felt like I was in a race against link rot. If I wanted to look for another job post "upgrade," I could have to rebuild most of my portfolio. I loved my job at The Pilot and was fortunate enough to have been given a lot of space to explore and experiment. Some stuff worked and some stuff didn't, but it was a unique circumstance that probably wouldn't be replicated elsewhere — a compounding issue which meant my new clips may never be as cool as my old ones.

Things ultimately turned out all right, and the destruction of everything I worked on in Virginia didn't come back to bite me as much as I feared. (It did create some hilarious bugs). I feel lucky in that regard. Custom elements wiped out by Arc still blew up and ruined the content they were embedded in. Sometimes those were pretty big stories. All of them affected more than just me.

I could solve a sliver of my problems by just recreating these projects on my personal site (either in code or in screenshots) — something I've thought of doing numerous times. But it fixes like 5% of the issue.

Each CMS change I've been a part of draws a line in the sand – effectively saying anything before this point is lost. Not only does creating a useful archive for readers (and journalists) get that much harder, but the history represented in those articles is another step removed from public knowledge. And I think that matters a great deal.