A (non-Swiftian) modest proposal

Update, March 2021: They did it!…sort of. DDOT did get a brand new shiny website this winter, but now none of the old project links work. Try searching for U Street Streetscape—even the search results’ links don’t work.

Borderstan’s still around, though….

DDOT needs an entirely new website, from the ground up.

This is admittedly not DDOT’s top need, nor even likely in the top 5. But it’s important.

For starters, DDOT’s practice of buying separate domains for each project needs to end. This is problematic for (at least) two reasons—it makes it difficult to tell what’s actually an official project site, and it means that either the District has to keep paying registration renewal fees or they disappear.

Take, for example, the 2012–3 U Street Project. Borderstan.com still exists, despite that site having gone out of business three years ago, but UStreetProject.com doesn’t. It’s no longer registered at all. (DDOT should probably consider themselves lucky it hasn’t been squatted by a porn site or something…)

Meanwhile, DDOT’s internal hosted page for the U Street Project, https://ddot.dc.gov/page/u-street-nw-streetscape-project, ends with a “Related Documents” link to a numbered ‘node’ rather than to a named page: https://ddot.dc.gov/node/463932. Though the U Street Project page says it leads to “Rehabilitation of U Street NW from 9th Street to 14th Street NW”, it actually goes to …[drumroll]… “Oxon Run Trail Rehabilitation Project - Appendix A”.

Not only is the broken link a problem in itself, the fact that the DDOT (and, I suppose, entire DC-dot-GOV) website infrastructure is built on numbered database nodes that only sometimes have names makes it impossible to track whether links are going to go where you think they will until you’ve actually followed them—which means the site is not only difficult to use, but difficult to maintain, which I’ll come back to.

It can also be difficult to find things in the DDOT website, especially if you only vaguely know what you’re looking for or what it might be called (doubly so if there are multiple ways to abbreviate or jargonize a term). There’s not much of a logical structure, and the DC.gov search function…let’s just say has interesting ideas of string matching and result relevance.

Now, having never been inside let alone worked on DDOT’s website infrastructure, I can only guess at what it takes, but having maintained a few websites—in and out of databases—I can guess it’s at minimum a giant pain in the ass. I can imagine that reconfiguring it would be…problematic.

So, better instead to blow it up and start over. A couple of suggestions: Give things real names. Have a defined structure with sensible levels, so people can find things. And for the love of god don’t just break every existing link and put a search bar on the 404 page; use Apache’s exquisitely powerful redirect infrastructure to make sure people can get where they’re going.

A bit late to get it going for the new decade—given the lead time on such things, probably a bit late even to have it up for the new year…but maybe if they start now, they can have it ready in time for the next Mayoral term to start in 2023.