ORAD taught a generation of dentists how to read. It is a teaching resource in oral and maxillofacial radiology, built by Dr. Stuart White, and for years it was where you went to see a finding described properly and matched to an image. The content is still good. The technology around it aged, the way technology does, and at some point a resource that everyone respects becomes a resource that fewer people can actually use.
I am rebuilding it with Dr. White. The interesting part of that work is not the code. It is deciding what not to change.
The temptation is to start over
When you inherit an old site, the reflexive engineering answer is to throw it away and rebuild from a clean slate. That answer is usually wrong for something like this, because the value was never in the platform. It was in the cases, the descriptions, and the specific judgment of the person who assembled them. If a rebuild loses that, it is not a rebuild. It is a different, emptier thing wearing the same name.
So the rule I set for myself was simple. Keep everything that made ORAD worth using. Replace only the parts that make it hard to use. In practice that line is harder to hold than it sounds, because modernizing the delivery keeps tempting you to also modernize the content, and the content does not need it.
What actually needed fixing
The problems were mundane and real. It needed to work on a phone, because that is where students read now. It needed images that load quickly and hold up when you zoom, because a radiology teaching resource that shows you a fuzzy thumbnail is failing at the one job it has. It needed a structure you can search and navigate, so that a student looking for a specific entity can find it without knowing where it was filed a decade ago. And it needed to be maintainable, so that the next person who wants to add a case can do so without excavating the old stack.
None of that touches the teaching. That is the point. The modern web part of the project is entirely in service of getting old, good material in front of people on the devices they actually hold.
Preservation as an engineering goal
I have started to think of this as a preservation project that happens to involve code. The instinct in software is to treat everything as replaceable, because most of the time it is. A teaching archive is the exception. The images and the descriptions are the asset, and they get more valuable, not less, as the people who made them move on. Building the new site well means building it so the content outlives this rebuild too, and the one after it.
There is a version of this work that is flashy: new features, a viewer with every bell, an interface that photographs well. I keep steering away from it. What ORAD needs is not to be impressive. It needs to load, to be findable, to look right on the screen in a student’s hand, and to still be there in twenty years. If I get those four things right and stay out of the way of the content, the rebuild will have done its job.