Maintaining Old Work

Jye Sawtell-Rickson · February 23, 2025

“We will update our paper list every two weeks and include all the following papers in the next version of our paper. Please Feel free to contact me in case we have missed any papers!” - (Guo et al., 2024)

I’ve seen a lot of research papers where authors claim that they are releasing a benchmark which will be continually updated and stand the test of time, or they’re creating a survey paper which will be updated with the latest papers.

Very few stay true to this. Very few. The benchmark is quickly beaten with a high score and then never seen again. Or a research paper list is maintained for a month and then the authors move on to other things.

One positive example that comes to mind is the ARC Prize which was released in 2019 and continually held prizes and has a second version coming in 2025, and a third version planned. Another classic example is ImageNet which has various competitions and various related datasets constructed.

I’m not expecting people to devote their lives to maintaining their previous work. One of the amazing things about practicing Science is our ability to learn new ideas and quickly move on from the ones that don’t still contribute value. I’d just ask authors to be more honest about this. If a project was a massive amount of work but will be static in time, then state that clearly. It doesn’t make the work any less valuable.

Twitter, Facebook