Why did RatSLAM fail?
I recently read about RatSLAM, a SLAM implementation that only needs a simple webcam to map and localize, instead of the usual high precision range finders and encoders.
It even has a ROS implementation, but it's been unmaintained for years and has largely been ignored. Why is this? The biggest headache with SLAM are the expensive and complicated sensor input it requires. I would think any method that could work with any cheap junky webcam would have a ton of applications, even if there's a slight drop in accuracy. What am I missing? Are there any practical limitations of the algorithm that the author doesn't mention?
Academic ROS packages fall by the wayside all the time. Sometimes the developer leaves graduate school and has no motivation to maintain the package. Other times the package is used by many after a process of silently forking and fixing the library. Sometimes the library fails to gain larger acceptance because the maintainer is not responsive to external pull requests. There are a host of reasons why this happens over and above performance concerns.