There's actually substantial value in access to such things. Rather than repeating the experiments yourself, you can go see where they led others without investing the time and energy to repeat the effort. If you find something that mostly resembles what you would have made, can get along with the license, and still wish to move forward - huzzah, a jump starting fork is born.
Abandonware IMHO is more appropriately applied to examples like Google's ever growing list of abandoned projects that actually had users, where the world noticed the abandonment of something useful.