A crash dumper does not have to be implemented with a reboot; this is just something Linux chooses to do. I have actually worked on kernel code that performs crash dumps to disk and over the network, if you find that kind of expertise relevant.