The editorial in the August issue of Nature Methods discusses an issue that comes up when computational biologists—or anyone else for that matter—wants to report a novel algorithm that biologists may want to use in their research. Specifically, whether or not to supply a named software implementation of their algorithm that biologists can use.
As part of our standard material sharing policy, Nature Methods generally requires that authors provide a useable software program implementing any new algorithm that is integral to a method they’re reporting. But we have never said anything about naming the software.
It recently came to our attention that there are a number of factors that act to discourage authors of new algorithms from naming a software implementation of their algorithm. As discussed in the editorial, this can lead to difficulties later on and in many cases providing a name for the software has benefits that outweigh the potential hazards. Read the editorial and then let us know what you think.