> the onus really should be on the user to record what version of software they're using and how they configured it. Recording version, etc. is fine, but what

> the onus really should be on the user to record what version of software they're using and how they configured it. Recording version, etc. is fine, but what use is it if it's not actually possible to install that software to reproduce whatever results were obtained? (This also extends to hardware, btw., but VMs can probably attain enough fidelity for that to not be an issue as long as the physical media are still readable.) Example: Reasearch done on Windows XP at some point in time past SP3. Unless you're prepared well ahead of EoL on Windows XP there's really no way to reproduce that environment. (I'll definitely grant that it's probably unlikely that the obscure details of the OS have that much impact on whatever results were obtained by programs running atop it, but it's at least moderately plausible.)