At the moment, batch tonemapping takes x settings-files and y HDRs and creates x*y output files.
Wouldn't it be better to use a system akin to that of ufraw, so that you have one "id" or "settings" file
for each HDR, which stores which gamma, tmo, options and output size to use?
If you have a bunch of HDRs, all taken under slightly different conditions, you have to experiment
with which tmo to use and what options, so most likely you end up with one settings file
per HDR anyway. If you then use batch tonemapping you create a lot of unnecessary files
that you have to sift through afterwards.
As an example: Say, you have ten HDRs and ten different settings, then batch-processing gives you
a hundred LDRs out of wich only ten were needed. That takes a lot of processing time and manual
interaction afterwards and somehow defeats the purpose of using batch processing in the first place.
I can see how both ideas could work. But to make them coexist... Now that's a task for a real usability architect :)