There are so many different names for what you would think are common descriptors. Just think about molecular weight with all its variations, including “MolWT”, “MW”, “mw”, “molwt”, and “mwt”.
By: Andrew Dalke [ Editor ] Asked
When I’m working with my clients I use their internal names. But now I have an idea for a descriptor calculation system which handles dependencies through a declaration system. I want to implement a few dozen/hundred different descriptors on top of it and I’m a bit stymied because for myself I don’t know what names to pick. (Yes, using free toolkits ;) )
I was thinking about using the Dragon descriptors since they are well defined (what with a book and all), but they seem a bit too terse. “Ss” is “sum of Kier-Hall electrotopological states” – that doesn’t easily jump to my mind, and it seems a waste of a rather short name!
I would like to keep all descriptors in the form [a-zA-Z][a-zA-Z0-9] so they can be used in most programs as variable names, but Dragon has things like “D/D” (distance/detour index) and “R1m+” (R maximal autocorrelation of lag 1 / weighted by atomic masses) and “nR#CH/X” (look it up yourself; functional groups #16). (There’s also plenty of uses of “%” and “()"s.)
Also, I think "nBr” is nicer than “nBR”.
Can you provide pointers to sets of descriptor names and definitions? At the least I could synthesize something from the combination.
For that matter, feel free to describe your standard list of descriptor names.