Consistency of n for atom properties

[vfscalfani]

There is currently a variety of inconsistency with n across atom properties and rings within the specification:

Hydrogens: Hn, where n is allowed from 0 to 9.
Charge: +n or -n where n is allowed from 0 to 15
Isotopes: n can be 0 to 999
Atom Class: n can be 0 to 9999
Rings: n can be 0 to 999.

Note that in all cases leading zeros are not allowed such as 01, so that part is consistent. I think it makes sense to make n consistent across everything. So, why not have n for all atom properties and rings be from 0 to 9999? Thoughts?

[merkys]

Personally I am fine with having "chemically-sane" limits. I do not think there is currently a need for three-digit hydrogen count, or four-digit charge and isotope number. But atom class and ring bond number indeed benefit from larger limits.

[vfscalfani]

For the consistency of n for atom properties, here is what we decided to do for now:

Hydrogens: Hn or Hnn (max 2 digits)
Charge: +n / -n or +nn / -nn (max 2 digits)
Isotopes: nnn (max 3 digits)
Atom Class: nnnn(max 4 digits)
Rings: n, %nn, %nnn, or %(nnn) notation. (max 3 digits)

In all cases, leading zeros are not allowed.