On a benchmark of 250251 compounds in the NCI00 database, mol2nam
is able to convert 223066 structures (89.14%) to names without BLAH.
Of these 223066 names, nam2mol is able to convert 192487 (86.29%)
back into structures.
This release includes a significant number of improvements to both
name generation and name parsing. For example, nam2mol now supports
more numbered locants, such as ‘N1-methylaniline‘ and for
‘Maybridge-style’ locant names such as N'1 (interpreted as
the more common N1'). These and similar changes have
increased the conversion rate for the 71367 compound names in the
2003 Maybridge catalog, from 69.51% in v1.4 to 80.80% in v1.5.
This release includes the ability to generate compound names in Japanese,
and much improved Spanish and Polish naming support. In order to better
support internationalization, APIs are now available to map from the
default ISO-8859-1 output to either 7-bit ASCII, UTF-8, HTML and for
Japanese locales, Shift-JIS or EUC-JP.
Although impossible in the general case, several improvements have
been made to Lexichem‘s compound naming such that the assigned names
are now more stable under arbitrary input ordering of atoms and bonds.