[
1S1o‘s CHEMTRAN
‘“ “Compatibilizes”
Fties of Encoded Chemical Structures
I
“Compatibilize”
and CHEMTRAN
are new words. Surprisingly,
“debabel-
ize” is not. Unfortunately
it not only
disparages the useful diversity of chemi-
cal “languages”
and codes,
but also
suggests
that CHEMTRAN
is another
chemical language. It isn’t. CHEMTRAN
aspires to become auniversal interpreter
of and between
various existing
sys-
tems–hence,
it compatibilizes.
The origins of CHEMTRAN
go back
to my original
interest
in
this pro-
blem.1-z Recently,
this culminated in
the successful completion of computer
programs which convert Wiswesser Line
Notation
into so-called “Dokumenta-
tionsring”
fragment codess by means
of
computer-generated
connectivity
tables.q These connectivity
tables pro-
vide the basic intermediate records re-
quired not only for the WLN-to-Ring
conversion, but also for conversion to
any other fragment code, as e.g., the
DuPont codes used for indexing chemi-
cal patents.
CHEMTRAN
can also
be used
to generate other connectivity
tables,
such as those used by Chemical Ab-
stracts. CHEMTRAN might be com-
pared to a common
computer
lan-
guage such as Cobol. Through “com-
pilers” Cobol compatibilizes
various
computers. Unlike Cobol,
interestingly
enough, CHEMTRAN
does not
in.
crease the computer running time re-
quired
to solve searching problems
November 15, 1972
>n the contrary,
the generation
of
“ragment codes or “screens” simplifies
earthing
enormously,
since
the in-
dividual fragment code, like a standard
:omputer
instruction,
pre-programs
)therwise
complex
programs.
In a previous series, 6-9 I described
why chemists,
pharmacologists
and
>thers need to search by substructures,
fragments, moieties,
etc. Space does
not permit me to repeat those basic
notions here. The diversity of problems
. .
:equwlng
such a capability
accounts
for the incredible variety of systems in
the literature. Tire irresistible intellec-
tual challenge
of
designing a new
chemical code must also be a factor,
since in so many cases existing methods
wou