what is the function naming convention?

On 12-11-04 8:16 AM, Bretschneider SIG-R wrote:

Dear R people,



In typing names of functions (built in or from a package) I often guess
wrong, and have to look the name up.
In other words, I don't understand the logic in naming functions (if there
is any):

R and its packages have been written by hundreds (thousands?) of people, and
they do not use consistent naming conventions.  Perhaps a convention should
have been enforced from the beginning, but it's too late now.

Duncan Murdoch

-  most names are plain, lower case:  cos,  plot,  sapply,  t,  toupper,
unserialize,  (etc)

-  some are capitalized:  Filter,  Machine,  Map,  NCOL,  RNGversion,  T
(etc)

-  many are dotted:  as.complex,  as.data.frame.array,  merge.data.frame,
write.dcf  (etc)

The manual "Creating R Packages" states that it depends on the classes and
instances. I couldn't find more hints.

And there's more:

-  using underscore characters:  check_tzones,  Cstack_info,
R_system_version  (etc)

-  using interCapping:  closeAllConnections,  rawToChar,  rowSums,
toString,  tryCatch,  writeLines  (etc)

-  using dots and intercapping:  as.Date,  julian.Date,  toString.default
(etc)



So, an entire zoo of function names.
Did I miss a system, or is it arbitrary (within the set of accepted
characters) ?
What is the best way to name one's own functions?

Thanks in advance,



Franklin Bretschneider
----------------------------
Utrecht University
Dept Biology
Kruytgebouw W711
Padualaan 8
3584 CH Utrecht
The Netherlands