?if

I am trying to introduce a class of senior undergraduates to the values and 
advantages of R for their statistical needs.  Unfortunately, little (and 
inconsequential) inconsistencies such as these do little to improve their 
confidence in the application.

You could tell them that the argument to ? (and to apropos) should 
*always* be quoted, which is internally consistent and is consistent with 
other functions, where strings do have to be quoted.  Unfortunately the 
documentation has examples with bare names, but most people don't read the 
documentation.

The fact that help() accepts bare names is a bit of a wart -- for example,
   text<-"lm"
   help(text)
which is annoying when programming.

 	-thomas