>>> Notes on reading the code and my very minor clean ups       <<<
>>>				Martin Maechler, Aug. 19, 2002  <<<

R:
==

- Replaced  Platform() by  .Platform
- the functions should not contain  require(tcltk)
  since the whole package should!

  --> removed this and added NEW file R/AAA.R  <<<<<<<<< CVS >>>>>>>>>>>>.

- Style: The functions have long argument names with capitals,
	 even for the first (and sometimes only) argument, e.g.
	 "aList", "toList", "valueFun" ...
  This is non standard (leads to less readable code IMO).


I have some doubts about the real usefulness of several of the provided
utility functions:
<<RG: I agree there is something not right in here >>

o as.character.function(fun) returns the function, not a character
o as.character.formula(fo) is complicated way of using
    format(fo)
    - It does *not* work as a method for "as.character"; since  standard R
      has already defined as.character for formula objects {*differently*}.
    - just wrong concept.

<<RG: I agree - Jianhua we need to fix this stuff up >>

o getContent() for almost all kind of objects returns the object itself
  instead of ``a representation of ..'' (as the help page states).
<<RG: I agree -- I doubt that it is needed in anyway >>

o  hasPrefix() and hasSuffix() seem wrong to me :
   why does
       (hasPrefix("xx"))("Dir/abc")
   give TRUE {it does because there's one "/" in the string}
<<RG: it should not! >>
<<why is the line "length(grep(exclude, x)) > 0 )" there at all?
  just because the string has a file separator does not mean that 
  it has "aPrefix"; this is either wrong or you are confusing two
  concepts >>

o  isCharacter() returns TRUE for an environment {why ??}
		 at least the function name is wrongly designed.
   - also it allows non-character arguments {giving FALSE},
     but most probably should signal an error instead!
<<why is this called isCharacter? why is there a try error? why not 
  simply do the sensible thing of checking to see if you have an 
  environment? >>

<<what is the point of stdType and stdView???? >>>

man:
====

- Some files should really be merged into one.
  * I did so for man/hasPrefix.Rd and
    man/hasSuffix.Rd <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< remove this from CVS !!

- \title{}  should be (almost) all capitalized; should not end in "."
	    shorter: it's a head line!
	    {eg.  don't say ``Functions for ....''}

- \arguments{.}:  The description in the 2nd \item{} entry
		  should start __lower__ case.

- value{} : Don't use \item{} _unless_ you return a list (where you use it
	    for the list components.

- \references{.}  mention R News article properly in only a few places


- Using \describe{...} for widget buttons -- looks much better now, but...
	     but button labels are still not okay.

      - if(interactive()) examples: use  source() {or many example()}
			  to test this interactively

      - stdType() {almost == mode() apart from "call" }
		   seems not too sensical to me.
        If you really really want it, then at least provide it in Biobase,
	so it might be used in much more places for UI consistency.

      - writeDir() really is a general useful utility nothing inherently to
		   do with "Dir"ectory -- hence has wrong name...
  <<RG: I agree -- not sure what it is doing but it is not named correctly>>


      - it seems that
	   listSelect() is the ``only really non trivial''
	function --

