------------------------------------------------------------
Remarks by Martin Maechler, Jan.-Feb.2002,
when first porting the Dec.1999 code from S-plus to R
------------------------------------------------------------

1) Question about naming :

	 Lasso2/Doc/Manual.ps
	 clearly talks about "Lasso2 : A S-PLUS library ...."
   As Berwin explained my, "2" is not a version number, and hence we won't
   have lasso3, hence we keep it (lower case).

2) I think it's a waste of space to have the GPL statement in every file.
   Would like to have it only in very few places if not one ("COPYRIGHT")
   alone {this is now done in many parts}.

3) Most man/*.Rd needed \keyword{} added.
   (2.1-0 : many still need!)

4) src/lasso.c and lasso.h
   -----------
 a) has many printf() -- this is not ok in R (think of Win Rgui);
     in R we have Rprintf() {and REprintf()} etc.
 b) contains its own QR code ( = last third of lasso.c)
    this is `bad' since we would rather want to use
	 LAPACK which might available in a much faster optimized version.
    I'd rather separate this part out.
 c) I've  replaced `long' by `Sint' and
    in lasso.h defined Sint (as include <R.h> would for R)
    IMPORTANT: replaced S_Plus by S_or_R  (both lasso.h & lasso.c)
	                ~~~~~~    ~~~~~~  ==> Berwin: change lasso.nw !

5) R/l1ce.q : (line 267) : I had to add " <- X.to.C.stds "; the
   if( standardize ) clause wouldn't have made sense otherwise!

  5b) similar bug fix in R/gl1ce.q

6) vcov() --- same as vcov() in MASS  (but ..)
	  --- or should we require MASS since that is now recommended, i.e.
	      available wherever R is.

	  --- or should  "base R" define the vcov(.) generic ?

7) man/*.Rd : There is now [Version 1.0-1, 2002-02-06] only one warning left :
	    - about 3 undocumented objects:
	      "force.loading" "opt.covmat"    "tib.covmat"

   man/aux.*   are renamed to
   man/zaux.*  since Brian Ripley remarked that filenames  AUX.*
	       are close to invalid on Windows.

8) I've started ./ChangeLog -- to document my changes a bit more
	        ~~~~~~~~~~~

Martin Maechler <maechler@stat.math.ethz.ch>	http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum  LEO C16	Leonhardstr. 27
ETH (Federal Inst. Technology)	8092 Zurich	SWITZERLAND
phone: x-41-1-632-3408		fax: ...-1228			<><
