06/03/2004 first pass
11/03/2004 chol, bdsmatrix, bdsI, familycheck, makefamid, makekinship working
12/03/2004 bdsmatrix.ibd,solve.gchol,solve.bdsmatrix working
14/03/2004 gchol(), solve(gchol.bdsmatrix) working
15/03/2004 debug coxme.fit
16/03/2004 place %*% in UseMethod working but overwrites system's
18/03/2004 use '%*%.default' to get around the problem of overwritting system '%*%'
19/03/2004 plot.pedigree working
21/03/2004 inform changes to Terry Therneau
23/03/2004 separate dmatrix/chsolve2 from bdsS.h
26/03/2004 force bdsmatrix to be a matrix but not a list, customize as.vector(bdsmatrix)
27/03/2004 gchol now gives gchol.bdsmatrix class if argument is bdsmatrix
28/03/2004 R CMD check kinship!
31/03/2004 confusion with full=T in solve.gchol, solve.bdsmatrix remains
06/04/2004 the use of get("%*%",pos=NULL,mode="function")(x,y) makes complaints gone
11/07/2004 fix problem in lmekin due to optim
25/07/2004 further fix on optim in coxme.fit, so no more complaints
25/08/2004 add keywords as required by R 2.x
27/08/2004 add version switch for R 2.x as required by setClassUnion("list or NULL",c())
31/08/2004 remove <s-expression> contained in title of coxme.control.Rd
01/09/2004 return of is.matrix.* and is.list.* for bdsmatrix,gchol.bdsmatrix, change eval()
02/09/2004 document is.*.*
03/09/2004 replace sys.parent() with parent.frame() in coxme/lmekin; uncomment !is.matrix in coxme.fit

1. documented or obvious

.h                      <- Sint and other changes
nlminb                  <- optim
.C()                    <- add PACKAGE = "kinship" as required by R 1.9.0
.First.Lib              <- add nlme, survival, methods
significant:            kinship.so and kinship/R/kinship directly debugged

2. sequence of statements

BDSmatrix.R             <- bdsmatrix.R
BDSmatrix_%_%.R         <- %_%.bdsmatrix.R

3. amend disturbing statements

UseMethod               <- add setGeneric
list or NULL            <- setClassUnion          
unique                  <- add 'missing' to 'bdsmatrix'
setMethod('diag',       <- add nrow=1,ncol=1
setMethod('['           <- add i, j
gchol.bdsmatrix.R       <- omit ... in setMethod('dim') and setMethod('show'

thus ... in the function arguments are dropped


4. other changes

assign                  <- envir as specified in R-FAQ
nDotArgs                <- length(list(...)
object                  <- from in places
diag                    <- setMethod("diag", "diag<-"
gchol                   <- as.matrix()
solve.gchol.bdsmatrix   <- temp@ <--> $
par()$"1em"[2]          <- par()$"cin"[2]/7
plot.pedigree           <- comment on segments at 118,119,124
align.pedigree          <- modify remove()
