Skip to content

Latest commit

 

History

History
40 lines (36 loc) · 2.49 KB

TODO

File metadata and controls

40 lines (36 loc) · 2.49 KB

##-*- org -*-

formally deprecate repChar() and bl.string() in favor of R (>= 3.3.0) strrep()

maybe add deparse1seq() from ~/R/Pkgs/DPQmpfr/R/dpoisEr.R ?

more to sessionInfoX(); chk openBLAS etc, see more moreSessionInfo() in robustbase,

~/R/Pkgs/robustbase/inst/xtraR/platform-sessionInfo.R

finish and export pkgLibs() R/unix/package-props.R

New utilities for MASS::boxcox() “diagnostics”, notably histogram nicely labeled

–> ~/R/MM/Pkg-ex/microbenchmark/seq-mb.R

Maybe ask Brian about willingness to improve: boxcox(*, plotit=TRUE) looks nice,

but the estimate + conf.interval computations are inside the plotting, not available as numbers

Better ylim for boxplot(<microbenchmark>), see, ~/R/MM/Pkg-ex/microbenchmark/simple-ex.R

mat2tex() is really not flexible enough; and there are better functions

in other packages, e.g. latex in ‘Hmisc’ –> deprecate this! [2005-04]

str_data() could become more flexible (class = “.”;

also (optionally) return class + dim()/length(); see ‘TODO’ in R/str_data.R

is.whole() R/misc-goodies.R conflicts with gmp::is.whole() – as it has a ‘tolerance’

move some demo/*.R to vignettes/*.Rmd – notably the prime-numbers (for those, see also below)

clean up R/prime-numbers-fn.R – move all non-exported functions to

(a new file) ./inst/primes-extra-fns.R

and source() that from ./demo/prime-numbers.R !!

Reverse Dep. status: 2014-06-16 http://stat.ethz.ch/CRAN/web/packages/sfsmisc/

Reverse depends: ascrda, catIrt, distr, ICEbox, lordif, modiscloud, plfm, polycor, random.polychor.pa

Reverse imports: CDM, Demerelate, FrF2, lokern, mcmcplots, ReliabilityTheory, simsalapar, sirt, TAM

Improve ps.end(): should return something useful

==> new ./tests/psend-ex.R which tests that 2 x {ps.do() … ps.end()} does work properly

eaxis() [ ./R/prettylab.R ] :

allow log2-scale and “2^…” labelling — or log_k and ‘k^..’

a range where the defaults of eaxis were completely bad

because of axTicks(*, log=TRUE) can be bad: curve(qgamma(1e-100, x, lower.tail=FALSE), 1e-110, 1e-70, log=”xy”, col=2, axes=FALSE) eaxis(1);eaxis(2)

Work on R/approx2.R, for now approx2.R – implement remaining “vectorized” part

2017-08-xx: in DESCRIPTION, use Authors@R with Werner, Andreas, .., Alain, ..

2008-10-22: Copied boxplot.matrix() to R-devel (2.9.0 to be)

Remove this from ‘sfsmisc’ eventually.