[R] Environmental problems.

[Rolf Turner]

On 26/02/14 00:56, Enrico Schumann wrote:

<SNIP>

>
> I did not follow your example, neither do I use the deSolve package; but
> why not pass an environment as an argument?
>
>    ## some iterative function that takes another fun as argument
>    outer <- function(fun, ...) {
>        for (i in 1:20)
>            fun(...)
>    }
>
>    ## create an environment ...
>    info <- new.env()
>    info$tt <- 0
>
>    ## ... and pass it as an argument
>    myfun <-function(e) {
>        e$tt <- e$tt + 1
>        cat("Iteration ", e$tt, "\n")
>    }
>
>    outer(myfun, info)
>    info$tt
>

Sounds like a good idea.  Tried it.  Got exactly the same phenomenon as 
with my idea.  The "tdone" variable stays equal to 0 and there is no 
progress report (as desired) until time gets up to 0.1.  Then tdone gets 
re-set to the value of tt *every* call, so that tt - tdone is always 
equal to 0 and there are no further progress reports.

I have attached a tar-ball and a zip archive both containing the 
necessary code to demonstrate the forgoing.  Unpack, source all the *.R 
files and then source "demoScript" to see it happen before your very 
eyes. :-)

cheers,

Rolf Turner