Pull dsBase package from github to your machine - this is all the server side datashield functions for base package
Pull the dsBaseClient from github to your machine - this contains all the client functions
Write R script for server function named in the conventional way xxxDS.R
Write an R script for the client function named in the conventional way ds.xxxxx.R
Place the server function script into the correct folder pulled from github e.g. dsBase/R/ on your own machine
Place the client function script into the correct folder pulled from github e.g. dsBaseClient/R/ on your own machine

Put the header at the top of the client R script (ds.xxxx.R script)

Example Header block

#' 
#' @title [insert something] 
#' @description [insert something]
#' @export 
#'

Example 2

#' 
#' @title creates random numbers 
#' @description This function is similar to R function \code{rnorm}.
#' @export
#'

On each Opal server log into the web interface and click on Administration > DataSHIELD and select either ASSIGN or AGGREGATE (depends on your function and what it does!)
Click +Method
Add a name of server side function "xxxxx.DS"
Select R script
Paste the server function code from your xxxxx.DS.R script into the empty box
Click save

Reload the new client library from your local machine. In R:

library(devtools)
devtools::load_all("filepath to dsBaseClient you have pulled from github on your machine")

Test the new function:

#load libraries
library(opal)
library(dsBaseClient)
library(dsStatsClient)
library(dsGraphicsClient)
library(dsModellingClient)
 
#log into to the test opals
data(logindata)
opals <- datashield.login(logins=logindata,assign=TRUE)
 
#run the function to test it
ds.rNorm1(samp.size=25,mean=0,sd=1,newobj="newObject",datasources=NULL)