[R] package under unix

Patrick Connolly p_connolly at slingshot.co.nz
Sun Jun 15 22:51:04 CEST 2008

On Fri, 13-Jun-2008 at 04:51PM +0200, cgenolin at u-paris10.fr wrote:

> Hi the list,
> I write a package for clusterizing longitudinal data using a non parametric 
> algorithm. I develop the package under windows. To be as user friendly as 
> possible, the package use some graphical procedure to "show" to the user 
> the evolution of the cluster construction, and to export the graph in a 
> friendly way.
> Here are some example : http://christophe.genolini.free.fr/kml
> Everything works fine... under windows.
> Unfortunately, it seems it does not work under linux. I first use the 
> instruction:
>> windows(5,5,xpos=0)
> which seems to be incompatible. Then I used :
>> if(getOption("device")=="windows"){windows(5,5,xpos=0)}else{}
> but it is non portable either.

Why not let users set the device they wish to use?  There's no need to
have it hard coded.  In fact, that's why I don't use certain (for now,
unnamed) packages because of that very issue.  Requests to the
maintainer gets no response.

> I do not know linux so it will be very hard for me to test and
> change my code.  On the other hand, I spend a lot of time to develop
> a graphical interface for exporting the result in a easy way, so it
> would be a pity to remove the code that deal with graphics.

Generally, I like to use a pdf device which can have any number of
pages.  There's no need to start a new device for each plot.  Windows
users generally don't work that way and are unaware of its advantages.
But an X11 device will also work for single plots and there's no need
to specifically set it in the package.

> Can someone help ?


   ___    Patrick Connolly   
 {~._.~}          		 Great minds discuss ideas    
 _( Y )_  	  	        Middle minds discuss events 
(:_~*~_:) 	       		 Small minds discuss people  
 (_)-(_)  	                           ..... Anon

More information about the R-help mailing list