[R] R as a server in client server computing

[Duncan Temple Lang]

> Reply-To: <jens.oehlschlaegel-akiyoshi at mdfactory.de>
> From: "=?iso-8859-1?Q?Jens_Oehlschl=E4gel-Akiyoshi?=" <jens.oehlschlaegel-akiyoshi at mdfactory.de>
> Date: Tue, 27 Jun 2000 14:49:59 +0200

> I like to have a continuously running R process, which can receive a
> dataframe from a client (over TCP/IP), does some processing, and sends some
> data back. What is the prefered way to do this? Using the socket interface?
> Using omega's CORBA stuff?
> Does anyone has example code for doing so?
> 
> Thanks for any help
> Regards
> 
> --
> Dr. Jens Oehlschlägel-Akiyoshi
> Analyse


Hi Jens,

Well, I think it depends on several factors. For example, 

a) what platforms are the client and server running on?

b) what language is the client written in?
   Is it R <-> R communication? or
   C <-> R
   Java, Python, Perl, ... <-> R

c) Is the set of data types and functionality known at the development
time of the client, or will it evolve? or is a quasi-generic interface to
R more likely?

d) How stable are the design and implementation plans and is it likely
that you will chose different languages for the client.

  
e) If you were to use the socket interface, what would the content of
the data communication look like?  XDR, the raw ASCII dump, XML?

 One of the good things about the CORBA and DCOM approaches is that
while there is an up-front cost of configuration and installation
(which is not really prohibitive), one then avoids low-level details
of communication and can leverage tools to query the current state of
the system.  This is especially important when things aren't working,
especially in the early stages. Debugging in a distributed environment
is perhaps the most challenging aspect of this entire endeavor and the
higher the level each node in the communication, the better.


 Perhaps if you suggest likely answers for a) - e), it might 
help to narrow the focus.

 D.


-- 
_______________________________________________________________

Duncan Temple Lang                duncan at research.bell-labs.com
Bell Labs, Lucent Technologies    office: (908)582-3217
700 Mountain Avenue, Room 2C-259  fax:    (908)582-3340
Murray Hill, NJ  07974-2070       
         http://cm.bell-labs.com/stat/duncan

      "Languages shape the way we think, and determine what 
       we can think about."        
                                      Benjamin Whorf
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._