[R] Re: R web interfaces

[Nathan Whitehouse]

cashaw at bcm.tmc.edu, 
    youngas7 at yahoo.com
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

Hi,

  First, does anyone think it is a good idea to put up
a web server site similar to Phillippe Grosjean's site
for GUIs?  

  To answer foobar's question-

  There are a several web 'interfaces', with different
focuses and degrees of sophistication.
  Apologies to the developers if I characterize them
wrongly.  They are all tailored to slightly different
purposes.

  CGI/(transaction based) -
       CGIWithR, RWeb & RCGI(unmaintained)

  TCP/IP Servers(sessioned)-
        RServe
       -RServe has a homegrown protocol for large
datasets

  Web/HTTP Servers(sessioned)-
       RZope/RStatServer
        -A Zope/Python/RPython facility for executing
scripts/application development

       Rho
        -A servlet application/application development
suite to selectively expose R code & manage
datasets/projects 

  There are three or so different things to consider
with these:
    1)How R-based processing is made available or
whether the client supplies them(like RWeb)
    2)The extent to which a framework/system is an
'application development system' as opposed to 'a way
to submit R code on the web'
    3)How R is running
        a)spawned from apache using CGI- one
spawn/transaction 
        b)an R process on the server that takes up
requests in a queue
        c)a set of processes that are spawned
dynamically and communicate through SOAP/CORBA etc.
These may exist on multiple internal computers.

   Although there are other issues, which we'd be
excited to talk to somebody about at UseR.

  Best,
  Nathan Whitehouse

=====
Nathan Whitehouse
Statistics/Programming
Baylor College of Medicine
Houston, TX, USA
nlwhitehouse at yahoo.com
work: 1-713-798-9029
cell:    1-512-293-5840

http://rho-project.org: rho- open source web services for R.
http://franklin.imgen.bcm.tmc.edu: Shaw laboratory, bcm.