Re: [S] Quick notes on using ESS

A.J. Rossini (rossini@biostat.washington.edu)
12 Oct 1998 10:17:57 -0700


(forwarded to the list, since this might be of general interest)
"Jeff Longmate"<jlongmate@smtplink.Coh.ORG> writes:

> Since comments are welcome:
>
> 1. Thanks.

You are welcome!

> 2. I'm using NT now so maybe there's no help for me, but I've
> recently taken up including S transcript segments as comments in
> latex documents. I was curious if there is any way of dealing with
> that sensibly, even in ESS on unix.

I use Noweb, which is a language-independent literate programming
tool. That, plus a noweb-mode for XEmacs, allows me to edit and
submit code, while documenting and indexing variable and function
definitions. The whole package isn't as global or robust as possible
(i.e. Noweb allows code chunks within chunks, but I've not finished up
the code to deparse appropriately).

For example:

\LaTeX code to discuss how to obtain estimates for
$\hat\beta$, which can be obtained by

<<Estimating Equation for \hat\beta.S>>=
<<find starting values for \beta.S>>
results <- ee(beta)
<<summarize \beta from estimating equation.S>>
@ %def results beta

which provides the following estimates (more latex describing
the output).

and, using C-c C-n on the "results" line, submits the code to the
desired process (but currently that must be specified, sigh...).

The latex is edited using AUC-TeX and the code is edited using ESS
(S-mode), and one just needs to have SPlus running in a separate
buffer.

The editing and document production works (except for process
communication, sigh...) on NT and W95, since one can get both XEmacs
and Noweb running under NT.

For XEmacs, one needs to look at the very recent betas (for 21.0...),
http://www.xemacs.org/
For Noweb, see:
http://www.cs.virginia.edu/~nr/noweb/

For Process communication under NT, it'll be about a year before we
integrate it (unless someone buys me a fast NT desktop with enough
capacity for developing, or someone else does most of the work). Some
work (extremely alpha) has been done by others.

For ESS mode chunk recovery (i.e. being able to dump chunks into a
running process, properly replaced with the actual S code), that's
coming soon...

best,
-tony

-- 
A.J. Rossini			
Rsrch Asst. Professor				Senior Biostatistician
Biostatistics, University of Washington		Epimetrics Corporation
Center for AIDS Research
rossini@biostat.washington.edu			rossini@epimetrics.com
http://www.biostat.washington.edu/~rossini/	http://www.epimetrics.com/
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu.  To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message:  unsubscribe s-news