-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Cyrus Harmon has used the same approach to interface R to Lisp
and I recently discussed this with him. He has gotten a long way.
I have developed a "lot" of inter-system interfaces and I
hadn't really considered going this route, i.e. interfacing
to the C routines in the R API or further still those
not necessarily "published" in the DLL. I tend to work
at higher level with a very thin engine that
connects the two systems. Instead of calling,
for example, Rf_parse() (or whatever) and the Rf_eval(),
I create a very general C routine that can call an arbitrary R function
and with arbitrary arguments. Then I can call the R function parse().
Then I can also call the R function eval(). And the nice thing about
this is that error handling is done correctly.
If you call Rf_parse() at the C-level, where will you return
to if there is an error.
I have waited to interface to .Net until there appeared to be a need
for it. But I had planned on using the approach for interfacing to
other systems, i.e. using the thin engine.
I will be interested to compare the relative merits of your approach.
You have access to every aspect of R and can build the interface
in either language. So in principal, you can do everything in
a variety of languages.
D.
Brian Vesperman wrote:> I am going to attempt to develop a wrapper for the R.dll library in the
.Net
> framework.
>
>
>
> I successfully interfaced with some very simple functions from the library,
> but nothing close to do some data analysis.
>
>
>
> I dumped all the "publicly" available functions and attributes
from the
> windows dll and found there are 2671 functions and attributes.
>
>
>
> I created an API document that just has the function signatures and what
> code file I found the function in; I have links to the file below:
>
>
>
> http://valueadded.wceruw.org/LinkClick.aspx?link=R-API.pdf
>
<http://valueadded.wceruw.org/LinkClick.aspx?link=R-API.pdf&tabid=36&mid=377
>
>>&tabid=36&mid=377 PDF file
>
>
> http://valueadded.wceruw.org/LinkClick.aspx?link=R-API.doc
>
<http://valueadded.wceruw.org/LinkClick.aspx?link=R-API.doc&tabid=36&mid=377
>
>>&tabid=36&mid=377 MS Word life
>
>
>
>
>
>
> If this information is useful to anyone else please contact me. If there is
> any interest to expand this document, I might setup some sort of help file
> Wiki where people could add to the documentation.
>
>
>
> If anyone would like to join my effort please contact me. Hopefully it
> doesn't become too difficult and I will be able to make some headway.
>
>
>
>
>
> Brian Vesperman
>
> Associate Information Processing Consultant
>
> Wisconsin Center for Education Research
>
> RM 772
>
> 1025 W. Johnson St
>
> Madison WI, 53706
>
> 608-265-5623
>
>
>
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
- --
Duncan Temple Lang duncan at wald.ucdavis.edu
Department of Statistics work: (530) 752-4782
371 Kerr Hall fax: (530) 752-7099
One Shields Ave.
University of California at Davis
Davis, CA 95616, USA
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (Darwin)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQFDRZNC9p/Jzwa2QP4RAl/MAJ0axT/rpc36jvcoInScMd+nXYQ8WQCdEjVB
d9aD4Yrc/5m5Twbu8n//n2o=gQdX
-----END PGP SIGNATURE-----