[R] hblm program: porting from Splus to R

[David Rindskopf]

I am replying to the email below my signature;
it is nominally about dmp files, but apparently
specific to problems with Bill DuMouchel's
hblm program.

I have been using this program
on Splus for years, and ran into some of the
same problems Jack did in trying to get it into
R.  I talked with Bill about it, and he didn't
know that there would be any problems; he
assumed that R was functionally identical to
Splus, and that the program should run.
Unfortunately, of course, that's not true.
I'm too much of a novice to figure it out, but
at least some problems seem to involve
scoping rule differences.  The NA coercion
apparently is something else, perhaps a character
being coerced to numeric?

As for legal issues, I can try to contact Bill
to be sure, but it seems obvious he intends the
program to be in the (noncommercial) public
domain.  The program remains on the AT&T
web site years after he left, and he seemed
perfectly fine with my attempt to port it to R.
At one time he had intended to develop a
commercial version, but as of the last time I
spoke with him (a couple of years ago) that
was on permanent hold.

If any of you are also interested in hblm,
I'd be glad to try to help resolve the technical
(and any legal) issues that arise.  The program
has wonderful graphics for meta-analysis,
though others are catching up now.

David
___________________________________

David Rindskopf
Educational Psychology
CUNY Graduate Center
365 Fifth Avenue
New York, NY 10016

(212) 817-8287
(212) 208-2690  (FAX)

____________________________________


Re: [R] Can anyone read a S-PLUS .dmp file for me?
Prof Brian Ripley
Wed, 06 Dec 2006 05:38:04 -0800
On Wed, 6 Dec 2006, John McHenry wrote:

> Anyone?

Yes, I can read it, but what use is that to you?  Given that it is someone
else's copyright work, I am not at liberty to redistribute a different
version:

   hblm and associated programs Copyright 1995 by William DuMouchel.
   Permission is given for not-for-profit redistribution of the hblm
   programs so long as this About.hblm variable is included unmodified.
   Please report problems, failures and successes of this program to
   [EMAIL PROTECTED]

The reason for the error is that it appears to contain invalid S data
frames.  (They are in invalid in R and in current S-PLUS, but they seems
also to have been invalid in S3.  Restoring dumps used to be one way to
create invalid objects, but many of the loopholes have been plugged in R.)

>
> John McHenry <[EMAIL PROTECTED]> wrote:    Hi WizaRds,
>
>  I tried reading the S-PLUS file
>
>  ftp://ftp.research.att.com/dist/bayes-meta/hblm.dmp
>
>  into R using
>
>  data.restore("hblm.dmp")
>
>  but I got an error:
>
>  Error in attributes(value) <- thelist[-match(c(".Data", ".Dim",
".Dimnames",
>  :
> row names must be 'character' or 'integer', not 'double'
> In addition: Warning message:
> NAs introduced by coercion
>
>  Does anyone know how to read this type of S-PLUS file into R? I am not
> familiar with it.
>  On http://cran.r-project.org/doc/manuals/R-data.html it is suggested that
>  "it is usually more reliable to dump the object(s) in S-PLUS and source
the
> dumpfile in R"
>  See also, http://tolstoy.newcastle.edu.au/R/help/05/12/18209.html
>
>  I don't know how this file was created. Could someone with S-PLUS access
> please see if they can read it?
>
>  Thanks!
>
>  Jack.