Skip to content

TRAMO-SEATS confusion?

4 messages · Erin Hodgess, Dirk Eddelbuettel, Allin Cottrell +1 more

Original
Erin Hodgess
# 
Dear R People:

When looking at the previous postings regarding TRAMO-SEATS,
I am somewhat puzzled.

Is it true that we CANNOT replicate TRAMO-SEATS because of 
licensing or ownership issues, please?

If not, would anyone be interested in an R version of it, please?

Thanks,
Sincerely,
Erin Hodgess
Associate Professor
Department of Computer and Mathematical Sciences
University of Houston - Downtown
mailto: hodgess at gator.uhd.edu
Dirk Eddelbuettel
# 
Erin,
On 15 October 2005 at 15:53, Erin Hodgess wrote:
| When looking at the previous postings regarding TRAMO-SEATS,
| I am somewhat puzzled.
| 
| Is it true that we CANNOT replicate TRAMO-SEATS because of 
| licensing or ownership issues, please?

Could you please 

i)  define 'replicate', and 

ii) cite any licenses to back up this claim, or simply show us what you
    find so confusing.

I suspect that you are misinterpreting this, and that tramo-seats is indeed
"simply" closed-source software that you can get re-distributed in binary
form (e.g. from the gretl website, see below), but not in source. 

| If not, would anyone be interested in an R version of it, please?

A couple of years ago I toyed with both the US Census' X12-ARIMA procedure
and its tramo-seats alternative by Maravall et al.  Given how atrociously
'1960s' the X12-ARIMA interface is, I toyed with building an R frontend given
that you get the Fortran code to X12-ARIMA straight from the Census site .
But then something else came up and I never pursued this ...

The closest I know to using x12 and/or tramo-seats from somewhat saner and
more modern software is via Allin Cottrell's gretl (cf http://gretl.sf.net).
And per my suggestion a few years back, Allin even hacked a 'gretl to R'
interface [ via mucking with ~/.Rprofile which isn't pretty but that is
another story ... ]

Hope this helps,  Dirk

| 
| Thanks,
| Sincerely,
| Erin Hodgess
| Associate Professor
| Department of Computer and Mathematical Sciences
| University of Houston - Downtown
| mailto: hodgess at gator.uhd.edu
| 
| ______________________________________________
| R-help at stat.math.ethz.ch mailing list
| https://stat.ethz.ch/mailman/listinfo/r-help
| PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Allin Cottrell
#
On Sat, 15 Oct 2005, Dirk Eddelbuettel wrote:

            

      
      
      
    

        
It's a few years since I concentrated on this, but my recollection 
is that the authors of TRAMO/SEATS were willing to grant access to 
the source to me as a developer, but I was not free to redistribute 
the source.  I was, however, able to produce working binaries for 
Linux and win32.

With both TRAMO/SEATS and X-12-ARIMA, it would be nice to be able to 
produce a "librified" version (i.e. code that can be called as a 
library from R or gretl or whatever), but from my point of view the 
binding constraint is that these programs are written in rather 
old-fashioned Fortran.  I tried, briefly, hacking on the T/S code to 
relax the fixed-memory constraint of no more than 600 observations, 
but found I was just breaking stuff so I stopped.

For gretl, I ended up accepting that T/S and X12A would remain as 
stand-along programs.  Gretl takes user input and sets up the 
command lines for these programs (both of which have rather 
byzantine command-line options), then parses the output files and 
feeds the relevant information back home.

If anyone would like to see how I approached this, look at tramo*.c 
in the "plugin" directory of the gretl code base.

http://cvs1.sourceforge.net/viewcvs.py/gretl/gretl/plugin/

Allin Cottrell

  
  


  


      

    

    

      
      

        


  

        

      Spencer Graves
    

    

      
      #
    

  


  

      
LICENSING AND OWNERSHIP

	  CRAN wants the source code for any contributed package, both for 
security reasons (protection against viruses) and for consistency wiht 
the GNU license, which requires distributors of software using GNU 
software to extend the GNU restrictions to their derivative software. 
I'm not an attorney, but it is my understanding that anyone contribiting 
a package to interface with TRAMO-SEATS would have to submit their 
source code, but not that for TRAMO-SEATS.  The user would be required 
to install separately TRAMO-SEATS.

WOULD ANYONE BE INTERESTED IN AN R DISTRIBUTION?

       I don't know this software, but time series analysis is an 
important topic, and I'd be shocked if no one else would be interested. 
  An interface would make it easier for (a) users of TRAMO-SEATS to 
migrate into R and (b) R users to explore the capabilities of TRAMO-SEATS.

       spencer graves
p.s.  Sundar Dorai-Raj and I are planning to develop a package to 
accompany Ruey Tsay (2005) Analysis of Financial Time Series, 2nd ed. 
(Wiley).  As part of this effort, we plan to invite people to send us R 
code for how they would reproduce various analyses in that book. 
Submissions would become part of the "FinTS" package, and might further 
contribute to comparisons of alternative R packages for time series.

        
Allin Cottrell wrote: