Re: [Bugs-r-devel] SF.net SVN: bugs-r:[119] trunk/R2WinBUGS

[Uwe L. <li...@st...>]

Dear Gregor,

I have reviewed your code, and I really think people who want to 
generate inits, data and model files in their working directory should 
do it another way, either by a separate argument (which I still dislike) 
or by calling a corresponding function explicitly.

R2WinBUGS:::bug() and BRugs:::BRugsFit() should really not do that 
(given the danger of overwriting or deleting already existing files there).

Better to save those files explicitly and later let the software do the 
tricks in the tempdir().

As said before, Brian and I have had some face to face discussion about 
possible issues.

I am really sorry about that, but I think in this case it is better to 
revert that change.

Best wishes,
Uwe





Gorjanc Gregor wrote:
>> 1. Brian and I tried hard (in January-February 2009, see the svn logs)
>> to make BRugs and R2WinBUGS not write into the working.directory unless
>> a file with full path is given explicitly. A software should not write
>> files into the current directory without being explicitly specified.
>> The working.directory shopuld be used to read files from, not to write
>> intermediate files to. If you want files to be created there: Can we
>> have an additional argument that tells it?
> 
> Ups. I was not aware of this efforts. I agree that messing up with "temporary
> files" is not nice. However, one does not specify working.directory in such
> cases. 
> 
> Let me clarify my intention. If we define working.directory argument in bugs(),
> then this function saves data and inits file in that place, while the model file is
> saved in some temporary place. This surely is not consistent. And if you need
> to do some testing it is hard to find a model file in the temp. folders. I often prepair
> data, inits and model files using bugs() function, because it is handy. This way one can
> create a complete set of files for someone that would like to try to run BUGS, but does
> not use R, say AT ;) Therefore, I thought that consistency could not do any harm.
> Anyway, the model file is being removed by deafult leaving the working.directory clean
> at the end.
> 
>> 2. Does path.expand() work on S-PLUS?
> 
> Good point! Do we have contact with S-PLUS developer (Brian?) to check this out?
> 
> gg
> 
> 
> ------------------------------------------------------------------------------
> Come build with us! The BlackBerry(R) Developer Conference in SF, CA
> is the only developer event you need to attend this year. Jumpstart your
> developing skills, take BlackBerry mobile applications to market and stay 
> ahead of the curve. Join us from November 9 - 12, 2009. Register now!
> http://p.sf.net/sfu/devconference
> _______________________________________________
> Bugs-r-devel mailing list
> Bug...@li...
> https://lists.sourceforge.net/lists/listinfo/bugs-r-devel