[Qgis-developer] SEXTANTE: R output?

[Alister Hood]

>Date: Fri, 12 Jul 2013 16:05:21 +0200
>From: Matteo Mattiuzzi <matteo at mattiuzzi.com>
>To: Victor Olaya <volayaf at gmail.com>
>Cc: "qgis-developer at lists.osgeo.org" <qgis-developer at lists.osgeo.org>
>Subject: Re: [Qgis-developer] SEXTANTE: R output?
>Message-ID:
>        <CADcazozd7Pb=ZWZdbqJ-DTpqaHZkNAh4g=CQyY_+8rtdn32xkg at mail.gmail.com>
>Content-Type: text/plain; charset="iso-8859-1"
>
>>Ok, fine then. Only in the canonical libpath. :-)
>
>Yes I guess it it the saver choice.
>
>>>I think it is (quite) easily possible to remove the need of R knowledge
>>>for a lot of situations, one of these it the installation and update of
>>>packages. It is much easier to share a Rscript (rsx) to another system if
>>>not installed packages are added automatically.
>
>>I agree with that, but what I mean is that the person that writes the
>>script should take care of all the installation/checking/etc code. He
>>should put all that sanity code in his plugin, not expect SEXTANTE to
>>do it, because that makes things less flexible, and it seems that we
>>cannot come to a good solution on that. That's what I meant.
>
>A script writer can care about the own environment but not so much on the
>one of others (just the loading can be handled with ease). R it to much
>modular it is not possible (for what I know) to provide a fully operation
>plugin if a package is not available locally.
>So probably every script need to have in its documentation which library
>has to be present, and a reference to a doc on how to install libraries in
>R. Probably this is a good solution.
>Or/And SEXTANTE provides with a script like the one I have send you with
>the capability to checkout for required packages. So if that script fails
>(the info which library has to be loaded comes from the sript.rsx), the
>info 'can not install package XXX' see doc... would help further to
>manually install.
>But as QGIS is now close to release (for what I know) it could be a
>development that happens after that... (but I don't want to make pressure
>on that!).
>
>I'm always thinking of SEXTANTE as a possibility more for non R-users.
>
>>Of course I like the idea of sharing scripts :-)
>Great count on me! (for all these things!)
>
>Yes I think Error handing (communication) is very important!
>and also an update script could be made that could be started from the
>SEXTANTE options and configuration.


Do any maintenance scripts need to be in the SEXTANTE configuration?
I was suggesting just listing them in the sextante toolbox with any other R scripts.

Regarding users sharing their R scripts, shouldn't these be shared through the same infrastructure as python scripts and styles and whatever else?  Someone was working on that, weren't they?

Regards,
Alister