Wellcome to OpenOffice.
It is really nice to have interested students for this project.
If you haven't subscribed to the [email protected] mailing list yet,
feel free to do so. Many important issues are discussed on the mailing list.
In the previous days I have already discussed some points of the Calc-R
integration (see the archives,
http://sc.openoffice.org/servlets/BrowseList?listName=dev&from=2007-03-01&to=2007-03-31&count=48&by=date&first=21&windowSize=20&selectedPage=2,
the "Summer of Code: R and Calc" topic). Prof. Neuwirth also wrote an
extensive article on R-Excel integration, see
http://dx.doi.org/10.1007/s00180-007-0023-6 (I believe the article is
free; I could access it through my home computer and I don't have a
personal subscription to Springer journals; if it isn't, please ask Prof
Neuwirth for a copy). Please see also the statistical wiki page
(http://wiki.services.openoffice.org/wiki/Statistical_Data_Analysis_Tool).
The "Links" section contains some further information on R integration
into various external programs.
Specific answers regarding:
POINT 2: I will try to emphasize in a later e-mail the R-statistical
functions I believe are crucial to access inside Calc. Though, of
course, other functions may be implemented as well and new ones
could/would be included in the list, as time will permit. Once the basic
framework is set, there shouldn't be any serious problem to implement
new ones. I will post an extended discussion under the previous topic
("Summer of Code: R and Calc").
POINT 3: the Calc-R integration should be platform independent (even IF
the code would have some platform dependency, there should be an
abstraction layer and most of it should be platform independent; a
connector should be of course available for every platform).
Unfortunately, I cannot offer much coding help. However, feel free to
ask any questions regarding statistics. I will be happy to help (though
it may take some time until I respond).
Sincerely,
Leonard
Krzysztof Filo Gorgolewski wrote:
Hi,
I've recently came across the proposed ideas for OpenOffice Google Summer of
Code. I've found integrating R with OpenOffice the most interesting - knowing
the power of R it could make a killer feature for OO.
I've got couple ideas and I would be glad if you could review them:
1. R for sure would be appreciated by advanced useres - therefore some binding
functions (for use in formulas) must be created. I believe the most sensible
way to do it is to create elastic framework for defining such functins (each
binding could consist of XML definition and Rscript). More functions could be
easily added with collaborative effort. What is more OOCalc RFunctions could
be named more intuitively than there R counterparts (in R many commonly used
functions are just more general ones used with specific arguments).
2. Statistics are for human beings. Creating wizards that step by step
(helping to choose the best method) helps
user to conduct meaningfull analysis is essential (for comparing means,
ANOVA, corrrelation and chi square at least). Yet again it would be great to
create xml based framework for creating such wizards but further
investigation is needed.
3. The main problem is what platform to use to connect R with OO. They both
have many binding but OO strongly relies on java - with little help of RSjava
we could call R functions from java ( http://www.omegahat.org/RSJava/ ). And
java is multiplatform and thus better then RDCOM. One thing that's need to be
checked is efficiency.
I believe that I'm prepared to implement this - apart from obvious computer
skills (I'm a full-time Computer Science BSc student) by which I mean fluent
programming skills in variety of languages (though C++ and Java will be most
useful for this project) I additionaly study Cognitive Science. That's where
I've learned how to handle data... (I've used SPSS, LISREL, Excel and of
course R). I'm not a statistician, but I believe that I cound introduce more
practical aproach. And I would get satisfaction if OO.o + R would be more
popular than Excel :)
Regards,
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
