There's a bunch of different ways that R interfaces/bindings can go, apparently. From what I can see, most straightforward way to interface with R would be to create a WorkSpace that can send WorkSpace input to an R interpreter and echo the output back to the workspace. You want something more than that, I assume. There's at least two Ruby interfaces to R that I could find. There's at least one Python interface. My cursory reading from a few months ago didn't make clear the difference between the library bindings, the interfaces and the GUIs, so there's probably a lot of overlap. Lawson On 3/9/12 7:29 PM, Hernán Morales Durand wrote:
Hi Lawson, I know from an experienced Java programmer, trying to make a binding to R, which is difficult to deal with particularities of connecting with R efficently (I could ask him for details if a student is interested in the project). I think R isn't comparable to what iGraph could provide, every day more studies results are presented in R and they are not necessarily described in network format. A front-end to R could be another interesting project idea though. Cheers,
Hernán
2012/3/9 Lawson English <lenglish5@cox.net <mailto:lenglish5@cox.net>>
A binding to R would be relatively trivial, I think. A binding to iGraph would be a bit more challenging and would allow for things like using Connectors to directly create/define iGraph data structures and use iGraph's API to modify their behavior.
Lawson
On 3/9/12 6:07 PM, Hernán Morales Durand wrote:
Name: A binding to R Level: Advanced Possible mentor: ? Possible second mentor: ?
Description
With the increased popularity of high-level statistical programming language and environments for data analysis like R, a way to interface this package is becoming a must have for Smalltalk, since it implements complex and unrivalled statistical techniques for data manipulation and presentation, like Analysis of Variance, Covariance, Time Series, Generalized Linear Models, Additive Models, Non-linear Regressions, Tree Models, Multivariate Statistics, etc. besides the many mathematical functions, which are used in fields from economics to medicine and engineering. It is estimated that R posseses about 2 million of users worldwide and more than 2000 add-ons and increasing everyday through repositories like CRAN and Crantastic.org
Technical Details
The student should know or be motivated to learn statistics with the R environment and language, and its fundamental workflow: importing and preparing the data, and finally running the analysis, and presenting the results. Dealing with R sessions and presentation of results (like vectors and plots) will be challenging too.
Benefits to the Student
The student will gain invaluable experience from two complementary environments, and his experience with the interface technology choosed will be useful for the many projects where Smalltalk needs help from external systems.
Benefits to the Community
The goal of this project is to build a wrapper to interface R, an open source statistical programming language, providing a whole range of missing functionality to Smalltalk. This binding could complement the R environment where a general programming environment is needed, attracting many statisticians, and will open Smalltalk to domain-specific areas as diverse as Clinical Trials, Finance and Machine Learning.
2012/3/9 Janko Mivšek <janko.mivsek@eranova.si <mailto:janko.mivsek@eranova.si>>
Dear Smalltalkers,
We just submitted the application to this year GSoC, now let we wait with fingers crossed for Google until next Friday to decide. I think we have quite some chance, specially because we collected and very well described 22 project ideas:
http://gsoc2012.esug.org/ideas
There is still a time for ideas, so if someone has still half finished one, please finish and publish it. Also authors of ideas please check if your idea is put correctly on above page.
Best regards Janko and Carla
-- Carla F. Griggio, Janko Mivšek Smalltalk GSoC Admin Team http://gsoc2012.esug.org
-- Hernán Morales Information Technology Manager, Institute of Veterinary Genetics. National Scientific and Technical Research Council (CONICET). La Plata (1900), Buenos Aires, Argentina. Telephone: +54 (0221) 421-1799. Internal: 422 Fax: 425-7980 or 421-1799.