Hacker News new | past | comments | ask | show | jobs | submit login
RStudio,a new IDE for R (rstudio.org)
271 points by agconway on Feb 28, 2011 | hide | past | favorite | 33 comments



Really impressed!

I was looking for the Buy button when I saw the link to github.


Very slick UI. How is this done? Is the main UI HTML/Javscript or QT? I guess it's good that I can't immediately tell.


Thanks! It's HTML/Javascript. (Note that you can use it as a webapp, we couldn't do that with Qt obviously)


But you make native dialog calls too?

Care to expand on how Qt, GWT, pure Javascript mix? I think you guys have really pulled off something great in terms of cross platform support and the web.


> But you make native dialog calls too?

Yes, sometimes we call back into native code, especially to do things like file dialogs that you really prefer to be native.

> Care to expand on how Qt, GWT, pure Javascript mix? I think you guys have really pulled off something great in terms of cross platform support and the web.

Thanks! We use QtWebkit as the container for our UI. QtWebkit has hooks to let Qt code execute JavaScript in the page, and you can also inject Qt objects into the window object so JS can call into it. GWT's extensible compilation pipeline actually lets us create stub interfaces in GWT that mirror the Qt classes we want to call, and automatically generate the glue code at compile time.

At some point I will blog about this in more detail at http://jcheng.wordpress.com. In the meantime feel free to peruse our source code at http://github.com/rstudio/rstudio and reach out if you have specific questions.


hm tried it on mac snow leopard and all i get is an empty white frame. :((

i have kde3 installed on the mac. might that cause some problems?


More likely a proxy server issue; make sure any HTTP proxy you use is bypassed for localhost and 127.0.0.1. If that doesn't work, come visit us at http://support.rstudio.org!


The comments on here and the screenshots made me install it, although I don't have any interest in R.

Slick indeed and fast as well, kind of screams for a dedicated project/repo for the UI components, as a starting point for other IDEs.


Wow... who knew that JJ Allaire, maker of (among other things) the ColdFusion web development system, also liked playing with R? http://www.rstudio.org/docs/about


AGPL license, interesting. I suppose the plan is to monetize with a hosting/SaaS service?

Edit: Found the answer on their blog:

> RStudio is also a company, and we plan to sell services (support, training, consulting, hosting) related to the open-source software we distribute.


I'd be curious to hear their calculus for why they thought this services route would be more lucrative than selling the product and services.

It seems like the type of product that people who would balk at purchasing it would be the same type to balkat most types of services listed (except maybe hosting).


For us, it was never about which approach was more lucrative. We want RStudio to be the de facto IDE for R. How could it be anything but open source?


Yes, but there are multiple open source licenses, including, but not limited to, GLP (which is used by R), AGPL, LGPL and BSD.


Also notice there are other non-open-source alternatives such as the IDE of Revolution.

I am very happy to see the way this is turning out...


The UI pretty similar to RKWard (http://rkward.sourceforge.net/), which is according to me the best IDE, especially for Linux users.

P.S. Someone should make a comparison between these two IDEs.


In my experience, RKWard frequently crashes when loaded with modest datasets (1700 data points or so for me). I wonder if RStudio has the same memory problems. If not, I'd be willing to give it a shot.


I don't have any problems, but I use it mostly as a code editor, I don't use the built in table/data editor.


Completely off-topic, but the first thing I thought of when I saw this was R-Studio, an old file/disk recovery software. I was surprised to see that the R IDE has a trademark on "RStudio" (in both US and Canada) given that the former has been around for at least a decade. I guess R-Tools just never cared to file for a TM?


Ianal but afaik (TM) is by definition non-registered (or filed). (R) is. (That's why I (tm) my commercial stuff but I don't (R) any.)


Looks good, but it's crashing on my Mac OS X (Snow Leopard). Any idea how I can debug/fix this issue?


We're on this over at support.rstudio.org: http://support.rstudio.org/help/discussions/problems/19-imme...


RStudio's binary is named 'rdesktop'. Oops! Although it's not in the default PATH, so it could be worse :)


It's strange and awesome how the linux world always works out these collisions even though there's no governing body or group.


Looks like there's a bug report already http://bit.ly/epH9kt


Didn't know anything about R until seeing this post, downloaded the Windows binary and within the hour I was producing graphs. Much easier to get into than MatLab (used it at uni, which is probably why I have an aversion to it!).

Really nice bit of kit (as is the R language), and it's free!


I use MatLab mostly for statistical purposes, together with EViews (statistical software geared towards econometrics). This could certainly ease the transition towards R and replace both aforementioned tools.


This is absolutely gorgeous. The only thing the standard RGui has on it is basic emacs/bash-ish keybindings (C-e, C-a, etc.) for navigation.


I was browsing through the web server implementation:

https://github.com/rstudio/rstudio/tree/master/src/cpp/core/...

Can anyone familiar with writing web servers in C comment on this? For instance, how full featured is it? How useful is the code for learning about web servers?


You're better off looking at the Boost ASIO examples here, which the RStudio code was originally based on. http://www.boost.org/doc/libs/1_45_0/doc/html/boost_asio/exa...


Cool stuff. But it seems lacking manual to run it. I would like to see manuals about how to build it, run it at command line and etc. Better docs can really add a lot attractions. Anyway, thank you for providing such as beautiful tool. It has been years that R under linux has no such a good IDE.


This looks great. Can't wait to dig into it some more.

Also excited about the timing... next week I'm starting a new position at a research co where most of my colleagues work in SAS (vs. my preference for R).

This may be just the thing to help win some converts :)


I find it strange, that I have to download and install a 40 Meg statistics package to run an IDE.


It's not that strange when you realize it's an IDE for a statistics processing language.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: