[R-SIG-Mac] An example of Cocoa GUI for R

Byron Ellis bellis at hsph.harvard.edu
Mon May 17 23:34:24 CEST 2004


On May 17, 2004, at 4:29 PM, stefano iacus wrote:

>
> On May 17, 2004, at 7:09 PM, Byron Ellis wrote:
>
>> On May 14, 2004, at 7:16 AM, stefano iacus wrote:
>>
>>> Finally, we are showing you a bit of work on the Cocoa GUI.
>>>
>>> Do not expect too much, it is just a preview. You have some nice 
>>> feature yet like code and file completion, but still don't expect 
>>> too much.
>>>
>>> Have a look here
>>>
>>> http://www.economia.unimi.it/R
>>>
>>>
>>> It works with R.app 1.9.0
>>>
>>> This is a joint work with Simon (mostly his work at the moment)
>>>
>>
>> Looks pretty good so far...I assume the "metal" interface for the 
>> console is because There Can Be Only One? (Though, really, you can 
>> have a console and a bunch of edit windows). But so far, I like it a 
>> lot---if nothing else my ingrained Emacs habits don't get in the way 
>> of using R :-)
>>
>>
>>
>>> When we will have time, we will also write some docs on how to write 
>>> your own gui for R using Cocoa, but don't expect any finished work 
>>> before R 2.0.0.
>>>
>>>
>>
>> You could also---it might be better, from the point of view of 
>> eventually writing documentation, to have those of us less intimate 
>> with your design struggle mightily (or not) for a time with only some 
>> source as our guide (I tried to get something running a while back 
>> just looking at the R source, but not very hard) to get an idea of 
>> the rough spots. One suggestion from just looking at the initial 
>> source would be to have the bundle name be a preference such that 
>> multiple GUIs can be present on a system (perhaps also searching 
>> ~/Application Support or something) and switchable---switching 
>> on-the-fly would be extra-cool but I understand that you have to save 
>> SOME features for later ;-)
>
>
> Indeed Byron, the final goal is you can have several  GUI (possibly 
> switching on the fly, it is not too hard in principle)...and we will 
> need someone to write a second one :)

*hides StatPaper code behind his back* What? Who? Me? :-)

> The problem is our "time" and writing docs require some. I think that 
> in june, probably after 1.9.1, we can have a documented  skeleton on 
> how to implement your own GUI.

Heh, in re-reading I see I edited out that actual verb in my first 
sentence there. The suggestion was to expose the source in truly-ugly 
form without any documentation and watch the pathetic flailing of your 
devoted fans. It would provide for some entertainment at least :-)

> I don't want to document too much what we have done since now because 
> what we have now  won't be like the final  design we have in mind for 
> several reasons. Among the others: the aqua module (like other gui 
> modules) will be absorbed in to R.bin and this is going to happen is 
> r-devel pretty soon.

Thats...interesting didn't they just split those modules out?

>
> stefano
>
>
>
>>
>>> stefano & simon
>>>
>>> _______________________________________________
>>> R-SIG-Mac mailing list
>>> R-SIG-Mac at stat.math.ethz.ch
>>> https://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac
>>>
>> ---
>> Byron Ellis (bellis at hsph.harvard.edu)
>> "Oook" -- The Librarian
>>
>> _______________________________________________
>> R-SIG-Mac mailing list
>> R-SIG-Mac at stat.math.ethz.ch
>> https://www.stat.math.ethz.ch/mailman/listinfo/r-sig-mac
>>
>
>
---
Byron Ellis (bellis at hsph.harvard.edu)
"Oook" -- The Librarian



More information about the R-SIG-Mac mailing list