[R-SIG-Mac] Cocoa Help System Totally Busted(tm)? And Other Random Notes (Not a Bug Report!)

Byron Ellis ellis at stat.harvard.edu
Tue Nov 2 07:48:09 CET 2004

On Nov 1, 2004, at 5:31 PM, Simon Urbanek wrote:

> Byron,
> if you read the docs and our comments, you would know that we 
> emphasized that the "?" hack is very temporary and the only reason it 
> was kept in the release was that a modification of R is necessary for 
> the clean solution and it was too late to do that for the 2.0.0 
> release.
> IMHO the way to go (and I successfully tested) is to re-define the 
> print methods for the help objects. As of htmlhelp - htmlhelp=FALSE 
> basically doesn't make much sense in OS X. It's ok for the console 
> version when the existence of a html browser is not guaranteed, but it 
> make no sense in the Cocoa GUI.

Sure, I arrived at roughly the same conclusion (and suspected that the 
hack was there for a good reason---though I apparently missed it in the 
documentation). I'd take it one step further and suggest that the help 
system in general Made Sense At The Time(tm) but that we have display 
environments rich enough that we really want to do on-the-fly 
calculations for a lot of these things (e.g., to eliminate the simlinks 
for HTML). A lot of Rd parsing stuff has been showing up in R for the 
quality control stuff, there's no reason it couldn't be harnessed to 
parse Rd files on the fly (I know the Gnome interface could to the same 
thing with Gtkhtml, as could Win32, though it already has the CHM 
system that works pretty well)---but doing that would mean a pretty 
major overhaul of the help system and I'm guessing there are higher 
priorities. (Dear Easter Bunny, Multiple evaluators/greenthreads would 
be "Da Bomb" ) :-)

> On Nov 1, 2004, at 6:19 PM, Byron Ellis wrote:
>> Only two-level nesting in the Workspace Browser doesn't work:
> The WB is pretty much pre-alpha thing and will need major re-writing. 
> Unfortunately that is a lot of work and other managers (with less work 
> involved) need to be finished first. I hope that I can weed out that 
> one this week so it'll be more viable for others to put their hands on 
> it.
>> dragging an element from the workspace browser to the console would 
>> paste it.
> With respect to this and the WB, you may have a look at JGR - it's got 
> both issues solved pretty well I guess. The only reason it's not 
> implemented in R.app is the lack of programming power :P So send us 
> the diffs and we'll be happy to include it ;)

I admire the look of JGR, but haven't had much time to play with 
it---but I see what you mean about the browser.

BTW, on a positive note, I think we just set some sort of record for 
"shortest time to prepare a grant" and the new Cocoa GUI had a lot to 
do with that---begin able to get the plot "just so" and hit Save As... 
was a real time saver.

> Cheers,
> Simon
> _______________________________________________
> R-SIG-Mac mailing list
> R-SIG-Mac at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
Byron Ellis (ellis at stat.harvard.edu)
"Oook" -- The Librarian

More information about the R-SIG-Mac mailing list