Skip to content

suggestion for make.names

3 messages · Brian Ripley, Peter Dalgaard, minka@stat.cmu.edu

#
Moved to R-devel, as it seems more appropriate to that audience and I 
think has a slightly more liberal policy on attachments.

You say
where, please?
On Tue, 17 Jun 2003, Tom Minka wrote:

            
Yes, we know and it is planned to allow that.  However, R is a volunteer 
project and that looks far from trivial and rather low priority.  Now if 
you have a patch for that we would be pleased to see it.
#
Prof Brian Ripley <ripley@stats.ox.ac.uk> writes:
It is not only non-trivial, but there is also a risk of running into
logical inconsistencies in some applications of model formulas. E.g.
in nls() and friends, the RHS is an arbitrary R expression and it
might involve true character strings -- think

y~lsoda(start, "Time since inj.", "myfun", parms=c(a=1,b=2),
        dllname="myfun.dll") 

where "Time since inj." could be a variable name, but the other two
are literals.

I'm rather tempted to suggest an extension to R's syntax involving a
"symbol quote" (the backtick, `like this`, would be available).
Haven't really thought it through though.
#
Peter Dalgaard BSA writes:
Maybe I() could be used to escape the true character strings?

Tom