Skip to content

Displaying Equations in Documentation

3 messages · Jarrett Byrnes, Rolf Turner, Brian Ripley

Original
Jarrett Byrnes
# 
I'm currently working on writing up some documentation for some of my  
code, but am having the darndest time coding in equations.  For  
example, the equation in the following:

\details{ Calculated the R Squared for observed endogenous variables  
in a structural equation model, as well as several other useful  
summary statistics about the error in thoe variables.

R Squared values are calculated as

\deqn{R^{2} = 1-\frac{estimated variance}{observed variance}}

Standardized error coefficients are then calculated as sqrt(1 - R^2).
}

While it shows normally using R CMD Rd2dvi, when I actually compile  
and load the package, displays as follows:

R^{2} = 1-frac{estimated variance}{observed variance}


I have also tried

\deqn{R^{2} = 1-\frac{{estimated variance}{observed variance}}}

and

\deqn{R^{2} = \frac{1-{estimated variance}{observed variance}}}

with the same result - Rd2dvi is happy, but the display is still wonky  
in practice.  I've also tried subbing in \eqn{R^{2}} in the rest of  
the text in a few places, but, again, it shows as R^{2}.  Is there  
something I'm missing about inserting equations into R documentation?

-Jarrett
Rolf Turner
#
On 26/08/2008, at 4:50 AM, Jarrett Byrnes wrote:

            

      
      
      
    

        
Displays, presumably, in a plain text or html environment.

	As my late Uncle Stanley would have said, ``What the hell do you  
expect?''

	Plain text and html, unlike LaTeX, do not have advanced mathematical  
display
	capabilities.

	The R documentation facility --- RTFM!!! (section 2.6 Mathematics)  
--- provides a way
	to allow for both possibilities.  You can do:

	\deqn{R^2 = 1 - \frac{estimated variance}{observed variance}}{R- 
squared = 1 - (estimated variance)/(observed variance)}

	Then in the pdf manual for your package you'll get the sexy  
mathematical display, but when you call
	for help on line and get a plain text or html version you'll see

		R-squared = 1 - (estimated variance)/(observed variance)

	which is the best that can be done under the circumstances

            

      
      
      
    

        
Yes.

		cheers,

			Rolf Turner

######################################################################
Attention:\ This e-mail message is privileged and confid...{{dropped:9}}

  
  


  


      

    

    

      
      

        


  

        

      Brian Ripley
    

    

      
      #
    

  


  

      
On Tue, 26 Aug 2008, Rolf Turner wrote:

        

            

      
      
      
    

        
That's not the whole story for HTML, as there is MathML.  At the time the 
decisions were made for Rd rendering, MathML was pretty much unsupported. 
That has changed slowly, and nowadays some browsers (e.g. Firefox, Opera) 
do support Presentation MathML -- unfortunately they often are short of 
suitable fonts.  AFAIK, Safari and IE (including Compiled HTML) do not yet 
support it.

So someone interested could rewrite the Rdconv code to make use of MathML, 
but for now the subset of R users who could benefit from it would be 
small.