Skip to content

Spelling (PR#6570)

7 messages · Peter Dalgaard, Roger D. Peng, Deepayan Sarkar +3 more

Original
Peter Dalgaard
# 
I came across this in connection with an unrelated issue
Error in beta[2] : object is not subsettable
Error in "[<-"(`*tmp*`, 2, value = 5) : object is not subsetable

One of the messages must be wrong, but I need a native English speaker
to tell me which one.
Marc Schwartz
#
On Tue, 2004-02-10 at 18:13, p.dalgaard@biostat.ku.dk wrote:
Peter,

To be clear, I presume that you are referring to either the single or
double 't' in the words 'subsettable' and 'subsetable'.

Note also that in R-lang.pdf, the following also exists:

10.4.3 Index constructions
...
The object can formally be any valid expression, but it is understood to
denote or evaluate to a subsetable object. 
...



A Google search suggests that both spellings are in use, though a quick
review would suggest that the double 't' is more common. 

For example:

at http://www4.ncsu.edu/~sgarg/l2h/node7.html, the following:

"Modularization and Subsetable systems"


Then of course, Omegahat has this:

http://www.omegahat.org/api/org/omegahat/Environment/DataStructures/Subsettable.html


Curiously, I have not been able to find either spelling in any
"mainstream" english language dictionary (ie. Oxford, Cambridge,
Websters). 

I could not find 'subsetting' in any dictionary either, though that word
with a double 't' also appears frequently in many technical references,
including ?"[":

"You can write methods to handle subsetting of specific classes of
objects..."


If you want my vote (FWIW), I would go with the double 't'. Keep in mind
that while I now live in Minnesota, I was born in Brooklyn, New York.
That may bias your opinion as to whether or not I am a native English
speaker...  ;-)

I will of course defer to any English majors who may be reading this.  

HTH,

Marc Schwartz
Roger D. Peng
# 
Unfortunately, with English it's conceivable that they're both 
correct.  But the double 't' sounds/looks *more* correct to me.
Of course, this is coming from a native English speaker who is 
generally unfamiliar with the rules of English.

On the other hand:

test --> testable?
arrest --> arrestable?
contest --> contestable?
detest --> detestable?
...

So maybe the single 't' is correct.  Did that clear things up?

-roger
p.dalgaard@biostat.ku.dk wrote:

            

      
      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Deepayan Sarkar
    

    

      
      #
    

  


  

      
On Tuesday 10 February 2004 20:25, Roger D. Peng wrote:

            

      
      
      
    

        
This is probably not a valid comparison, since test->testing whereas 
set->setting. 

Oxford (OED) lists the word 'settable' but not 'setable'. subsettable (or 
subsetable) is obviously an artificial word (we use subset as a verb), but 
presumably invented derivatives of 'subset' should behave in a manner similar 
to those of 'set'.

Deepayan

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Marc Schwartz
    

    

      
      #
    

  


  

      
On Tue, 2004-02-10 at 21:14, Deepayan Sarkar wrote:

            

      
      
      
    

        
OK...wait a minute.  I just found the following at:

http://grammar.uoregon.edu/spelling1.html

[Go Ducks!]

To Wit:

DOUBLING A FINAL CONSONANT WHEN ADDING A SUFFIX: from "Correct Spelling
Made Easy" (p. 69)


1) The word must end in just one consonant.
        
Compel (l + ed or ing) = Compelled; Compelling

But Not : Resist (+ ed or ing) = Resisted; Resisting
        
        
2) There must be only one vowel before the final single consonant.
        
Refer (r + ed or ing) = Referred; Referring

But Not: Appear = Appeared; Appearing
        

3) The last syllable of the verb must receive the accent.
        
Commit (accent on mit) = Committed; Committing
        
But Not: Profit = Profited; Profiting
        
        
4) The suffix to be added must start with a vowel--in order to double
the final consonant.
        
Defer (r + ed) = Deferred

But Not: Defer + ment) Deferment


So, if these rules are correct, it should be "subsetable", as 'subset'
meets 1, 2 and 4, but fails 3, as the accented syllable according to
various dictionaries is 'sub' and not 'set'.

Is it too late to change my vote?

Thus, 'testing' fails due to rule 1 ('st'), whereas it is 'setting' due
to 'et'.

HTH,

Marc

<I now return you to your regularly scheduled programming
of...well...programming...>

  
  


  


      

    

    

      
      

        


  

        

      Paul Gilbert
    

    

      
      #
    

  


  

      
Roger D. Peng wrote:

        

            

      
      
      
    

        
I expect this case is more like "upset," and I think there are four 
important rules.

1/ Is it being used as a noun, adjective, or verb?

2/ It may be done differently in England than it is in the colonies, but 
you can't count on it.

3/ No matter which way you do it you are certain to upset someone.

4/ If you do it both ways you may appease everyone, but more likely you 
will upset everyone.

HTH,
(I vote for two t's.)
Paul

  
  


  


      

    

    

      
      

        


  

        

      Thomas Lumley
    

    

      
      #
    

  


  

      
On Wed, 11 Feb 2004 p.dalgaard@biostat.ku.dk wrote:

        

            

      
      
      
    

        
English (OED or Chambers) will be subsettable, American might well have a
single t. I notice this most in words with double l: eg American uses
`modeling' rather than `modelling'.

R usually tries to use English (but is a little schizophrenic about OED vs
Chambers, eg for words ending in -i{s,z}e

	-thomas