Skip to content

[Bioc-devel] exprSet and friends

5 messages · Robert Gentleman, Vincent Carey, Wolfgang Huber +2 more

Original
Robert Gentleman
# 
Hi,
   I would like to get some comments on whether we can safely make all 
exprSet methods defunct, or of folks think we want yet one more cycle? 
I have found a few more lurking around recently, and I think it would 
help us to get them all expunged, sooner rather than later.

   thanks
    Robert
Vincent Carey
# 
i have tried to get rid of mine.  i see no role for them so
i am in favor of defunct.

---
Vince Carey, PhD
Assoc. Prof Med (Biostatistics)
Harvard Medical School
Channing Laboratory - ph 6175252265 fa 6177311541
181 Longwood Ave Boston MA 02115 USA
stvjc at channing.harvard.edu
On Thu, 13 Sep 2007, Robert Gentleman wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Wolfgang Huber
    

    

      
      #
    

  


  

      
Hi Robert,

my vote would be to leave them around for this release and .Defunct them 
straight afterwards in the devel branch. We have 10 days until the 
release schedule says "Deadline for packages passing R CMD check without 
error". Many of the packages currently look quite good on the daily 
check page. Some authors might now be away or busy during the next ten 
days and it seems harsh to (possibly) impose making consequential 
changes to their packages within that deadline.

But I'll be happy to be convinced otherwise - progress is good and needs 
to encouraged

  Best wishes
	Wolfgang

            

      
      
      
    

  
  


  


      

    

    
3 days later

    

      
      

        


  

        

      Keith Satterley
    

    

      
      #
    

  


  

      
I agree with Wolfgang. Don't make changes so close to a release.

I'm not sure what effect expunging all exprSet methods would have. I currently 
need the exprSet class definition and accessor methods to get expr and se.expr 
data from exprSet objects. Users may have saved there data previously using 
affylmGUI. A user may wish to open that data file again using a current version 
of affylmGUI. affylmGUI supports this by opening the old data file, getting the 
expr and se.expr data from the exprSet object. affylmGUI will then save the data 
in an updated version of the affylmGUI data object, if the user requests it. I 
don't wish to alter this process before the R 2.6.0 release date, as I would 
need to make several changes and need to do adequate testing of these changes,

cheers,

Keith

========================
Keith Satterley
Bioinformatics Division
The Walter and Eliza Hall Institute of Medical Research
Parkville, Melbourne,
Victoria, Australia
=======================

        
Wolfgang Huber wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Henrik Bengtsson
    

    

      
      #
    

  


  

      
Hi,

what about creating a package 'BioCDeprecated' and move those classes
and function there?  Packages making use of deprecated
classes/functions can then depend on that package.  This will make it
clear for everyone that there are deprecated things being used.  It
would also clean out the code and the documentation of a few BioC
packages.  BioCDeprecated can be used for other purposes in future
release cycles as well.

A safe step towards the above is to extract the above package and make
Biobase dependent on it.  That should be totally transparent for all
packages as is today.  When that is working (by looking at the nightly
builds), one could scan packages for 'exprSet' and add BioCDeprecated
to Depends: in DESCRIPTION.  As a final step one removes the
dependency Biobase -> BioCDeprecated and then nightly builds would
tell us what remaining packages has to have their DESCRIPTION updated.

Cheers

Henrik

        
On 9/13/07, Robert Gentleman <rgentlem at fhcrc.org> wrote: