[Bioc-devel] Request to add 'normalize' to BiocGenerics

cstrato cstrato at aon.at
Wed Feb 20 20:42:16 CET 2013


Dear all,

Looking at the alphabetical list of functions/methods in BioGenerics I 
have the feeling that all functions are really generic, i.e. general, 
with the exception of function 'annotation' which in my opinion should 
be part of BioBase as method for one or more of the BioBase classes.

The same holds true for function 'normalize' which should also have been 
a method for one or more of the BioBase classes. However, my feeling is 
that today (about 10 years later) it is too late to put 'normalize' in 
one of the BioC base classes. Furthermore, I do not see any connection 
between affy::normalize() and EBImage::normalize(). However, this is my 
personal opinion, and if people would like to have a generic method 
'normalize' implemented, this is fine with me.

If I remember correctly, 'normalize' was first introduced with the affy 
package to preprocess Affymetrix arrays. In the case of my package xps 
'normalize' is only a wrapper to S4 method xpsNormalize(), however, my 
S4 classes are completely independent of the BioBase classes. My 
intention was to use wrapper functions such as 'normalize' etc so that 
people using the affy package do know immediately the meaning of  the 
functions in xps.

I agree with Laurent that "all functions 'normalize' do perform an 
action that can be described with that verb".

Furthermore, since many packages use 'normalize' since many years, it is 
too late to force package authors to change the names of their 
functions. This will break the code of most of the users of these packages.

Best regards,
Christian



More information about the Bioc-devel mailing list