Re: [boost] boost::nil

From: Beman Dawes (bdawes_at_[hidden])
Date: 2001-06-08 12:08:31

At 09:05 PM 6/7/2001, Douglas Gregor wrote:

 >Hello all,
 >	The function library currently under review contains a type 'nil_t' with 
an
 >
 >instance 'nil'. It is a replacement for "0" or "NULL" in any code.

Someone who I guess didn't want to post to the list sent me this:

 >The problem is that there is no portable definition of nil. nil has been
 >defined by a bzillion different people to mean different things.
 >Stroustrup's D & E discusses the problems with this. CORBA has a nil,
 >Czarnecki & Eisenecker have a nil in the software from their book
 >'Generative Programming' and many other class libraries have a nil. This
 >seems like a bad idea IMHO.

I'm inclined agree with him/her.

One further problem is that when it is discovered that a library contains a 
nil or null or other 0 replacement, all further discussion is about that 
replacement.  The library itself recedes into the background.

It seems to me that 0 replacements polarize people with hardly any 
benefit.  You might be able to defuse some of the discussion by changing 
the name to something less likely to act as a lightening rod.  But that is 
about the best you can hope for.

--Beman