JdBP>> It isn't a reason for a C++ Standard because standardising a
 JdBP>> language does nothing for portability in this respect.  The
 JdBP>> presence or absence of a C++ Standard is irrelevant to whether or
 JdBP>> not there exists an "up-to-date" implementation for any given
 JdBP>> machine architecture.
 BP> But why it take so long to create one? 
One what ?  A C++ standard ?  It didn't take all that long, as standards go.  
What makes you think that it took a long time ?
 BP> Why it did not appear years ago?  In thah case "portability" would 
 BP> be out of question.  
Standardisation is *de*scriptive, not *pre*scriptive.  In almost all cases, 
standards are adopted either alongside or after real-world implementations 
exist.
Standardisation does *not* imply an easy or effective mechanism for universal 
portability.  The *only* way that that will happen is if either all platforms 
are homogenised, or consumers require that vendors support the same language 
across all platforms and exercise market forces (i.e. voting with their 
wallets) in order to ensure that the vendors do so.  The former case usually 
indicates monopolisation of all sectors of the market by one company, and 
usually indicates stagnation.  The latter case is what I am implying by 
saying that one should choose Borland C++ or Watcom C++ over Microsoft C++ if 
one wants to develop with the same C++ language features in both 16-bit and 
32-bit: if you don't buy Borland's or Watcom's compilers, which *do* do the 
job requried, but instead buy Microsoft's compiler, which does *not*, then 
you are *saying* that you want one thing, but by your actions exercising 
market pressure so that the exact reverse happens.
 BP> Now every developer make his own quirks. 
For two reasons: history and added value.  
 BP> It would not be needed, just implement what is standardized. 
It is a comment, all too often heard in many fora from those who haven't 
looked at what the standard actually standardises and considered why it does, 
that if one merely implemented what was in the standard one would have all 
that one needed.  Clearly this is not so, otherwise there would be no need 
for the standard to specify that certain areas are to be defined by the 
implementation.  There would also be no complaints (as there *are*) from 
people who complain that important areas of C++ programming are not addressed 
by the standard.
 ¯ JdeBP ®
--- FleetStreet 1.19 NR
---------------