Re: What is wrong with OO ?

[rmartin@oma.com (Robert C. Martin)]

In article <5aae85$ad8@masters0.InterNex.Net>, clovis@wartech.com wrote:

> In <5aa73v$p14@news3.digex.net>, ell@access5.digex.net (Ell) writes:
> 
> >I find the key aspects of C++ to be within the parameters of the
> >_progressive_ OO paradigm and practice.  In many ways C++ has led and
> >still leads in the expression and implementation of many significantly
> >useful OO concepts.
> 
> I concur with this, but think it does not go far enough.
> 
> ANY language, up to and including direct object code, is just fine.
> 
> The real problem is personal discipline -- designing completely before coding.
> 
> Wirth's paradigm is STILL at the basis of all good design and coding practice.
> 
> Data Structures + Algorithms = Programs.
> 
> This whole thread somehow ignors that basic reality.  The most aggressive 
> paradigm still respects Wirth's basic rule.  One MUST have structured
data (otherwise
> exact access is impossible, and all algorithms produce garbage).  One
MUST have
> algorithms -- otherwise, the data is nothing more than an electronic,
fixed book
> which one can't even access reliably.
> 
> Any language which provides both these features is sufficient.
> 
> All we're really discussing is HOW to provide data structures and algorithms.
> 
> There isn't a real paradigm shift until the basic model itself changes. 
All we're
> really talking about is how to represent the data structures and
algorithms, that is,
> how to go about implementing them.

The paradigm shift in OO comes about because of some very different
separation strategies.  Prior to OO, data was accessed from data structures.
Any algorithm that needed that data had knowledge of the data structures.
In OO, this changes.  Some algorithms are closely associated with 
a data structure and have intimate knowledge of it.  However, other algorithms,
that use the data within the data structure have no knowledge of the data
structure.  They are isolated from the data through abstract interfaces. 
When
the data structure changes, these algorithms do not change.  Moreover, these
algorithms can make use of many different kinds of data structures without
knowing it.

Thus there is a fundemental change.  We still create data structures, and
we still create algorithms.  But in OO we also create interfaces that
separate some algorithms from some data structures.  Then we must decide which
algorithms know about which data structures, and which should be kept separate.
It is that decision that is called OOD.

If you don't believe that this is a fundemental paradigm shift, then consider
that this simple trick of separating algorithms and data structures creates
an equation that is quite different from Wirth's:

program = Interfaces(algorithms,data_structures)

Where the algorithms and data_structures are independent variables that
have no knowledge of each other, and the interfaces are the mapping functions
that bind the two together to create certain programs.

-- 
Robert C. Martin    | Design Consulting   | Training courses offered:
Object Mentor       | rmartin@oma.com     |   Object Oriented Design
14619 N Somerset Cr | Tel: (847) 918-1004 |   C++
Green Oaks IL 60048 | Fax: (847) 918-1023 | http://www.oma.com  

"One of the great commandments of science is:
    'Mistrust arguments from authority.'" -- Carl Sagan