Re: Simple inheritence question

Robert C. Martin wrote:
> On 25 Jan 2006 18:03:16 -0800, stevenwurster@xxxxxxxxx wrote:
>
> >Robert C. Martin wrote:
> >> >
> >> >I think these classes are legitimate in an Object-Oriented breakdown of
> >> >a Scrabble system.
> >>
> >> What makes you think that?  I agree that they are the obvious nouns in
> >> the problem statement.  But are they the right partitioning of
> >> behavior?  Just what *are* the behaviors that the program must
> >> express?  The computer must shepherd the players to make the correct
> >> moves.  The computer must supply a random source of tiles.  The
> >> computer must ensure that players are playing tiles on valid squares.
> >> The computer must ensure that words are real.  The computer must score
> >> the words, and keep score for the players.
> >
> >I did not choose them because they are nouns.  To do so is just as
> >wrong as assuming that software models cannot be pretty.  The two are
> >not mutually exclusive.  (Although I will agree that business
> >requirements tend to mandate the development of models that "get the
> >job done" but are amazingly ugly.  Then we see these ugly models get
> >turned into things that are even more ugly over time.  What fun.)
>
> IMHO software models *must* be pretty.  I consider it a matter of
> professional ethics to create software designs that are simple,
> elegant, expressive, and malleable.  Business problems can be complex
> and irregular; but the software that solves those problems should be
> well crafted.  We should not suffer ugly software to survive.
>
> On the other hand, (and I am not accusing you of this) it is all too
> common to find software designers who create pretty models that are
> not coupled to the business problems.
>
> See www.waterfall2006.com for more information.
>

If the software model solves the business model then the business model
must be consistent....i.e. there must be some sort of embedding of the
business model in the software one.......you cannot loose the ugliness
of the business model.

.

Relevant Pages

  • Re: Simple inheritence question
    ... >Robert C. Martin wrote: ... >I did not choose them because they are nouns. ... not coupled to the business problems. ...
    (comp.object)
  • Re: AMD planning 45nm 12-Core Istanbul Processor ?
    ... Typical answer of leftist elitists. ... That's exactly what Robert proposes, ... Not only unknown dynamics, but unknowable dynamics; ... and I don't have to support your business model. ...
    (comp.sys.ibm.pc.hardware.chips)
  • Re: AMD planning 45nm 12-Core Istanbul Processor ?
    ... That's exactly what Robert proposes, ... Not only unknown dynamics, but unknowable dynamics; ... and I don't have to support your business model. ... risk off onto the end user. ...
    (comp.sys.ibm.pc.hardware.chips)