[ajug-members] design theory

Les Hazlewood les at hazlewood.com
Mon May 5 10:07:50 EDT 2008 

Just a side note - I hope no one ever really creates an Interface
named BombInterface or IBomb or anything crappy like that.

Bomb _should_ be refactored into an Interface itself.  Then, if you
only have one implementation, call the class DefaultBomb or something
like that.

I know there are others that think this isn't a big deal (the
Microsoft world loves IBomb and similar hungarian notation
constructs), but I absolutely _despise_ it.   The reason is that
because it inherently 'undercuts' the true principals of Object
Orientation.  An interface exists as a behavioral specification: one
component literally "interfaces" with another component via that
specification.  In true OO form, the component making the call isn't
_supposed_ to know if its an interface or an abstract class or a
concrete implementation.  The more succinct, noun-matching names are
inherently more self documenting too.

That is the beauty of interface-driven design and polymorphism.  The
very use of Hungarian notation in interface and class names themselves
shows either a lack of understanding of these principals or a laziness
in choosing to ignore these principals in code.  Don't get started on
bad habits!

Let the flame war begin... :)

On Sun, May 4, 2008 at 9:30 PM, Bentley, Jeff <jb5561 at att.com> wrote:
> Ok. I'll bite.
>
>  Imagine you have some service you need to implement called
>  NuclearWeaponControl.
>  It has one method called "public void detonate(GPSCoordinates
>  coordinates) throws CantDetonateException".
>  The method is supposed to throw the exception in 500 different
>  situations.
>
>  You are given a Class call Bomb with one method named
>  blowUp(GPSCoordinates coordinates) that will explode a bomb at the given
>  position.
>
>  Cool. So you write a class called NuclearWeaponContol with the detonate
>  method that calls Bomb.blowup(..)  and are feeling like a stud.
>
>  Boss says... Yo dude... Have you tested that method out to make sure
>  that it works to the specs?
>  Mcdaniels says.... Uh.... No... Can I detonate a few bombs to check it
>  out?
>  Boss says... No
>
>  In walks the high priced spring advocate that says... No problem. I can
>  write some code to test that method behavior with out actually firing
>  off a nuke.
>
>  Boss fires Mcdaniel and hires the spring advocate.
>
>  Spring advocate then extracts an interface called BombInterface from the
>  class Bomb.
>  Spring advocate uses spring to inject a TEST implementation of the
>  BombInterface into future test cases.
>  The TEST implementation of BombInterface is called BombTest.
>  Spring advocate then writes a series of junit tests to exercise
>  NuclearWeaponControl that uses the injected BombTest vs Bomb.
>
>  The spring advocate gets hungry and orders a pizza. Mcdaniel delivers
>  the hot pizza and gets a $10 tip from the high priced spring advocate.
>  :-)
>
>  The spring advocate finishes 500+ junit tests that exercise the 500
>  exception situations defined in the requirements and banks the check
>  from the boss.
>
>  The boss is happy, the spring advocate is happy, the interface newbie
>  learns a lessson,
>  jeff
>
>
>
>
>
>  -----Original Message-----
>  From: mcdaniel [mailto:521 at ofig.org]
>  Sent: Sunday, May 04, 2008 5:51 PM
>  To: ajug-members at ajug.org
>  Subject: [ajug-members] design theory
>
>  can someone tell me what design to an interface means- what i mean by
>  this is that i know what it means by definition- because i read about it
>  often- most oop examples are oak tree extends tree, implements growable,
>  and then tree extends plantae... etc -then i see lots of  code right up
>  front as extends jframe, extends awprx.border etc; so...  and this may
>  be basic'ish, but what's a persons' example of design to the
>  interface(s); by the way- i am enjoying learning java (i've programmed
>  for years, but it has all been procedural) and tks for any (theory)input
>  [i think that i just casted theory-traits onto what was an 'input'
>  object]
>
>  Shivraj Patil wrote:
>  > answer is Spring oriented.
>  > Design using Interfaces than Classes,
>  > it gives us multiple inheritance and polymorphism benefit.
>  > Enjoy!
>  >
>  > On Fri, May 2, 2008 at 9:51 PM, mcdaniel <521 at ofig.org
>  > <mailto:521 at ofig.org>> wrote:
>  >
>  >     i understand, this was not a question- like, from a test- it's
>  >     simply me
>  >     wondering...  -i understand why it looks that way though...
>  >
>  >     b
>  >
>  >     Chris Fowler wrote:
>  >     > On Fri, 2008-05-02 at 21:38 -0400, mcdaniel wrote:
>  >     >
>  >     >> can anyone please answer my question?
>  >     >>
>  >     >
>  >     > Not good following a mailing list post asking for an answer to
>  the
>  >     > question.  You'll not get many answers that way.
>  >     >
>  >     > I did a quick google and found some answers
>  >     >
>  >     >
>  >
>  http://saloon.javaranch.com/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic&f=2
>  7&t=002061
>  >
>  <http://saloon.javaranch.com/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic&f=
>  27&t=002061>
>  >     >
>  >     > http://www.google.com/search?hl=en&q=Extend+Threads+vs
>  >     <http://www.google.com/search?hl=en&q=Extend+Threads+vs>
>  >     > +Runnable&btnG=Google+Search
>  >     >
>  >     >
>  >     >
>  >     >> barclay
>  >     >>
>  >     >>  >>>>
>  >     >>
>  >     >> 5/1/2008; barclay wrote:
>  >     >>
>  >     >>> why would one ever "extend Thread" vs. "implementing
>  >     Runnable"? -are
>  >     >>> there speed gains/losses?
>  >     >>>
>  >     >>>
>  >
>  ------------------------------------------------------------------------
>  >     >>>
>  >     >>> _______________________________________________
>  >     >>> ajug-members mailing list
>  >     >>> ajug-members at ajug.org <mailto:ajug-members at ajug.org>
>  >     >>> http://www.ajug.org/mailman/listinfo/ajug-members
>  >     >>>
>  >     >>>
>  >     >> _______________________________________________
>  >     >> ajug-members mailing list
>  >     >> ajug-members at ajug.org <mailto:ajug-members at ajug.org>
>  >     >> http://www.ajug.org/mailman/listinfo/ajug-members
>  >     >>
>  >     >
>  >     >
>  >     > _______________________________________________
>  >     > ajug-members mailing list
>  >     > ajug-members at ajug.org <mailto:ajug-members at ajug.org>
>  >     > http://www.ajug.org/mailman/listinfo/ajug-members
>  >     >
>  >     >
>  >     >
>  >
>  >     _______________________________________________
>  >     ajug-members mailing list
>  >     ajug-members at ajug.org <mailto:ajug-members at ajug.org>
>  >     http://www.ajug.org/mailman/listinfo/ajug-members
>  >
>  >
>  >
>  >
>  > --
>  > Regards,
>  > ShivRaj Patil
>  > 4685, Apartment # K4
>  > Atlanta, GA-30338
>  > Cell: (908) - 720 0285
>  > ----------------------------------------------------------------------
>  > --
>  >
>  > _______________________________________________
>  > ajug-members mailing list
>  > ajug-members at ajug.org
>  > http://www.ajug.org/mailman/listinfo/ajug-members
>  >
>
>  _______________________________________________
>  ajug-members mailing list
>  ajug-members at ajug.org
>  http://www.ajug.org/mailman/listinfo/ajug-members
>
>  _______________________________________________
>  ajug-members mailing list
>  ajug-members at ajug.org
>  http://www.ajug.org/mailman/listinfo/ajug-members
>

More information about the ajug-members mailing list