Re: What is wrong with OO ?

[drush@zakalwe.raleigh.ibm.com]

Elliot (Ell@access1.digex.net) wrote:
>I'd like to know what it is the "object gurus" have said/done, or haven't
>said/done that makes you two distrust them to design systems.  It
>certainly wouldn't be for lack of experience as virtually all of the ones
>I know of have extensive experience in medium and large system
>development. 

Actually, my contention (as was the original poster's IIRC) was with the
local, self-proclaimed, object-gurus. Most of the people who have published
books seem to have learned enough to be worth listening to. 

The local toads, OTOH, I have found to be unable to listen[1], unable to value
ideas that did not originate with them[2], unable to code (either cleanly or
efficiently[3])... On the gripping hand, every discipline has its own set of
power games: right now objects are hot, so they attract this kind of behavior.

>Do you all have any objections to object gurus other than that most of
>them see OO in the real world, or human cognition?  Do you all object to
>the way most object gurus emphasize that planning should lead coding? 
>What what concretely is you alls beef?

Regarding Objects & "real-world" modeling: I have also seen this as primarily
a political ploy. Yes there are objects in the real world, but (nearly) as soon
as you start doing analysis you are working in an abstracted realm. The real
question is how to structure your abstractions, and that can frequently be
done in many different ways.

The "real-world" is the system that you're replacing. To inflict that structure
on the system you're designing can be a *BIG* mistake.

Happy Hackin'
	david rush
	mailto:kumo@intercenter.net
	flamesto:/dev/null

	I bill $100/min for reading commercial e-mail.


[1] Once I asked the same question, three times in three different ways of
    the same individual. Each time I got the same 20 minute lecture which
    did not answer my question, but consisted on an exposition of his pet
    project. BTW - I was finishing/maintaining an OOP project that was 2
    *years* over schedule that he'd designed, and then gotten bored with.

[2] My boss at the time developed a great strategy, that I have yet to be
    able to perfect - he got the self-proclaimed (his own words) "Object Head"
    to come up with the same design ideas through very careful coaching.

[3] Same OH, designed & implemented a server process that used 12 threads
    internally (where he architecturally could have used 3), by missing an
    essential design idea from having never built server-side software
    before. He got the OOA all wrong, but it was thouroughly OO...and the
    cross object interactions still made the thing unmaintainable. After two
    years of maintaining this thing that *never worked right* I finally got
    him to realize that he had botched the design. It was too late, he was
    advocating a framework change where we would no longer need it.