Re: Some Laws

From: mAsterdam <mAsterdam_at_vrijdag.org>
Date: Fri, 24 Sep 2004 14:48:34 +0200
Message-ID: <415417ad$0$36861$e4fe514c_at_news.xs4all.nl>


Laconic2 wrote:

> robert wrote:
>

>>... the relational database is really an object
>> store without the froo-froo.
>>it is data and behaviour rolled into one.  

Data in a database is data with contraints on behaviou?r. The constraints just put outer limits on the behaviour.

>>...with low cost DBs, there's no
>>reason to lock yourself into some application code.  the truth is in the
>>data, and any application (or none at all) can access and update it
>>without worrying about hurting it:  it knows the rules to protect itself
>>from knucklehead users, and knuckleheader coders.  that's a great thing
>>to have.

>
> Excellent point!
>
> The truth is indeed in the data, and not
> only in the process that formed it.

Emphasize: *only*, some is.

> I was a programmer for 20 years before going into databases. And I had what
> one analyst called the "process centered mindset." He was very patient with
> me in explaining the difference between the "process centered" view of the
> world, and the "data centered" view of the world. Learning databases was,
> for me, learninng a whole new way to think.


IPO               &                    IDO

IDO: Input process - Data(base) - Output process. IPO: Input data - Process(base) - Output data.

IPODIPODIPODI
(hard to draw a circle in text -- something like this:)

                   iPo

            dOp/pId    pOd/dIp

                   oDi

dOp/pId: the databases' Output process, the Process' input data pOd/dIp: the procesess Output data, the databases' input process


> My disappointment with the track record of object oriented thinking is this.
> In theory, "object centered thinking" could reunite "process centered
> thinking" and "data centered thinking" in a very nice way, preserving the
> best of both worlds, and reconciling them with each other.

Same feeling here.

I don't *want* to sound misterious, but the next does, no doubt. Sorry for that.
I have a hunch about this along these lines:

(1) To get the cabbage, goat and wolve across the water, at first you have to leave the goat behind.

> But the people
> who are "turning Java into COBOL" are reducing "object oriented thinking" to
> merely "process oriented thinking", using fancy terminology.

(2) When trying to take the wrong things with you in one crossing you end up without goat. This would map to: trying to reconcile the wrong combination of concepts will get you to lose conceptual integrity, and render a non-working system. Ideological frontiers do that to you.

(Intuitive, very incomplete:)
IDO:
semantic networks, state, contraints, datatypes, integration, monologue (datawarehouses), meaning...

IPO:
hierarchy, datatypes, dynamics, order, encapsulation, dialogue(UIs), use ...

In-between: persistance/storage, directories/dictionaries, transitions, views, ORBs, protocols, ...

Just a hunch.

> BTW, I'd like to say a few kind words about COBOL. It was a really good
> attempt at defining a "higher level language", one that would create code
> worth reading by non coders. Somebody should revive that attempt with an
> OOPL that doesn't look like portable assembler.

Amen to that. Received on Fri Sep 24 2004 - 14:48:34 CEST

Original text of this message