Re: D7 Compiler Bug
From: J French (erewhon_at_nowhere.com)
Date: 10/05/04
- Next message: J French: "Re: TImage and GrayScale"
- Previous message: Rob Kennedy: "Re: TBytearray to variant"
- In reply to: Bjørge Sæther: "Re: D7 Compiler Bug"
- Next in thread: VBDis: "Re: D7 Compiler Bug"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Tue, 5 Oct 2004 07:08:23 +0000 (UTC)
On Tue, 5 Oct 2004 02:18:21 +0200, "Bjørge Sæther"
<bjorge@hahaha_itte.no> wrote:
>J French wrote:
>> On Sun, 03 Oct 2004 17:21:46 -0400, L D Blake <not@any.adr> wrote:
>>
>> <snip>
>>
>>> OO just doesn't fit my head very well... I just don't think that way.
>>> You do.. and that's just fine with me. Really, no grumbles on this
>>> end at all.
>>
>> I'm not so sure about that, the chances are that you have been
>> applying OO principles for many years
>> - we certainly were - well before Windows
>>
>> What is now called 'OO' is really a re-working and re-presentation of
>> a very old approach
>Writing OO code in a language that doesn't support it does not seem like a
>very good idea. Unless OOP is so outstanding that it would justify nearly
>any kind of ugly hack ;-)
I can't agree with you here
We found that the methods we used vastly simplified applications,
encapsulated functionality and gave us 'hidden flexibility'
There were no real 'hacks' or code contortions, it was just the
logical way to design the code.
>Noone is talking about a revolution here, only a *slight* improvement. The
>magnitude could possibly resemble that of proper use of character case
>compared to all upper- or lowercase. ;-)
>No, seriously, it's simply a model of grouping things. I believe I'm still
>evolving a sense of whether classes are "pure", in the meaning "whether
>something belongs in class A or B". I often divide classes into
>"type"-classes and "data" - classes, where the structures of the "types"
>control the instantiation of the "data" - classes. I just cannot see how I
>could have myself trained this way if writing non-OO code. It's like when
>constructing a house: "What elements do I connect, and what leans on what ?"
Yes, that makes sense
>While I do write non-OO code all the time, there is typically a movement
>towards objects as the project grows. The downside is of course that the
>decision on creating a class model may involve quite some work. But again
>and again I conclude that the error was writing all of that spaghetti
>instead of doing it properly from the beginning.
Often one starts something as a single instance, and then realizes
that it would make more sense wrapping it up and being able to
maintain multiple instances.
Also, however much one plans, one often 'discovers' something while
coding which (should) necessitate a partial re-write.
Also, one of the advantages I find from OO as it is now implemented,
is that one can 'hide' code and data away in the certainty that one
will not be tempted to re-use it 'a la bolognese'
>Regards,
>
>Bjørge Sæther
>bjorge@haha_itte.no
>-------------------------------------
>I'll not spend any money on American Software products
>until armed forces are out of Iraq.
>
>
>
- Next message: J French: "Re: TImage and GrayScale"
- Previous message: Rob Kennedy: "Re: TBytearray to variant"
- In reply to: Bjørge Sæther: "Re: D7 Compiler Bug"
- Next in thread: VBDis: "Re: D7 Compiler Bug"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]