Re: Help Constructing Fictional Cross-Religious Movement
- From: whheydt@xxxxxxxxxxx (Wilson Heydt)
- Date: Thu, 8 Sep 2005 00:07:52 GMT
In article <9dEte38Hw-B@xxxxxxxxxxxxxxxxx>,
Kai Henningsen <kaih=9dEte38Hw-B@xxxxxxxxxxxxxxxxx> wrote:
>whheydt@xxxxxxxxxxx (Wilson Heydt) wrote on 26.08.05 in <ILurrC.DJJ@xxxxxxxxxxx>:
>
>> In article <m264tsjoak.fsf@xxxxxxxxxxxxxxxxxxxxxxx>,
>> Charlton Wilbur <cwilbur@xxxxxxxxxxxxxx> wrote:
>> >>>>>> "WH" == Wilson Heydt <whheydt@xxxxxxxxxxx> writes:
>> >
>> > WH> In article <m2fysxy35q.fsf@xxxxxxxxxxxxxxxxxxxxxxx>,
>> > WH> Charlton Wilbur <cwilbur@xxxxxxxxxxxxxx> wrote:
>> >
>> > >> Does this articulate the difference satisfactorily? It's not
>> > >> that object orientation offers things you can't do with
>> > >> subroutines and functions, but that it offers a way to break up
>> > >> large programs into smaller subprograms and minimize the
>> > >> interdependencies between those subprograms.
>> >
>> > WH> I'm afraid it doesn't, becasue I could do everything you've
>> > WH> described (inclduing hiding the actual implementation) in
>> > WH> COBOL using subroutines. If I can do the same things you've
>> > WH> described using subroutines, what is the functional
>> > WH> difference?
>> >
>> >Ease of implementation, especially once you get into method dispatch,
>> >inheritance and introspection (the ability of objects to know what
>> >they are and what they can do).
>>
>> No argument there. IF you play to any languages strengths, it will
>> do a better job than if you play to it's weaknesses--and COBOL is
>> certainly ill-adapted to do things that OO languages are good at.
>
>I'm sure I heard about object oriented COBOL.
Sometimes referred to as "ADD 1 TO COBOL".
>> Yes...but... While business processes some business processes may
>> be more tractable to being written using OO methods (how are OO
>> languages at writing reports?), I have yet to see a clearly
>
>Report writing sounds like a library feature, at least if it goes beyond
>the trivial, not like a language feature. As for language feature level,
>C++ still has printf, but it also has operator<< which can be made to
>output arbitrary types in arbitrary ways. (Incidentally, the Boost library
>collection has a combination.)
COBOL has a major, and almost never used feature called (oddly
enough) REPORT-WRITER specifically intended to make writing reports
easy. Almost all reports written in COBOL are hand coded, however,
and most systems have more reports than any other single type of
program.
>> exlicated data type that is nearly univerasl in business
>> applications demonstrated in an OO language: fixed point number
>> with decimal fraction.
>
>You can, of course, do it. I'm actually currently doing something slightly
>similar: a time data type (or rather, a combination point-in-time and time-
>duration library with assorted minor other stuff) which is based on a 64
>bit nanosecond counter, and meant to work on both Linux and Windows. First
>implementation was as C data types (with a bit of Objective-C help); I'm
>converting it to a C++ data type, which should make the interface slightly
>more natural but means I need to implement stuff like addition and
>subtraction.
Yabbut.... I COBOL, I don't *need* to write or have someone else
write library routines. All I have to do is declare a variable to
have a specifc number of decmial places and I can do all the
arithmetic on it I want, e.g.:
77 FIXED-NUMBER PIC S9(5)v9(4) COMP-3.
--
Hal Heydt
Albany, CA
My dime, my opinions.
.
- Follow-Ups:
- Re: Help Constructing Fictional Cross-Religious Movement
- From: Brooks Moses
- Re: Help Constructing Fictional Cross-Religious Movement
- Prev by Date: Re: Help Constructing Fictional Cross-Religious Movement
- Next by Date: Re: A newcomer here
- Previous by thread: Re: Help Constructing Fictional Cross-Religious Movement
- Next by thread: Re: Help Constructing Fictional Cross-Religious Movement
- Index(es):
Relevant Pages
|
