Board index » delphi » Re: An idle rant

Re: An idle rant


2005-01-18 03:11:11 AM
delphi68
Quote
In South Africa the word used is "orientated" (people look at you funny
if you say "oriented").
In normal speech that is true both here in the UK and at home in
Australia. In the specific context of programming, it is "oriented". I'm
sorry, but it is true.
Quote
Search for "Object Orientated" on the web and you will find quite a few
entries.
As you will for many other spelling errors :-)
Quote
Do the same thing in Amazon.com and again you will find books
that even have this in the title- e.g. "Design patterns: Elements of
reusable object orientated software" (Eric Gamma)
Again, a spelling error. I have that book in front of me, and that is
most definitely not the title, nor has it ever been. In all the
literature, the term is "object oriented" :-)
I actually did that search on amazon.com and got 7 entries, only 3 of
which had that in the title, and none of those had a image of the book
cover. However, if you check, you will find that there are gazillions of
"object oriented" entries, and all of those 3 titles come up, and all
have "oriented" on the cover.
Quote
Personally, I don't care which is used.
I shouldn't either, I suppose, but it bugs me for some completely
irrational reason.
Cheers,
Jim Cooper
_______________________________________________
Jim Cooper XXXX@XXXXX.COM
Falafel Software www.falafelsoft.com
_______________________________________________
 
 

Re: An idle rant

Craig Stuntz [TeamB] writes:
Quote
No, just that left alone the programmer might read The Onion instead
of programming....
hehe
 

Re: An idle rant

"Jim Cooper" <XXXX@XXXXX.COM>wrote
Quote
>They still program in straight "C" as they
>claim C++ (object orientated proramming) gives them nothing they need.
That's "object oriented" :-) And perhaps they're looking for the wrong
things. OOP is intended to make code more maintainable,
Which doesn't always work, for it can add complication without any
advantage. I think it is more apparent than scientific and engineering
non-RAD programming where the natural paradigm is operators operating on
data and binding data and code may not usually be a good idea. For instance,
sometimes, an array of numbers is just an array of numbers and you can do
radically different things to it. In such situations, creating and
maintaining a meaningful object hiearchy will consume a huge amount of your
time, even limit you and make maintenance harder, increase memory
requirements and slow down performance without any benefits. If you document
well and write clear code, strictly procedural programming can be just as
maintainable as OOP (and if you don't document and code well, OOP can be
just as hard when you have a soup of 100s or thousands of objects that
interact in complex ways).
That said, OOP is great, and almost mandatory for visual RAD type business
stuff nowadays.
Quote
which is the
most important virtue code can have.
 

Re: An idle rant

rick writes:
Quote
That means people sitting in front of it clicking, typing, etc. NO! WE
WANT AUTOMATED TESTING, the experts told me. I think that company went
out of business.
Back in my *** (name withheld) days we called that "Gorilla Testing". ;-)
We would get free pizza and such and in return had to hack and try
everyting that god forbade just to see how stable the application behaved.
Needless to say, we discovered a /lot/ of flaws (StrToInt exceptions etc
etc etc)...
--
Ben
 

Re: An idle rant

Quote
Which doesn't always work, for it can add complication without any
advantage.
Of course, there is definitely a skill involved. Good procedural
designers do not necessarily make the transition that quickly.
Quote
If you document well and write clear code, strictly procedural
programming can be just as maintainable as OOP
Well, that is not my experience. There are techniques that are more
difficult to apply to procedural code.
Quote
That said, OOP is great, and almost mandatory for visual RAD type business
stuff nowadays.
Hmmm. I'd argue that OO is at least as useful elsewhere, myself :-)
Cheers,
Jim Cooper
_______________________________________________
Jim Cooper XXXX@XXXXX.COM
Falafel Software www.falafelsoft.com
_______________________________________________
 

Re: An idle rant

"Ben Hochstrasser" <bhoc@tiscali123^H^H^H.ch>writes
Quote
Please bear in mind that the following is /not/ meant as D2005 bashing.

In my jobs I came across many different kinds of developers. One thing had
all in common: They worked with above-average well-equipped, fast machines
with dazzling graphics and ultra-fast server attachment. There's nothing
wrong with that.

Unfortunately, many of their products proved to be of rather poor
performance when run on an average office PC with average memory, disk and
graphics. It often proved to be unusable when connecting via the internet.
Today I run across this behaviour far more than the years before.
Well said!
I used to develop on a Pentium 166 until the middle of last year. Now I run
a 2ghz machine but I feel fairly secure in the knowledge that it is unlikely
to be outdated for a few years. I was teased horribly for using a 166mhz cpu
but as you say, if it'll run well on a p166, it'll run well on most things.
I look at the specs on some of the games I see around and they can use
massive quantities of memory, top-end graphics cards etc. The thing
everybody is missing with their headlong rush for the latest technology is
that the vast majority of users don't upgrade their computers until they
jolly well have to. Personally, I feel that we should concentrate on
producing software for the majority rather than the ultra expensive end of
the market.
I see so many software houses rushing to produce software for the upper end
of the market while forgetting the golden rule that the software must
actually work before it is released. Releasing it and using paying users as
beta-testers is, in my opinion, excruciatingly dishonourable. I recently ran
foul of this with my quest to find software that will write a vcd, svcd or
video dvd, turning digital photos into a slideshow that can be seen on TV.
As examples : Roxio's "easy CD and DVD creator" would crash the computer
spectacularly with its drag to disk feature and took 8 hours before it would
announce that it couldn't fit 1.5gb of files onto a 4.7gb DVD. Another
example: Sonic's vCD creator - goes through all the motions then crashes
with some program-generated error message despite the fact that (again) I
was trying to write less data than would fit onto a CD (in this case 300mb).
Photo2DVD will allow me to orient a photo vertically but will then stretch
it to fit the screen width instead of putting a border. As the latter is
shareware, I cannot imagine they get many customers. These are the milder
bits of bad software. A company calling itself Direct-Soft inc managed to
code their AVI to MPEG recoding software so badly (I cannot imagine how they
managed it) that it destroyed XP's capability to do a system restore and
corrupted the system so badly the only answer was to install XP onto another
hard drive and then copy the data accross before reformatting the original
drive.
It's giving programming and software a very poor reputation. Oddly enough, I
never hear of this from mac software so the next computer I buy for my own
personal use (as opposed to for development) will almost certainly be a Mac.
As I see it, the strength of the Mac is that its market share is smaller,
leading to fewer but better software developers and more dedicated users.
 

Re: An idle rant

Rhys Sage writes:
Quote
It's giving programming and software a very poor reputation. Oddly
enough, I never hear of this from mac software so the next computer I
buy for my own personal use (as opposed to for development) will
almost certainly be a Mac. As I see it, the strength of the Mac is
that its market share is smaller, leading to fewer but better
software developers and more dedicated users.
My mac (yeah I have one, I didnt buy it, it was err, a donation from a
job I did they gave me it) its a laptop, its I believe a g4, it has I
think 128mb memory, maybe more, Im not firing it up to find out, and a
6gb disk. Mac apps tend to be much smaller than its PC counterpart, or
even its linux counterpart with OSX.. I assume that a lot of it is the
OS underneath it has more feature built in so less app is needed to
generate it.. Ive yet to code on my mac, I thought about it, all the
OSX dev stuff is on it (so yes the disk is basically full).. But
perhaps because they are used to working with smaller confines they
have more predictable levels of computer kit, eg, g3, g4.. and can be
sure it will work at a reasonable rate if tested on it.
 

Re: An idle rant

Liz writes:
Quote
Mac apps tend to be much smaller than its PC counterpart, or
even its linux counterpart with OSX.. I assume that a lot of it is the
OS underneath it has more feature built in so less app is needed to
generate it..
One reason for this are simply the huge runtime dlls hidden in the O/S.
Just think of the small VB(Visual Basic) applets vs. the "fat" delphi apps, just because
MS bundled their VBRUNxxx.DLL while Borland couldn't...
--
Ben
 

Re: An idle rant

"somebody" <XXXX@XXXXX.COM>writes
Quote
"Jim Cooper" <XXXX@XXXXX.COM>wrote

>>They still program in straight "C" as they
>>claim C++ (object orientated proramming) gives them nothing they need.

>That's "object oriented" :-) And perhaps they're looking for the wrong
>things. OOP is intended to make code more maintainable,

Which doesn't always work, for it can add complication without any
advantage. I think it is more apparent than scientific and engineering
non-RAD programming where the natural paradigm is operators operating on
data and binding data and code may not usually be a good idea. For
instance,
sometimes, an array of numbers is just an array of numbers and you can do
radically different things to it. In such situations, creating and
Usually an array of numbers has some meaning <G>. If you treat them as a raw
list of numbers, then no, encapsulation won't help you at all. But if you
treat that array for what it is, tt may be a chromatogram, or a list of
energies, or whatever, then more often than not you will find that there
are a finite number of operations that are performed on that dataset.
For example, if you look at a raw chromatogram, you will see an array which
has one value for each reading taken, and you'd plot the the values
sequentially, (with the X value being the array index or "scan number").
But you still need to know what the scan interval is, and the scaling factor
for the values, and the instrument the data was acquired from, plus the date
and time and analyst, etc. This all goes into a header. Then you can
perform operations on the chromatographic data, such as smoothing, peak
detection, etc. Creating a TChromatogram class to keep this all together
makes the process SO much easier - trust me, I did it the non OOP way first,
and it became a nightmare. you will find similar results from most other
scientific and engineering data.
And for you non-chemists out there <G>
www.google.com/search&lr=&oi=defmore&q=define:Chromatogram
 

Re: An idle rant

Jim Cooper writes:
Quote
I shouldn't either, I suppose, but it bugs me for some completely
irrational reason.
LOL!
 

Re: An idle rant

Jim Cooper writes:
Quote
>Search for "Object Orientated" on the web and you will find quite a few
>entries.

As you will for many other spelling errors :-)
Good point. However, in this particular case I believe that it is only
"Object Oriented" because of the impact that the USA has on the rest of
the world regarding computer terminology. If you applied this logic to
everything computer related it would be wrong to call the colours on the
monitor colours. You'd have to call them colors. :-)
Cheers,
Kevin.
 

Re: An idle rant

Quote
Indented!!! Doesn't mean that is what happens, and from what I have seen here
at various times, OOP really made no difference. OOP promised alot of
things - but again I think it was all hype. There's no way anyone can
convince me OOP code is more maintainable than non oop code. it is all down
to the skill of the programmer in the end.
yup, actually OOP can make code less readable, -
like those endless inherited inherited etc..
 

Re: An idle rant

Quote
Maintainable code makes optimisation
easier, IME. At least in the OO world.
huh?
 

Re: An idle rant

Quote
However, in this particular case I believe that it is only
"Object Oriented" because of the impact that the USA has on the rest of
the world regarding computer terminology.
Oh definitely. Which is why we have computer "programs". However, the
usage has stuck in all the literature, in both cases.
Quote
You'd have to call them colors. :-)
Yeah, I hate that. And the "Math" unit :-)
Cheers,
Jim Cooper
_______________________________________________
Jim Cooper XXXX@XXXXX.COM
Falafel Software www.falafelsoft.com
_______________________________________________
 

Re: An idle rant

Jim Cooper writes:
Quote
Oh definitely. Which is why we have computer "programs". However, the
usage has stuck in all the literature, in both cases.
I avoid having to use "programs" by calling them all "applications"
rather! :-)
Cheers,
Kevin.