Board index » kylix » Re: The future..

Re: The future..


2004-01-24 01:54:31 PM
kylix1
Quote
That is the problem, I have seen much too often in the past where I worked
more closely with C/C++. Pointers being passed around everywhere and it is
very easy to lose track where the first allocation took place.

Ahhh the classic C+ programmer disease. I was stuck in that mode for
a few years. Hard to break old habits :)
Quote
Well, different strokes for different folks. I actually like the OOP
structure MUCH BETTER.. IT enforces better use of design.

True but it sucks having to code a class when what you really want to
do can be encapsulated in a simple function.
Quote
The only exception is when dealing with primitive types. Then Object
allocation can be a waste of precious memory (bet I get attacked on that
one <G>).

No you are dead on. The fact the java doesnt have true primitive
types hurts its performance as I understand it. Then again, the
object allocation/deallocations in C# and java are heavily optimized
and are just moving pointers around most of the time i'm sure. Thats
the one place they mop up C++ with the standard allocator :)
Quote
Please note, I am not against either C or C++. They both have their places
and especially when doing low level work, they are a better option than
Java. For general Networked based development and/or general application
development, Java is better overall, IMHO, since it inherently enforces
structure and it lends itself to much more rapid development and
deployment. Of course if you add xplatform, Java definitely wins. C is
portable as long as you can avoid GUI and processor specific instructions.

Couldnt agree more. I'd never build a desktop GUI or server side app
in C at this point. Its just to prone to bugs and security lapses.
There's still lots of life left in C though on other platforms.
C++ still makes a fine all around langage for just about anything.
Between Boost, STL, and wx/qt you can pretty much build an app that
will run on just about any platform. It's not as consistent as java
but it doesnt come with all the warts either. My three biggest gripes
with java are lack of deterministic destructors, JNI, and the fact the
each process has to have its own VM.
 
 

Re:Re: The future..

Mike Margerum wrote:
Quote
Ahhh the classic C+ programmer disease. I was stuck in that mode for
a few years. Hard to break old habits :)

LOL!! I litrally hated to have to come behind someone that was pointer
happy. I felt like they were playing "Dirty Harry" and I was "making their
day" <G>. They were certainly making mine, and I was like the singer in the
song, "Ain't no sunshine when your gone" :)
Quote

True but it sucks having to code a class when what you really want to
do can be encapsulated in a simple function.

Yes I can agree with that. You have to get used to the class stucture part
and how I kinda "fake" regular function calls, is to create a class that
acts like a class template. Not difficult to do, just a little different.
I, of course try and group common functionality and not all things fall into
specific groups (as we all know, and those who are {*word*60}ed to UML will
soon discover :)). Of course I have seen some OOP Java programmers who are
as bad at normalizing classes as some DBAs do databases :). In these cases,
you have to ask the question, "How do I get there from here?" Answer? Step
into my maze :):)
Quote
>The only exception is when dealing with primitive types. Then Object
>allocation can be a waste of precious memory (bet I get attacked on that
>one <G>).
>
No you are dead on. The fact the java doesnt have true primitive
types hurts its performance as I understand it. Then again, the
object allocation/deallocations in C# and java are heavily optimized
and are just moving pointers around most of the time i'm sure. Thats
the one place they mop up C++ with the standard allocator :)

Well Java does have primitives, for the more common types. int, long, float,
double, char,and boolean are all primitives iu Java. They also have their
class countertypes, Integer, Float, Long, Double, Boolean.
The only thing that I wish Java handled better is Strings. There are some
pretty powerful functionality with String classes, but if you are not
careful, typecasting and concatenation can cost you some real heap space.
Of course NET, in its Java clone fashion, suffers from this same malady.
Quote

C++ still makes a fine all around langage for just about anything.
Between Boost, STL, and wx/qt you can pretty much build an app that
will run on just about any platform. It's not as consistent as java
but it doesnt come with all the warts either. My three biggest gripes
with java are lack of deterministic destructors, JNI, and the fact the
each process has to have its own VM.
I understand the deterministic destructor part, but theoretically, whenever
an object goes out of scope, it should be gced in the nearest order or pass
of the next gc cycle. I know it is hard to determine exactly when that GC
operation occurs, but there is way of forcing gc operations. Of course, you
are probably using or missing, I should say, dd as when declaring Object
scope and the object (upon losing scope), not only calls its destructor,
but also performs other clean up operations. For example, closing a file
when a file object has gone out of scope.
The latter type example is about the only time I can see this as being
advantageous however. The very nature of a gced language inherently frees
the object once it is out of scope, or at least makes it available for gc.
You can, of course force gc, or do as I generally do, set the object to
null in the finally or finalization section of a class.
 

Re:Re: The future..

Mike Margerum wrote:
Quote
C++ still makes a fine all around langage for just about anything.
Between Boost, STL, and wx/qt you can pretty much build an app that
will run on just about any platform. It's not as consistent as java
but it doesnt come with all the warts either. My three biggest gripes
with java are lack of deterministic destructors, JNI, and the fact the
each process has to have its own VM.
BTW, it would seem you are used to working with extreme memory limitations
on embedded devices. In this case, i can certinly understand why you are
more dispoosed toward C and C++. I am sure you have a better understanding
of this type of low level memory management, than do I. My C/C++ is rusty,
to say the least, and extreme memory management I can truly say, is no
longer my forte.
Keep going however, you may bring back the good ole days of 16-64K machines
when this was vitally important, even to the PC programmer :).
 

{smallsort}

Re:Re: The future..

"Mike Margerum" < XXXX@XXXXX.COM >wrote in message
Quote
China's already playing hardball. They are coming up with a competing
standard to 802.11 and if you want to sell wireless hardware there you
have to implement it and worse yet pay royalties to one of 6 "offical"
companies with rights to the IP. They dont respect anyone elses IP
but we are supposed to honor theirs. Things are gonna get real
interesting over there.
It seems that China is having fun toying with western companies; seeing how
far they will go just for the *hope* of gaining access to their market.
They hold all the cards .... for the time being.
 

Re:Re: The future..

JQP wrote:
Quote
"SiegfriedN" < XXXX@XXXXX.COM >wrote in message
news:4007f8e8$ XXXX@XXXXX.COM ...
>It is always a dilema today to decide what to invest your future into.
>Time is too precious to waste on technology that may not be around in 3
>years time.

I couldn't agree more. Time is limited so use it wisely.

One simple rule to keep in mind when evaluating technology --- it mostly
only moves forward, occasionally sideways and hardly ever backwards.

To facilitate a lateral move, the transition must be relatively simple,
pain
free and easily justifiable in economic terms. The available evidence
shows that the growth of the Linux server has been primarily a lateral
move from proprietary Unix, mostly Solaris.

On the desktop, moving from Windows to Linux does not fit the lateral move
criteria IMO. For the average user, it is a step backwards in terms of
usability. China aside, for most of the rest of the world, the benefits
aren't worth the effort under current conditions and I'll bet money that
it won't happen to any significant extent.
China alone? You had better update your list of countries <G>. Other
countries moving to Linux (Desktops and Servers) include:
Japan
Korea
Israel
Brazil
India
Germany
 

Re:Re: The future..

"pnichols" < XXXX@XXXXX.COM >wrote in message
Quote
China alone? You had better update your list of countries <G>. Other
countries moving to Linux (Desktops and Servers) include:

Japan
Korea
Israel
Brazil
India
Germany
Yes, with all these "countries and nations" moving over, Linux may finally
break through the 3% barrier this year and overtake Apple in terms of
desktop marketshare.
But then again, people have been predicting this for years and it hasn't
happened yet.
 

Re:Re: The future..

JQP wrote:
Quote
"pnichols" < XXXX@XXXXX.COM >wrote in message
news: XXXX@XXXXX.COM ...
>China alone? You had better update your list of countries <G>. Other
>countries moving to Linux (Desktops and Servers) include:
>
>Japan
>Korea
>Israel
>Brazil
>India
>Germany

Yes, with all these "countries and nations" moving over, Linux may finally
break through the 3% barrier this year and overtake Apple in terms of
desktop marketshare.

How do you know that Linux has not already taken over Apple's marketshare?
If you are going by "sold copies only", you lose. One does not have to
purchase a copy of Linux for every work station and server they own. In
fact, they do not have to purchase a single copy period.
By most estimates, Linux has surpassed Apple's meager desktop marketshare
some time ago. The only studies that show differently, are doing so by a
"purchased" Linux type of study, which is quite inaccurate.
For instance, we are running Linux on all of our workstations and servers
and have only purchased one copy at any given time. That means for 7
workstations and notebooks, and for 3 servers, we register as purchasing
one copy which means 1 machine is running Linux, which is totally
inaccurate, since we are actually running Linux on 10 machines. If you ask
anyone else that is running Linux in their company and even homes, that
would reveal similar figures to much greater actual installed base figures.
This does not even account for those running Fedora core, Debian, or
Mandrake from a download where they purchased nothing. In fact, it is quite
impossible to give an accurate count of those running Linux. It is easier
with Windows, since nearly every machine you purchase MUST come with
Windows, such as those from Dell and Gateway, even though Windows may be
uninstalled as soon as they receive the notebooks or Desktops.
We have five such machines that came with Windows, and ran Windows only long
enough to make sure they functioned properly. Now they all run Linux, and
guess what? Microsoft gets to claim that these machines are all running
Windows, when in reality, they are not running Windows, but rather, are
running Linux.
Besides, these countries represent many more people than exists in the US
and Canada combined.
Quote
But then again, people have been predicting this for years and it hasn't
happened yet.
It has already happened. Think about these populations for a moment. You
might want to pull out your Worldbook's and check the populations in these
countries.
 

Re:Re: The future..

"pnichols" < XXXX@XXXXX.COM >wrote in message
Quote
How do you know that Linux has not already taken over Apple's marketshare?
The last survey I saw was done by IDC (International Data Corp).
Various Linux distro vendors (Suse for example) have commented that they
accept IDC as the best available numbers.
But I'm sure you have something better<g>.
Quote
It has already happened.
In your mind.
Quote
Think about these populations for a moment. You
might want to pull out your Worldbook's and check the populations in these
countries.
Does it list the % that have ever touched a computer? Population is
totally irrelevant.
 

Re:Re: The future..

JQP schrieb:
Quote
"pnichols" < XXXX@XXXXX.COM >wrote in message
news: XXXX@XXXXX.COM ...

>How do you know that Linux has not already taken over Apple's marketshare?


The last survey I saw was done by IDC (International Data Corp).

Various Linux distro vendors (Suse for example) have commented that they
accept IDC as the best available numbers.

But I'm sure you have something better<g>.


>It has already happened.


In your mind.


>Think about these populations for a moment. You
>might want to pull out your Worldbook's and check the populations in these
>countries.


Does it list the % that have ever touched a computer? Population is
totally irrelevant.


Note: China has in the meantime more internet user, as whole humans in
USA, in Germany there are more as 44% online, did you know this?
regards
Gerhard
 

Re:Re: The future..

JQP schrieb:
Quote
"pnichols" < XXXX@XXXXX.COM >wrote in message
news: XXXX@XXXXX.COM ...

>How do you know that Linux has not already taken over Apple's marketshare?


The last survey I saw was done by IDC (International Data Corp).

Various Linux distro vendors (Suse for example) have commented that they
accept IDC as the best available numbers.

But I'm sure you have something better<g>.


>It has already happened.


In your mind.


>Think about these populations for a moment. You
>might want to pull out your Worldbook's and check the populations in these
>countries.


Does it list the % that have ever touched a computer? Population is
totally irrelevant.


Note: China has in the meantime more internet user, as whole humans in
USA, in Germany there are more as 44% online, did you know this?
regards
Gerhard
 

Re:Re: The future..

JQP schrieb:
Quote
"pnichols" < XXXX@XXXXX.COM >wrote in message
news: XXXX@XXXXX.COM ...

>How do you know that Linux has not already taken over Apple's marketshare?


The last survey I saw was done by IDC (International Data Corp).

Various Linux distro vendors (Suse for example) have commented that they
accept IDC as the best available numbers.

But I'm sure you have something better<g>.


>It has already happened.


In your mind.


>Think about these populations for a moment. You
>might want to pull out your Worldbook's and check the populations in these
>countries.


Does it list the % that have ever touched a computer? Population is
totally irrelevant.


for your information:
country Internet-User(bill.) all people(bill.) Internet-user in %
1. Japan 86,6 127,3 68%
2. Norwegen 2,8 4,5 63%
3. Dänemark 3,3 5,4 62%
4. Kanada 18,3 30,5 60%
5. USA 160,4 281,4 57%
6. Niederlande 8,4 16,1 52%
7. Süd-Korea 24,0 47,0 51%
8. Australien 9,0 18,7 48%
9. Singapur 2,0 4,1 48%
10. Finnland 2,3 5,2 45%
11. Hong Kong 3,0 7,0 43%
12. Israel 2,5 6,2 40%
13. Taiwan 8,8 22,0 40%
14. Irland 1,5 3,9 39%
15. Deutschland 29,7 82,4 36%
16. Großbritannien 20,4 60,0 34%
...
24. China 297 1292 23%
Global e-Commerce-Report 2001
 

Re:Re: The future..

JQP schrieb:
Quote
"pnichols" < XXXX@XXXXX.COM >wrote in message
news: XXXX@XXXXX.COM ...

>How do you know that Linux has not already taken over Apple's marketshare?


The last survey I saw was done by IDC (International Data Corp).

Various Linux distro vendors (Suse for example) have commented that they
accept IDC as the best available numbers.

But I'm sure you have something better<g>.


>It has already happened.


In your mind.


>Think about these populations for a moment. You
>might want to pull out your Worldbook's and check the populations in these
>countries.


Does it list the % that have ever touched a computer? Population is
totally irrelevant.


next information:
In 89 days the EU will have more then 500 millions peoples,
and i think in 5 years are more then 200 millions online in the EU, and
LINUX has much more acception in europe in governments and companies.
Some local governments will change to LINUX (is in plan now, or in the
phase of implementation).
.....
next information:
"Most interest in Linux has been in Australia, China, Japan and Korea
and this momentum is coming from governments rather than enterprises,"
said Robin Simpson, Gartner Australia's research director.
.........
Linux and Linux Desktop will grow and grow in the next years, more then
every other thing,
thats my believe.
regards
Gerhard
 

Re:Re: The future..

JQP schrieb:
Quote
"pnichols" < XXXX@XXXXX.COM >wrote in message
news: XXXX@XXXXX.COM ...

>How do you know that Linux has not already taken over Apple's marketshare?


The last survey I saw was done by IDC (International Data Corp).

Various Linux distro vendors (Suse for example) have commented that they
accept IDC as the best available numbers.

But I'm sure you have something better<g>.


>It has already happened.


In your mind.


>Think about these populations for a moment. You
>might want to pull out your Worldbook's and check the populations in these
>countries.


Does it list the % that have ever touched a computer? Population is
totally irrelevant.


next information:
The United Kingdom and Russia this week become the
latest governments to embrace Linux and open
standards, joining more than 20 U.S. states and 175
governments around the world.
In two separate deals illustrating the unprecedented
adoption of Linux by governments, IBM has announced
that it is working with the UK and Russia on Linux
projects that will further establish Linux in the
public sector.
regards
 

Re:Re: The future..

Hello,
for me and many others, i think,
Kylix is much more important as .NET for the next years,
and especially for 2004.
1)
As i write classic Client/Server programms with databases
(and this is i think 70% of the Delphi-programms)
i am not really interested in the .NET thing.
(show 4000 records in a .NET Grid? it's horrible,
because there are 4000 instances of an object in .NET)
The other thing is: If you develop .NET why with Delphi 8.NET?
Its nearly the same as Visual Studio from Microsoft. And you remember
the troubles with winapi-programming (documentation in C++, and many
other prolbems, and integration of microsoft office, thats was allways
better with visual Basic, sorry, but this is the true).
AND it's only a dream to update your existing
source-code (100.000's of lines) wiht Delphi .NET.
Therefore you have to start at ground zero. Many years of development
are lost in the trash (1000's of good VCL-components too). NO THANKS.
Longhorn will be Win32 compatible for many years, on the other side
longhorn will be dead before born (and this is not a developer wish,
this is market situaion).
conclusio: Win32 will live 10 years from now, i am sure,
and if .NET will be really an alternative in some years, it's a
question. (Asp.NET? lol, PHP, java, Apache, MySQL and many other things
are very good ways to make good web-services)
2)
Linux: 2004 will be the year of linux, and 2005, 2006 much more.
Many companies and goverments are changing from windows to linux.
Not, because Linux is so much more better as Windwos.
There are two reasons: 1)They don't have the money to follow microsoft
marketing plans. 2) Microsoft thinks: all computers will be online do
work with microsoft and pay for any thing they need ->totaly depentence
from one company: NO and NO!
conclusio: To port existing Delphi-Windows-Programms to LINUX is much
more interessant for delphi developers, as to go to .NET.
AND: If there is a really good thing to do so (and this is possible,
look Kylix), THE DELPHI DEVELOPERS will pay for it !!!
What we need is a Kylix 7, with an IDE that is a real LINUX IDE,
and where Delphi 7 code can be 1:1 used (without win32 special things,
thats clear).
I can understand Borland, to support early .NET. It's a new technology
and Borland have to suppurt this. But i got feedback from some old
Delphi-Developes, that they change now from Delphi to Visual Studio
(Microsoft) if they will make .NET development, because they see no
advantance to stay with delhpi .NET. (It's expensive too, there are only
little diffrences, but some heavy bad things (not supporting pocket
pc's, and so on).
But i cannot understand that Borland don't support for
the Delphi-developers Linux. Delphi' developers are Windows developers,
till 2004, and Borland don't make much money with Kylix. But now this
things changes. And exactly in this moment Borland let us alone.
Give us (Delphi developers) a good thing to work on Linux too,
PLEAAAAAAAAASE!
That we (delphi developers) can make Win32 and Linux applications would
be a nice thing! To make NEW .NET applications with Delphi too, is also
a nice thing!
This are 2 important reasons do be a DELPHI developer in the future!
But if the 1) is not possible, then we have to search for a new
Developing Tool on Linux. Delphi crossplatform-developing is then
lost, and if i have to decide to have .NET developing on one side,
and LINUX-developing on the other side, i will take for bouth sides
the best tools, and this will be then Microsoft tools for microsoft
world, and linux tools for Linux world. end of the story.
I am a fan from Delphi since 1.0 (and before Turbo Pascal) and i could
make my job very good with Borland till now.
I thank Borland for the good years with our wonderfull Delphi!
(hope that Borland don't forget us, but this hope will be smaller and
smaller)
best regards
gerhard
 

Re:Re: The future..

JQP wrote:
Quote
"pnichols" < XXXX@XXXXX.COM >wrote in message
news: XXXX@XXXXX.COM ...
>How do you know that Linux has not already taken over Apple's
>marketshare?

The last survey I saw was done by IDC (International Data Corp).

Various Linux distro vendors (Suse for example) have commented that they
accept IDC as the best available numbers.

But I'm sure you have something better<g>.

I just gave you one real world examp[le where the figures will not work. I
think Mike and R.F. gave you the same type figures that I gave in their
companies on even a larger scale.
So, since MS was preinstalled on three of my machines, then MS can claim I
am running 3 Windows machines (when actually neither of these machines is
running Windows). Linux is only running on one of my machines, according to
IDC, when in acutality, it is running on 10.
I know MOST Linux users buy a new copy for each machine they are running as
do the Enterprises that use Linuxc right (according to your idealistic
survey quote). The problem is, I know absolutely NO ONE who does this,
Quote
>It has already happened.

In your mind.

In reailty, you can look and see every single solitary report about
countries running Linux and abandoning Windows. But you will not believe
these governments own reports, so what can I or anyone else say that would
convince you.
Quote
Does it list the % that have ever touched a computer? Population is
totally irrelevant.
If you took even modest figures, that would say only 20% of the people in
these countries used computers, you would be somewhere in the 200,000
million plus range. I would think that to be relevevant.
If you assumed only 10% of the people actually owned a computer, you would
be well over 100 million. That is definitely a lot of computers.
Of course MS, is not worried about these countries changing over at all, are
they? That is why Bill Gates and Steve Ballmer have been flying to these
countries and giving grants, pow wows and attempts to make special deals
with them, is it not? Is that why MS agreed to let China, India, and others
have access to the MS code, because these markets are totally unimportant
and irrelevant?
We all know that these countries choices in Software is unimportant, so why
don't you share your superior knoweldge of marketshare, and be a benevolent
guy and convey this to MS? I am sure Gates and Ballmer would appreciate
knowing that these countries and marketshares do not count and are totally
irrelevant. It could save them alot of flight time and they would no
longer have to offer give away software programs. :):)