Board index » kylix » Re: just thinking..

Re: just thinking..


2005-01-08 09:28:29 PM
kylix1
Quote

Lets hope for Kylix 4!

I don't believe in K4 any more :(
I suppose the future of RAD targeting the PC (and more {at least the
PDA}) is .NET.
What I hope for is as a first step a good way to have .NET programs
created with Delphi run in Mono (more realistic now as with Novell {and
IBM behind the curtain} strong companies are supporting Mono. As a
second step we might see the Delphi IDE run on .NET instead on the
Windows API and with that (or later) also on Mono (and thus on Windows,
Linux with different processors, Mac and IBM big iron).
-Michael
 
 

Re:Re: just thinking..

Michael Schnell wrote:
Quote
I don't believe in K4 any more :(
Don't know what to say but , I will buy K4 if it comes out in next 6 months,
else I'll abandon borland and use fpc & lazarus, or qt designer + gcc , or
at least vi & gcc ;)
Quote
I suppose the future of RAD targeting the PC (and more {at least the
PDA}) is .NET.
I've already heard that few years ago when Java was "in".
 

Re:Re: just thinking..

Michael Schnell < XXXX@XXXXX.COM >wrote:
Quote
>Lets hope for Kylix 4!
>
I don't believe in K4 any more :(

I suppose the future of RAD targeting the PC (and more {at least the
PDA}) is .NET.
MS .NET is window-centric moving target, will take years to mature and it is encumbered with patents like: www.internetnews.com/bus-news/article.php/3453371
I do not see how anybody in a right mind will prefer bloated, buggy, expensive .Net solution to make bloated, buggy and slow applications just to become more dependent from Microsoft, giving up native development, fast, reliable and easy coding with Delphi or Kylix. I don't want to be be lock in, I want to have a choice. I want my programs to be faster and smaller and based on truly open standards and protocols. What do you want?
Andreas
 

{smallsort}

Re:Re: just thinking..

Quote
Don't know what to say but , I will buy K4 if it comes out in next 6 months,
else I'll abandon borland and use fpc & lazarus,
Hope for many of us, but it's not likely that Lazarus is up to the
usability standard of -say- D7 within the next six months.
Quote
or qt designer + gcc , or
at least vi & gcc ;)

good for you, if you don't need to support a large code base that has
been written in Delhi.
Quote

>I suppose the future of RAD targeting the PC (and more {at least the
>PDA}) is .NET.


I've already heard that few years ago when Java was "in".
.NET outran JAVA for Desktop software desktop is very short time.
-Michael
 

Re:Re: just thinking..

Quote

MS .NET is window-centric moving target, will take years to mature and it is encumbered with patents like: www.internetnews.com/bus-news/article.php/3453371
Maybe somebody owns a patent that describes that it might be possible to
use a computer program to convert readable text into a computer
program and calls this compilation....
Quote
I do not see how anybody in a right mind will prefer bloated, buggy, expensive .Net solution to make bloated, buggy and slow applications just to become more dependent from Microsoft, giving up native development, fast, reliable and easy coding with Delphi or Kylix. I don't want to be be lock in, I want to have a choice. I want my programs to be faster and smaller and based on truly open standards and protocols. What do you want?

I was told that .NET desktop programs take longer to _load_ but are not
executing than natively compiled software.
Moreover see:
www.apcmag.com/apc/v3.nsf/0/22F356ECB65244E2CA256F6A00107C19
 

Re:Re: just thinking..

zeljko < XXXX@XXXXX.COM >wrote:
Quote
>I suppose the future of RAD targeting the PC (and more {at least the
>PDA}) is .NET.

I've already heard that few years ago when Java was "in".
In the 80s, we had e.g. USCD Pascal and Basic, in the 90s Java and now .NET. It was always said these interpreted byte code language are highly portable and allow the use of the generated programs on a lot of platforms. However, people forget that the compiler backend is the smallest part of cross platforming. More important is the native OS support and the GUI. But even that makes only a little part of libraries.
Examples:
FPC compiler (because I know for it all those numbers ;))
total: 300k lines
ARM specific code: 6k lines
SPARC specific code: 8k lines
x86_64 specific code: 2k lines
FPC's code quality is at the same level as gcc's on these platforms. From what I know about the gcc sources, the ratio between processor specific and common code might be the same.
FPC rtl:
total: 400k lines
linux specific: 10k lines
win32 specific: 10k lines
Lazarus/LCL
total: 1.200k lines
GTK specific code: 50k lines
Win32 specific code: 20k lines
You see that the processor, os and gui specific parts are very little compared with the whole code base, so where is the use of such byte code interpreters except that they an extra layer which can be buggy and that you don't have native application regarding look and feel etc.? Borland should know better, they should look at the 80s with USCD and Turbo Pascal :)
 

Re:Re: just thinking..

I think much functionality of the VCL will be moved to the .Net layer.
Borland has fewer to maintain.
The customer has to pay with "performance".
"Florian Klaempfl" < XXXX@XXXXX.COM >schrieb im Newsbeitrag
Quote

zeljko < XXXX@XXXXX.COM >wrote:
>>I suppose the future of RAD targeting the PC (and more {at least the
>>PDA}) is .NET.
>
>I've already heard that few years ago when Java was "in".

In the 80s, we had e.g. USCD Pascal and Basic, in the 90s Java and now
.NET. It was always said these interpreted byte code language are highly
portable and allow the use of the generated programs on a lot of platforms.
However, people forget that the compiler backend is the smallest part of
cross platforming. More important is the native OS support and the GUI. But
even that makes only a little part of libraries.
Quote

Examples:

FPC compiler (because I know for it all those numbers ;))
total: 300k lines
ARM specific code: 6k lines
SPARC specific code: 8k lines
x86_64 specific code: 2k lines
FPC's code quality is at the same level as gcc's on these platforms. From
what I know about the gcc sources, the ratio between processor specific and
common code might be the same.
Quote

FPC rtl:
total: 400k lines
linux specific: 10k lines
win32 specific: 10k lines

Lazarus/LCL
total: 1.200k lines
GTK specific code: 50k lines
Win32 specific code: 20k lines

You see that the processor, os and gui specific parts are very little
compared with the whole code base, so where is the use of such byte code
interpreters except that they an extra layer which can be buggy and that you
don't have native application regarding look and feel etc.? Borland should
know better, they should look at the 80s with USCD and Turbo Pascal :)
Quote

 

Re:Re: just thinking..

On 2005-01-08, Michael Schnell < XXXX@XXXXX.COM >wrote:
Quote
>Don't know what to say but , I will buy K4 if it comes out in next 6 months,
>else I'll abandon borland and use fpc & lazarus,

Hope for many of us, but it's not likely that Lazarus is up to the
usability standard of -say- D7 within the next six months.
Neither is _any_ .NET product on Mono. I think 2 years is already positive
for that.
Quote
good for you, if you don't need to support a large code base that has
been written in Delhi.
Did you already test run your codebase with Delphi.NET.
Quote
>>I suppose the future of RAD targeting the PC (and more {at least the
>>PDA}) is .NET.
>
>I've already heard that few years ago when Java was "in".

.NET outran JAVA for Desktop software desktop is very short time.
.NET is still in its infancy. It is hard to predict what will happen. However
I think the non-Windows .NET world will be extremely amateuristic for a long
way to come. Sure, Mono and dotgnu will slowly progress, and maybe very
clean console or winforms programs will run in time.
But medium and large scale VCL Delphi apps via the .NET and VCL.NET route?
No way, I think VCL.NET is doomed medium long term anyway. The D2005
introduction demo sessions made pretty clear that Borland considers VCL.NET
for porting and legacy reasons, and new apps should use winforms. (all the
demoes did, though sometimes they said "this also works with VCL.NET", and
even less often "also works with win32")
If mono and dotgnu ever come to age, VCL.NET will already be phased out.
 

Re:Re: just thinking..

Quote
>>or qt designer + gcc , or
>>at least vi & gcc ;)
>>
>good for you, if you don't need to support a large code base that has
>been written in Delhi.


Did you already test run your codebase with Delphi.NET.


Of course not. But supposedly it would be easier than converting it to
either of designer + gcc or vi & gcc.
-Michael
 

Re:Re: just thinking..

Quote
Neither is _any_ .NET product on Mono. I think 2 years is already positive
for that.

Not true. AFAIK, Novell already released a desktop solution based on Mono.
-Michael
 

Re:Re: just thinking..

1) MONO does work in Windows, too. So the cross-platform alternative is
not necessarily .NET framework vs MONO, but can be Mono on any platform.
2) The "Borland .NET drama" was not handled:
- Borland Delphi shows the paradigms that allow for .NET.
- MS hires the main Delphi Architect away from Borland and he helps to
develop .NET and C# (A language that is Delphi with C++ - similar
instead of Pascal derived syntax)
- Borland (supposedly as an exchange) is supplied with early
informations on .NET
- Borland develops Delphi for .NET (and then Delphi 2005). Same can
output CIL code that runs on .NET
- due to lack of working together between the Mono Group and Borland,
the Delphi generated CIL assemblies don't run on MONO (unless they are
extremely simple).
- So Borland is giving up on platform independancy, while they at the
same time stop serious support for Kylix, their native code Linux RAD
Delphi tool.
-Michael
 

Re:Re: just thinking..

Michael Schnell wrote:
Quote
- due to lack of working together between the Mono Group and Borland,
the Delphi generated CIL assemblies don't run on MONO (unless they are
extremely simple).
Lack of working together, or required by contract with Microsoft ? ;-)
Micha
 

Re:Re: just thinking..

Quote
Lack of working together, or required by contract with Microsoft ? ;-)

I try to just state the facts. In what I read from several Borlanders
there were no hints about bounds laid on them.
-Michael
 

Re:Re: just thinking..

Quote
Novell already released a desktop solution based on Mono.
called "iFolder3"
-Michael
 

Re:Re: just thinking..

g'day siegfried,
Quote
..what a Kylix 4 could be.

Both a Windows and Linux dev tool i.e. A Windows IDE and a Linux IDE
(like Delphi 7 Studio was for CLX).
you might want to check out Lazarus:
www.lazarus.freepascal.org/
Linux, Windows and mac are all supported. 64 bit is supported on Linux and will
eventually be supported on windows.
imho, it is not *quite* ready for prime time, but it is getting very close.
my own experience is that if you write an application from scratch in Lazarus
you can probably get it to work, but if you try to port an existing one there
are currently too many "gotchas". eg: no scanlines for tbitmap, tfont ignored in
things like labels, etc.
Borland seems to have focussed solely on dotnet to the expense of all other
platforms (win32, win64, linux, mac, etc).
cheers,
Mat