Board index » cppbuilder » Re: Interesting Borland commentary on the future of Win32 programming

Re: Interesting Borland commentary on the future of Win32 programming


2004-05-14 03:02:15 AM
cppbuilder24
"Pete Fraser" < XXXX@XXXXX.COM >wrote in message
Quote
Under .NET any .NET source file can interact with any other .NET source
file
directly, No COM/OCX etc.
This means that VB/C++/C#/Delphi source can be mixed in a project.(If
that's
something you want!)
Also, the target platform is not important (as long as a .NET CLR is
available for it such as in MONO for linux-when it's finished).
If I understand things correctly, you can compile code from your language of
choice into some intermediate language, that gets interpreted by the CLR.
The code works with framework classes in .Net. Any kind of low level work is
thus out of reach from you (unless you use unmanaged C++ code?) You lose
lots in performance as well. On the plus side you gain garbage collection an
maybe you write products quicker? You can mix source languages (is that good
or bad?) Maybe security can be better managed?
I don't see the draw though, other than to keep up with M$ latest tech. My
feeling is that it's a scheme to continue milking computer users for more
cash. What problems is it solving (or trying to solve), and who (if anyone)
*should* use it?
Simon.
 
 

Re:Re: Interesting Borland commentary on the future of Win32 programming

Gord wrote:
Quote
>When the IBM PC came out I was doing some programming on a CPM machine and
>an Apple II+. When I decided to learn more about programming I looked around
>and although I liked Apple, businesses used the PC. So, I went to where the
>work was. Most of by current customers wouldn't know DOT NET from a fishing
>net. I simply provide a solution to their problem and I'll use whatever
>tools fit best into their environment. I guess I'm a lemming ;-).

Yup, when I first wrote my main program it was on a main frame and mostly for
personal use. I then then ported it to a macintosh. After people started
getting wind of it (mostly engineers and mining companies), they wanted to buy
it - but not on a Mac - on a PC. All my present and prospective customers use
Windows, none have *nix.
Exactly the same here, but many of them are still running 98SE and have
no plans to change that. XP is just OTT without a complete change of
hardware, and replacing 25,000 machines?
SO I repeat the statement - why be a lemming - I STILL want a Builder*5*
that actually works :)
--
Lester Caine
-----------------------------
L.S.Caine Electronic Services
 

Re:Re: Interesting Borland commentary on the future of Win32 programming

Yes, that's correct except that apparently you *don't* lose performance.
M$ said at a recent ACCU conference that the time penalty was only about 15%
and could be faster on Centrinos.
However, whether that's actually true we shall have to wait and see.
Rgds Pete
"Simon D" < XXXX@XXXXX.COM >wrote in message
Quote
"Pete Fraser" < XXXX@XXXXX.COM >wrote in
message
news:40a3910c$ XXXX@XXXXX.COM ...
>Under .NET any .NET source file can interact with any other .NET source
file
>directly, No COM/OCX etc.
>This means that VB/C++/C#/Delphi source can be mixed in a project.(If
that's
>something you want!)
>Also, the target platform is not important (as long as a .NET CLR is
>available for it such as in MONO for linux-when it's finished).

If I understand things correctly, you can compile code from your language
of
choice into some intermediate language, that gets interpreted by the CLR.
The code works with framework classes in .Net. Any kind of low level work
is
thus out of reach from you (unless you use unmanaged C++ code?) You lose
lots in performance as well. On the plus side you gain garbage collection
an
maybe you write products quicker? You can mix source languages (is that
good
or bad?) Maybe security can be better managed?

I don't see the draw though, other than to keep up with M$ latest tech. My
feeling is that it's a scheme to continue milking computer users for more
cash. What problems is it solving (or trying to solve), and who (if
anyone)
*should* use it?
 

{smallsort}

Re:Re: Interesting Borland commentary on the future of Win32 programming

Gord wrote:
Quote
If a lemming means providing a product for customers that are there (Windows users) as
opposed to providing a product for customers that aren't there (*nix users), then I
guess I'm a lemming.

I agree. I just like the word play.
 

Re:Re: Interesting Borland commentary on the future of Win32 programming

"Simon D" < XXXX@XXXXX.COM >wrote:
Quote
If I understand things correctly, you can compile code from your language of
choice into some intermediate language, that gets interpreted by the CLR.
The CLR does JIT compiling; it's never interpreted. Even Mono, which
has an interpreted mode, converts the IL (intermediate language) into
a different intermediate language just to make interpreting *easier*
(not faster, just easier).
Quote
The code works with framework classes in .Net. Any kind of low level work is
thus out of reach from you (unless you use unmanaged C++ code?)
C# has an unsafe code option, allowing you to use pointers directly in
your functions. This means that the assembly containing the code needs
to be trusted, however (for example, the default security policy means
that an assembly using unsafe code can't run from a network location
like a network drive or the Internet, for example - either the policy
must be changed or the whole app copied to the local drive).
Quote
You lose lots in performance as well.
As an example of performance loss, the goal for Managed DirectX C#
samples is 98% the speed of the C++ samples. In my cursory
examination, some of the samples run faster in C# than they do in C++.
The loss is a long way short of 'lots'.
--
Barry Kelly
People often find it easier to be a result of the past than a cause
of the future.
 

Re:Re: Interesting Borland commentary on the future of Win32 programming

Barry Kelly < XXXX@XXXXX.COM >wrote:
Quote
"Simon D" < XXXX@XXXXX.COM >wrote:
>You lose lots in performance as well.

As an example of performance loss, the goal for Managed DirectX C#
samples is 98% the speed of the C++ samples. In my cursory
examination, some of the samples run faster in C# than they do in C++.
The loss is a long way short of 'lots'.
I should also point out that memory costs seem increase by about 60%
or so, however - that seems to be de rigueur for garbage collected
languages.
--
Barry Kelly
People often find it easier to be a result of the past than a cause
of the future.
 

Re:Re: Interesting Borland commentary on the future of Win32 programming

"Barry Kelly" < XXXX@XXXXX.COM >wrote in message
Quote
Barry Kelly < XXXX@XXXXX.COM >wrote:

>"Simon D" < XXXX@XXXXX.COM >wrote:
>>You lose lots in performance as well.
>
>As an example of performance loss, the goal for Managed DirectX C#
>samples is 98% the speed of the C++ samples. In my cursory
>examination, some of the samples run faster in C# than they do in C++.
>The loss is a long way short of 'lots'.

I should also point out that memory costs seem increase by about 60%
or so, however - that seems to be de rigueur for garbage collected
languages.

--
Barry Kelly
Thanks for the extra info Barry. One final query. What (if any) type of
development is .Net aimed at? Is it just about distributed/multi-tier apps?
I don't write such apps, so maybe that's why .Net is pretty meaningless to
me.
TIA
Simon.
 

Re:Re: Interesting Borland commentary on the future of Win32 programming

Distributed/multi-tier is the primary target. However .NET is useful
in many other areas. If they fix some issues, it has the potential of
handling even some real time operations.
Simon D wrote:
Quote
Thanks for the extra info Barry. One final query. What (if any) type of
development is .Net aimed at? Is it just about distributed/multi-tier apps?
I don't write such apps, so maybe that's why .Net is pretty meaningless to
me.