Board index » delphi » Re: .NET myths debunked.

Re: .NET myths debunked.


2004-01-05 07:25:52 AM
delphi249
marc hoffman writes:
Quote
try running a (non-trivial) linux application build under any given
distribution under any other/newer distribution. alternatively, check the
kylix groups for the issues people are having with running
Kylix/Kylix-made-apps under recent linux distributions. this gives a whole
new meaning to the phrase "DLL hell".
That is a consequence of how Borland chose to implement Kylix with Qt and WineLib.
They could have built on top of lower level APIs though it would have taken more work
on their part.
Frm a Charlie Calvert Dec 23, 2003 post in the "Borland Layoffs?" thread:
There is some truth to what you say. However, Borland specifically
targetted particular versions of WineLib and QT, two libraries that
everyone knew were going to change. In the Windows world, this is the
equivalent of tying your product to a specific version of MS Office 1997, or building
a Delphi tool that specifically targets a particular version of Delphi, such as
Delphi 5. If you do such a thing, then of course you are going to have problems when
the platforms you targeted are upgraded.
There is nothing irrevocably wrong with building on top of QT or
WineLib. I am not implying that was necessarily a mistake. But if you do
this, then you have to accept that it is going to cost money to maintain
your code. That maintenance price is not inherent in Linux, it is
inherent in QT and WineLib. (Not that the calls in these libraries
change all that much. In fact, they don't. Particularly in the case of
QT. But Borland linked tightly to specific library releases of these
products.)
 
 

Re: .NET myths debunked.

Randall Parker writes:
Quote
Are there any warning dialogs popping up when you try to run DOS apps
from the 1980s? There are probably some DOS apps that are
incompatible. But many still run.

How about warning dialogs for Win16 apps?

Win32 is going to last for decades. As for its days being numbered:
Many of us will probably die first.
I think you're confused. Win16 was a technological improvement over
DOS, as Win32 was an improvement over Win16. Win16 provided access to
more memory and a standardized GUI interface (whereas with DOS you were
forced to roll your own, use EMS/XMS or a DPMI interface). Win32
expanded upon that by providing for a flat memory model that forever
rid us of dealing with segments, as well as forcing code to 32-bit
(e.g. - use the extended CPU registers rather than the old 16-bit
registers).
Because of that, there was no reason to put a "warning" dialog up
because there was nothing fundamentally unsafe about the older code.
DOS apps existing in virtual 8086 sandboxes, and Win16 apps existed in
yet another sandbox, but they were just as speedy as their Win32
counterpart, they just lacked easy access to more memory and a simpler
memory model.
.NET, OTOH, is being propped up as being safer (no more pointers, a
garbage collector to minimize the chances of memory leaks, and the
ability for admins to limit what kind of code can be ran). .NET brings
nothing new technology wise to the table other than a half-hearted
attempt to make something that can be cross-platform (and I say
half-hearted because .NET code is likely only going to be
cross-platform when it is ran on systems that implement WinForms and the
other Windows specific portions of the framework).
As far as Win32 being pushed to the back burner, one need only visit
Microsoft's Longhorn SDK online to note that there's either a) no
coverage of new Longhorn Win32 API's or b) there simply are no new
Longhorn Win32 API's--
longhorn.msdn.microsoft.com/
Everything documented there seems to indicate that it is only accessible
via .NET (or via Win32 somehow connecting with .NET namespaces, which
one imagines would be slower than simply doing the same from a .NET
app).
Anyways, it is delusional to think Microsoft won't play off safety and
security as a prime reason to avoid non-.NET apps. They already (on
Windows XP and Windows Server 2003) provide warning dialogs for
unsigned driver installations, I imagine popping up warning dialogs for
non-.NET apps would just be a logical extension of that (and of course,
anything signed or shipped by MS as part of Longhorn or Office would
likely be listed in any exception list to the rule).
Let me repeat something I have said elsewhere so there's no mistaking my
beliefs-- I do NOT believe that Win32 code is unsafe, I DO believe that
Microsoft will market Longhorn that way however, and I DO believe
consumers are just gullible enough to believe it.
Will
 

Re: .NET myths debunked.

Are there any warning dialogs popping up when you try to run DOS apps from the 1980s?
There are probably some DOS apps that are incompatible. But many still run.
How about warning dialogs for Win16 apps?
Win32 is going to last for decades. As for its days being numbered: Many of us will
probably die first.
Kevin writes:
Quote
How about after Longhorn?

I can not help wondering how long it will be before a Win32 app starts up
on a Microsoft O/S with a warning dialog. Since Microsoft has received
quite a bit of flack about security... and they've put all this effort
into creating a platform (.NET) where you can set policies as to what a
user can and cannot run, isn't it just a matter of time before they try
to stop us running Win32 apps altogether? This may not yet be
implemented in Longhorn, but maybe the Longhorn+1 release?? This will
work in a similar way to the way the warning dialog in Outlook works.

Fact is, Win32's days are numbered.
 

Re: .NET myths debunked.

Marco Caspers writes:
Quote
Kevin writes:

[snip]

>Fact is, Win32's days are numbered.

<<<<<<<

But not because of Microsoft, but because of AMD bringing 64 bit to the
home desktop at sub US$1500 pricing tags...

AMD's 64 bit processors do not break 32 bit code. So Win32 will live on and on.
 

Re: .NET myths debunked.

Nick Hodges (TeamB) writes:
Quote
Dave Jewell writes:


>>Fact is, Win32's days are numbered.


And indeed they are. ;-)
The darned universe is eventually going to become so used up and spread out (or is it
going to implode?) that there will not be much if any life forms to run Win32 apps.
So, yes, Win32's days are numbered. Heavy sigh.
 

Re: .NET myths debunked.

"Pete Goodwin" <XXXX@XXXXX.COM>writes
Quote
>
>If you feel MS is offering you something you want/need in .NET that you
>don't have in Win32/64 then you now have multiple paths to get to
.NET...
>
>But you still are not forced to go (by MS).

Huh? If there's something I need or want in .NET that isn't in WIN32/64,
then how do I get it without .NET? My whole point here is that you're
saying you need .NET after all.

I think there is a fundamental difference between what *you* need as a
developer and what .NET offers in terms of new technology. The simple fact
that you can deploy a delphi 7 app on longhorn, as is, should at least bring
into focus your needs versus your desires.
There is also the aspect of catering to the demands/needs of *customers*,
which others in this thread have pointed out as a driving force for the
adoption of .NET... One way or another the marketing prowess of MS will
make itself felt.
Quote
>I am curious. Can you name your top three features in .NET that is
>compelling you to develop in .NET over Win32 now?

The new GUI in Longhorn? I am guessing here...

Ok. Perhaps you percieve this as something you need now or will need... I
have to admit XAML appears to be a major improvement over anything we
currently have. It seems to me marketing *better* technology, and
marketing it well, is always a sure-fire strategy to success...
<g>
Developers are going to need concrete reasons to switch from tried and true
technology that works..... Marketing is crucial, but in this case, it is
by no means the deciding factor. Speaking for myself I will deploy
natively compiled applications whenever my analysis tells me its a superior
alternative... And I will develop strategies to Inter-Op with natively
compiled applicatons whenever that appears to be the superior option.
Native Win64 compilers will be *very* relevant to the future of software
development IMO.
-d
 

Re: .NET myths debunked.

"Will DeWitt Jr." <XXXX@XXXXX.COM>writes
Quote
Anyways, it is delusional to think Microsoft won't play off safety and
security as a prime reason to avoid non-.NET apps. They already (on
Windows XP and Windows Server 2003) provide warning dialogs for
unsigned driver installations, I imagine popping up warning dialogs for
non-.NET apps would just be a logical extension of that
I think there's a lot in what you say, but of course there's a big
difference in frequency between the occasional unsigned driver installation
on the one hand and repeated, day-in, day-out usage of numerous Win32 apps
on the other. I agree that there will be some sort of security warning
because Microsoft will want to inject some FUD and persuade consumers away
from their existing legacy apps. However, I can not believe that a warning
dialog will come up every time you run a Win32 app. This would be
profoundly irritating to most people, and I suspect that it would be *so*
irritating that most folks would simply not want to use Longhorn on that
basis. To my mind, a more likely scenario is that you will be able to
configure local security policies for Win32 apps, and that will include
whether or not a warning dialog appears. I imagine that most folks will
simply switch this off, just like most folks routinely switch off Remote
Desktop, Remote Assistance and that kind of stuff.
Quote
(and of course, anything signed or shipped by MS as part of Longhorn
or Office would likely be listed in any exception list to the rule).
If Microsoft provide a 'per-application' exception list facility , then one
would hope that it is available to administrator accounts so that
(non-Microsoft!) legacy apps can be added to the list.
Personally, I am convinced that the Win32 API will be around for a *long*
time, and it definitely wouldn't be in Microsoft's interests to make the
running of Win32 apps *too* irksome.
Dave
 

Re: .NET myths debunked.

Will DeWitt Jr. writes:
Quote
Anyways, it is delusional to think Microsoft won't play off safety and
security as a prime reason to avoid non-.NET apps. They already (on
Windows XP and Windows Server 2003) provide warning dialogs for
unsigned driver installations, I imagine popping up warning dialogs for
non-.NET apps would just be a logical extension of that
I was thinking along these lines as well. The "unsigned driver" dialog
is the exact type of thing they're likely to implement for Win32 apps in
Windows versions after Longhorn. The reason I say "after Longhorn" is
that they don't really want to dissuade people from using Longhorn
(being the first Microsoft O/S to include .NET). They probably will
have policies in Longhorn that can be turned on by network admins that
will display these warnings.
If Microsoft implements this warning dialog I'd think that they
would have a checkbox on the dialog to say "do not display this message
next time for this application" or something like that. That would
remove the annoyance factor while still informing users of the "safety
issue" associated with running non .NET apps.
Cheers,
Kevin.
 

Re: .NET myths debunked.

Captain Jake writes:
Quote
By the way, I saw a unicode-compliant replacement for TStringList on the CD's
for either Delphi 7 or D8.
OK. Thanks for the tip, only problem is that I have already implemented
my own TStringList. :-)
 

Re: .NET myths debunked.

Dennis Landi writes:
Quote
"Pete Goodwin" <XXXX@XXXXX.COM>writes
news:3ff81ecf$XXXX@XXXXX.COM...

>If new API's aren't created for WIN32/64 then I would say Microsoft are
>locking developers into .NET for the future - if I can not access any of
>the features of Longhorn except by .NET, then WIN32 is dead for me.
>
>Pete Goodwin

If you feel MS is offering you something you want/need in .NET that you
don't have in Win32/64 then you now have multiple paths to get to .NET...

But you still are not forced to go (by MS).

I am curious. Can you name your top three features in .NET that is
compelling you to develop in .NET over Win32 now?
Taken from:
www.fawcette.com/vsm/2004_02/magazine/features/rjennings/
"If you aren't using managed code for all new Web and Windows apps, it's
time to refactor your development strategy. Win32 will go into
maintenance mode in the Longhorn timeframe."
It looks as though the major push from .NET is for WIN32 to stop dead in
its tracks and .NET to take over. That is, any new features will be via
.NET and not WIN32.
Pete
 

Re: .NET myths debunked.

"Pete Goodwin" <XXXX@XXXXX.COM>writes
Quote
>I am curious. Can you name your top three features in .NET that is
>compelling you to develop in .NET over Win32 now?

Taken from:

www.fawcette.com/vsm/2004_02/magazine/features/rjennings/

"If you aren't using managed code for all new Web and Windows apps, it's
time to refactor your development strategy. Win32 will go into
maintenance mode in the Longhorn timeframe."

I didn't ask you to quote the trade press. You didn't answer my question.
I asked *you* to name any features you needed in .NET that you don't have
now. You didn't.
Quote
It looks as though the major push from .NET is for WIN32 to stop dead in
its tracks and .NET to take over.
"Stopping dead in its track", while a colorful phrase, lacks the precision
necessary to actually respond to it. Win32/Win64 will continue to exist and
function on desktops and the enterprise for at least the next 5 years and
possibly much longer. So it certainly isn't stopping dead in that sense.
Plus, there is a "term of art" in the .NET taxonomy known as "unmanaged"
code. On windows, its Win32/64 natively compiled code, and it is alive and
well in the .NET milieu.
Use .NET because it has something you want or need. You are not being
forced.
Quote
That is, any new features will be via
.NET and not WIN32.
Now you are quoting me....
Time to move on.
 

Re: .NET myths debunked.

Dennis Landi writes:
Quote
"Pete Goodwin" <XXXX@XXXXX.COM>writes
news:XXXX@XXXXX.COM...


>>I am curious. Can you name your top three features in .NET that is
>>compelling you to develop in .NET over Win32 now?
>
>Taken from:
>
>www.fawcette.com/vsm/2004_02/magazine/features/rjennings/
>
>"If you aren't using managed code for all new Web and Windows apps, it's
>time to refactor your development strategy. Win32 will go into
>maintenance mode in the Longhorn timeframe."

I didn't ask you to quote the trade press. You didn't answer my question.
I asked *you* to name any features you needed in .NET that you don't have
now. You didn't.
At the moment none. But when a new feature is added to .NET that appears
useful, I am unable to use it with WIN32 but must use .NET instead.
Quote
>It looks as though the major push from .NET is for WIN32 to stop dead in
>its tracks and .NET to take over.

"Stopping dead in its track", while a colorful phrase, lacks the precision
necessary to actually respond to it. Win32/Win64 will continue to exist and
function on desktops and the enterprise for at least the next 5 years and
possibly much longer. So it certainly isn't stopping dead in that sense.
Plus, there is a "term of art" in the .NET taxonomy known as "unmanaged"
code. On windows, its Win32/64 natively compiled code, and it is alive and
well in the .NET milieu.
However, no new development will go into WIN32/64 - all the new stuff
will go into .NET. that is why I described it as "dead in its tracks".
Quote
Use .NET because it has something you want or need. You are not being
forced.
It looks to me as though if I want to stay current, I *am* being forced
to use .NET.
Quote
>That is, any new features will be via
>.NET and not WIN32.

Now you are quoting me....

Time to move on.
Pete
 

Re: .NET myths debunked.

Dennis Landi writes:
Quote
I am curious. Can you name your top three features in .NET that is
compelling you to develop in .NET over Win32 now?
I can not name three, I have barely scratched the surface of .NET (well,
rather, I have barely scratched the surface of exploring WinForms and the
FCL), but I will name one interesting thing-- the SysLink control
available with .NET in any OS that can install .NET. The SysLink
control is minor, of course, and you can simulate it in Delphi with a
little effort, but it is an example of a UI element that is only
available via .NET on all OS's. (FWIW, the SysLink control IS
available in Windows XP and greater as a common control, but there is
no redistributable component to get the control to function on earlier
OS's-- you're locked in to XP or greater if you access it via Win32).
I think situations such as this will become much more commonplace with
Longhorn and future Microsoft OS's.
Will