Re: D8: first impression and uneasy feeling... (long)
From: Adrian Gallero (agallero_at_netscape.net)
Date: 17 Feb 2004 04:46:37 -0800
Stephane Grobety wrote:
> Would someone care to comment on that ?
Well, I think the big problem here is that we tend to put on the same
pack server apps and gui apps. Maybe because we are so used to delphi,
and having one tool for everything.
I think people that defends .net here is people doing mostly server
apps. Why? because .Net is great for web apps! GC is a must on a server
that has to run 24x7. Installing a .net framework is no problem on a
server. Memory and CPU are no problem either. And a bad winforms
implementation (not using hardware acceleration??) isn't either, you
don't use winforms.
Now, people who attacks .Net is probably doing GUI apps. Why? Because
.Net is not so good for GUI staff. GC is nice to have, but not a must
if the user is turning off the pc each day. Installing the .net
framework can be a pain if there are thousand of machines, or you want
to distribute shareware. You can't tell your user to "go and buy a
better machine", and " ah! and btw, get rid of that wintoy98!" And you
have to use the non so performant winforms.
All in all, I think it is an easy desision. If you are doing guis,
stick with win32 for a while. If you are doing servers/web apps, don't
wait a second, go to .net now.
Borland has an incredible potential here, but it has to realize there
are 2 markets here, and support both. So, if you want to do gui, you
can chose d/win32/64. If you want to do web, choose d/net. Everyting
with similar tools and a common framework. And you can do a 3 tier app
with a .NET server and a win32 client. Wouldn't that be great?
If borland really goes that way (and I hope they will, and wish they
would say it louder) I think it is the best reason to staying with them.
I just can't understand things like the first open letter (saying
something like we were 100% ready for .NET) or borland not saying
clearer what they are going to do. Yes, I know sometimes they say
something about commitment and a D7 patch here, but I mean something
much bigger. I would expect something like the borland motto beeing
something like "Supporting .NET and Win32" But well, I really don't
know and hope borland is right on what they do. They must have studied
this a lot more than me.
Nobody knows the future. Some people thinks .NET is going to be the
next dos->windows change. Everyone using win32/win64 API will be a
cobol-dino-programmer. Some others think it is going to be the dos->os2
change. Even when os2 was much batter technically than dos/win3, it
just dind't made it.
If you ask me, I think .NET evolution is going to be like in Java. It
will not be the "end-all revolution like dos-win" (as Java was supposed
to be when it started), and it won't be the "os2 complete failure
thing" some other people expects. It will be something in between, a
slow growth, that might end up completely replacing or not the windows
api. And with a fast adoption on servers, and not so fast on clients.
(does this remembers me java?.. why would it be ;-) )
After 4 years, I think it is perfiling this way. But borland might give
us the best of the worlds, whatever it is. Is .NET a complete success
and it replaces completely win32? Ok, we have d.net. Is a complete
failure and replaced by microsoft .pet? Well, you still have delphi for
win32 :-) Is in not a complete failure/success and it coexists with
win32/64? Well, you have a delphi for each of those, and can use them
depending on your needs.