Re: Why .NET vs. Win32?
From: Brent S. (brentatatmosoftdotcom)
Date: Wed, 8 Oct 2003 10:51:15 -0600
"Bob Dawson" <email@example.com> wrote in message
> "Brent S." wrote
> > Yes, but what are the advantages to the vast majority of users. .NET
> > is being portrayed as a toolkit, and I can select from a lot of those
> > already. There is very little is being mentioned of the benefits to my
> > customers.
> Directly and immediately for my customers, none--just as I don't know or
> care what tool sets my auto mechanic uses.
> In the long run, however, .NET offers a more powerful, consistent, and
> flexible programming model than Win32 does--so it's sort of like your
> mechanic suddenly having power tools and diagnostic devices he never had
> before. Eventually, as those tools are mastered and come into play, the
> craftsman can work faster and produce better results: all other things
> equal, the craftman with the better tools wins.
All of this is very well and good, and you are helping me make my point.
.NET has more in common with Visual Studio than it does with an OS or
platform. Why lock ourselves in to the .NET platform by taking the bait of
compiler features and toolkits. This can all be done by more than one
vendor outside of .NET, without any of the penalties of .NET. JIT etc. has
its place, but not everywhere.
> > I have windows whether I use .NET or not. What feature is there in
> > .NET that is as revolutionary besides portability?
> Revolutionary? not much. .NET is more a synthesis/integration than a
> revolution, and not of UI features so much as of programming and software
> design concepts. But specifically: reflection, consistent error/exception
> handling, better diagnostics, a class library that rivals or betters Java,
> advanced language interoperability, assemblies in place of DLLs (like
> packages but even better). Just as the VCL is a major factor in Delphi's
> greatness, so the FCL is a major piece of .NET's value. The FCL dwarfs the
> > The various flavors of Windows NT are not a shell. .NET
> > always will be.
> Not according to MS. Anders Heilsberg said as long ago as last BorCon that
> the Windows API would be 'retired.' That is MS's official intention (How
> fast they can make it happen is _very_ open to question).
What Windows API has ever been retired, unless by retired you mean no longer
expanded. This is the exact same API they just converted to 64 bit! It
doesn't look like it is going to be dead for a long time. MS will tell you
one thing, but they won't introduce incompatabilities if it means upsetting
.NET was created out of Java jealousy. Microsoft was scared to death that
Java hosts would replace their operating systems. This was FUD, generated
and felt by MS. Java wasn't about to replace all of the natively compiled
applications out there. MS couldn't write their own version of the Java VM
and exclude all others so they decided to write a replacement which would
support more languages and have more functionality. They roped a few other
features into the mix, security not being the least of which, but I am still
waiting to see how this can actually work. Their intent was to create a
Windows OS that would transcend the PC and move out onto the net(.NET). I
think these decisions were made under the FUD Java created, and trumped
actual market research data. Yes .NET is neat and has its place, but MS is
trying to force all of its development market segments into the same bottle.
If they can do this then there are obvious benefits for Microsoft, but
should they do this? There are also drawbacks when you try to dictate to
your customers rather than catering to them. The end result is the only one
that matters, so MS should be careful.
(The following paragraph has nothing to do with the discussion at hand, but
I just had to say it.)
Despite all of the naysayers out there preaching about Borlands doom they
are far from it. Borland has an interest in .NET too. Do you think the
Borland Patent payoff was a one time event. Borland still holds many
patents used in .NET, and the success of .NET means the success of Borland.
This is funny because the same could be said of Java, and to a minor extent
Win32. Borland is catering to its customers much better than MS can.