Re: How about the future of VC++/MFC

And Sun, with its multiple incompatible releases of Java, makes sure we can't bear to be
locked into their proprietary tools.

And all those versions of Unix, with names that ended in -ix but which couldn't call
themselves "Unix" because of the AT&T licensing, and with their various different and
incompatible C compilers, shells, and utilities, made sure we couldn't move from Apollo to
Sun to HP, but that was OK because it was Unix, which Can Do No Wrong.

I guess none of you ever worked on IBM mainframes, where the lockin was fairly strong. Or
used RCA, Burroughs, NCR, or any other vendor.

There are more Windows boxes out there now than the sum of ALL computers made by ALL
previous vendors since the first computer was built. This is "lock-in"? I can buy a
computer computer which is orders of magnitude faster than a mainframe for less than I
used to pay for a single disk pack (which held 40MB. The drive the pack was mounted on
was about $35,000. We had eight drives. I signed the purchase order). So if I'm "locked
in" to a product that costs me less per year than the old mainframes cost in electricity
per month (our DEC-20, about a 286-power machine, cost us $1900/month just in electricity!
In 1981 dollars...).

My entire capital outlay for 12 machines (three of them laptops), a superfast B&W laser
printer, a superfast color laser printer, two scanners, three UPS units, networking
support, wireless LAN, a terabyte of online memory, tape backup units, and three massive
flat-screen displays would not have bought a single 40MB disk drive in 1980. What do I
care about lockin? Yet there are people who seem to think this matters. It doesn't, in
nearly all cases. There are exceptions, but we are talking about a few tens of thousands
of cases compared to about a billion installed computers. People treat this as if it
matters, but it rarely matters. In fact, WIndows has done more for portability than any
operating system in history (I recall one organization I worked for that had a "portable"
product, with so many #ifdefs in it that the code was nearly unreadable, and most of them
to get around fatal compiler bugs in the 30 different Unix platforms we were supporting)
By comparison, WIndows is a remarkably uniform concept.

Everyone talks about "proprietary lock-in" as if it matters any longer. It doesn't matter
in the slightest. This reflects an attitude that made sense only when hardware and
software cost money. The difference between the cost of Windows and the cost of Linux is
nonexistent; effectively, both are free, but Windows utlimately costs less. (I don't use
free open-source software because I can't afford it).

Let's see: its OK to be locked into Linux, but not OK to be locked into Windows? What
have I missed here? The hardware is nearly free, the software is nearly free... (and
don't talk to me about purchase price; I'm talking about TCO), so what does it matter? If
you want to work in an environment that represents the best thinking of 1965, by all
means, feel free to use Linux. I lived in Unix for 15 years and nothing made me happier
than to never have to see it again.

Let's see: it's OK to be locked into Java but not into MFC? Don't tell me about
portability; most portability of Java has little, if anything, to do with the choice of
platform.

Oh, yes, and we know that open architectures are moral and right and closed architectures
are evil. And Microsoft is evil and Apple is good. Have you ever tried to create a
third-party backplane device for a Macintosh? Which bus did you choose? And what was its
lifetime? I spent a couple years programming Macintoshes, and one of the things I
learned was that everyone who said how great they were had never programmed one. Those of
us who were trying to program them recognized they were years behind Microsoft's MS-DOS
environment in terms of developer support. And you couldn't build portable apps on the
Mac.

A friend who used to do Java applets said they could only program in a very limited subset
of Java 1.0 because otherwise the applets wouldn't run on the client machines. Yet I can
compile a well-written 16-bit MFC program and it runs immediately in Win32 (providing the
programmer didn't do anything brain-dead like store two handles in a DWORD), but I can't
compile a correctly-written Java program from the previous release in the current release
because there are so many incompatible changes. But Sun is Good and Microsoft is Evil. I
guess I have a lot of trouble understanding this.

Oh, yes, it's OK for Apple to steal from Xerox, but evil for Microsoft to steal from
Xerox, and it is *definitely* OK for X-Windows to steal from Xerox and KDE and Gnome to
steal from Apple and Microsoft. I find this all a bit confusing...

There are some topics that I react badly to, and the issue of "lock-in" is more than silly
these days.

I guess I've seen so many of these things come and go that I just cannot get excited about
the issue any longer.
joe



On Tue, 28 Nov 2006 15:41:47 -0600, BobF <rNfOrSePeAzMe@xxxxxxxxxxx> wrote:

On Tue, 28 Nov 2006 12:44:32 -0800, Tom Serface wrote:

I think the C# group ...

<$.02US>

I think MSFT is trying to save their azz from the OS tools. By creating a
new language, and migrating core functionality to require it (XAML), they
keep us locked in buying their proprietary tools.

Oh, and compete with Java ... :-)

</$.02US>
Joseph M. Newcomer [MVP]
email: newcomer@xxxxxxxxxxxx
Web: http://www.flounder.com
MVP Tips: http://www.flounder.com/mvp_tips.htm
.