On Sep 23, 9:28 am, Hadron <hadronqu...@xxxxxxxxx> wrote:
> Matt <m...@xxxxxxxxxxxxxxxxxxxxxxx> writes:
> > Roy Schestowitz wrote:
> >> Two Views of Enterprise Open Source
> >> ,----[ Quote ]
> >>http://www.computerworlduk.com/toolbox/open-source/blogs/index.cfm?bl...
>
> > I don't know if I can be very broken up about the demise of Borland.
> > I think they really played the same game as MS in regard to
> > programming languages.
Actually, that was one of Microsoft's biggest concerns about Borland.
Borland not only made sure that they were source code compatible with
BSD UNIX (Linux wasn't out until later), but they even provided source
to any modules that were "unique" to Borland. This was especially
true with C, C++, and Java.
This was probably one of the reasons that Microsoft targeted Borland,
because Microsoft wanted total control of the desktop. Borland
offering developers the ability to take UNIX code and compile it on
Turbo C or Turbo C++, left the door wide open for MS-DOS and Windows
applications that ran as well, or better, on UNIX.
Microsoft really couldn't have that, so they decided to challenge
Borland directly with their own compiler (actually a compiler the got
by acquiring another company), and they created a whole new set of
APIs designed to be completely incompatible with UNIX.
When that didn't work, Microsoft started making changes to Windows,
designed to break applications created using Borland's compiler.
Eventually, the developers just had no choice but to "give up" on
Borland, because they couldn't afford the lost of millions of
customers every time Microsoft fired another torpedo.
For those who don't know, a torpedo is a deliberate change in low
level code designed to sabotage a product that calls that low level
code. Microsoft has used torpedoes against Stacker, Borland,
Netscape, Word Perfect, DR-DOS, Lotus 1-2-3, Lotus Notes, OS/2, and
numerous other competitors to Microsoft products or shovel-ware. I
wouldn't be surprised to hear that Microsoft was using torpedoes
against McAffee, Nortan, and Symantic Anti-virus and Anti-Spy-ware to
drive them out of the Vista market.
Microsoft has often been caught and judges often rule that these
torpedoes are a form of sabotage of the competitor's products, at
which point, Microsoft usually offers a carefully worded settlement
that involves a small amount of cash (relative to the $billions at
stake if Microsoft loses it's monopoly power and monopoly extending
power), and gives Microsoft permission to continue the practice, but
the settlement is so well crafted by Microsoft's brilliant lawyers,
that the plaintiff usually doesn't know that they've just signed a
settlement that will drive them into bankruptcy.
A bit like what the DOJ did in the Antitrust Lawsuit.
Eventually, Microsoft's settlement and contract tactics became a model
for other companies, including banks, mortgage companies, and other
important businesses - along with some of Microsoft's other bad
habits.
> > There was a similar attack on standards
> > through non-standard language features and extensions, the desired
> > effect being to lock developers in to their compilers. The guy who
>
> What standards? And you think anything has changed? The great majority
> of code out there is not "standard". You wish death on gcc too?
Actually, Borland made it their goal to be compliant with standards of
BSD Unix. They also used an API that was compatible with the Athena
Widget set.
Later they adopted the TK toolkit, which hid the Windows vs X11
interface. The library code was proprietary, but it was API
compatible.
> > The guy who
> > built Turbo Pascal for Borland is now at MS, being the designer of C#.
I hear that Microsoft is trying to hire away all of the key adobe
developers as well. If Microsoft can't get the company's product for
a cheap price, and they can't get the company for a cheap price, then
they can hire away the top talent and have them form a company, who
can sell it to Microsoft for a cheap price. IIRC, there was a thread
in COLA a few years ago about the Microsoft interview questions, and
several of the questions were related to integrity. If you had too
much integrity, you didn't get hired.
> > Borland was very good about pricing though, at least in the mid 80's,
> > selling Turbo for $50 while MS compilers cost hundreds.
Microsoft's C++ compiler, even Visual C++, were very cheap until they
had crippled Borland. Then they were able to jump the price from $49
to $199, then to $299, and so on. Most serious Microsoft developers
get the MSDN subscription which goes for what? About 1500/year? And
MSDN developers also have to pay Microsoft royalties for run-time
rights once the product gets to market.
> Because the built a cheap quality innovative product. And yet you are
> happy they failed? Borland C++ 2.0 was a wonderful product.
Very true. I loved Turbo C++. I also liked Turbo Pascal, Turbo
Prolog was really interesting (write a complex database in 2 lines of
code). I also liked JBuilder. I was using CORBA when it came out,
and the Borland toolkit made CORBA development very easy. Later, they
offered J2EE with CORBA compatibility, which meant you could call C or
C++ services from Java clients, and vice versa. And again, the price
for all that functionality was very reasonable.
Borland actually did get some traction in the Linux community with
it's Kylix product, which again let you write one source code package,
and compile versions for both Linux and Windows.
> > Borland came late to the game and so didn't have a chance to compete
> > in the OS domain.
Very true. Borland's products did work very well with DR-DOS,
as well as MS-DOS and OS/2. But CP/M was pretty much extinct by the
time
Borland entered the market.
|
|
