Re: Opportunity passed by

Quite right! That's the scenario that would have fit perfectly into my
plan
for world domination if I were trying to grow that business and managing
the
project:

a primarily server-oriented system

Agree! First pass would have included basic ui vcl to target standalone
desktop, with intentions for version 2+ to target the browser instead. An
enabler technology for both developers and businesses who wanted to use
Linux systems. Market cultivation. Inception phase would have included
platform agnostic considerations.

That means Kylix would be competing with Java, and I doubt they would have
gotten at least as far as they have had they tried such a thing. They would
have still ended up abandoning it. Basic UI VCL (like Basic Swing
development
in many of today's Java IDEs, but worse maybe, especially since it's a
proprietary
framework [guaranteeing people would cuss and throw tomatoes at it]), and
in order to target the browser Borland would probably have had to reinvent
the wheel and develop a web-based framework a'la JSF to allow easy
development. Unfortunately, third-party vendors would not be porting their
components to this because it would be asking to lose money. Borland's
track reckord with improving the VCL in the past few years would lead me
to believe that Kylix users would still be using the same 15-20 web controls
that they were using since K2 was released had they gone this route.
Nevermind,
it would be dead just like the other one, anyways.

Word spreads quick when your name is "Borland Software Corporation" and
you produce an IDE as buggy and maginally useful as Kylix. They should have
coded the thing in Qt/wxWidgets from scratch and just had another IDE look-
and-feel, cause the WINE solution was less than optional. It seemed like a
quick
hack. There are open source software products that offer better robustness
with
just as many features as Kylix did. It was horrible. I still haven't
upgraded my
Linux distro because Kylix won't run on the next version. That isn't the
case with
the other 150+ Linux development tools.

JBuilder was successful because it got in fast (when Java IDEs were still
not free
and terribly expensive and didn't offer that much bang for your buck, but it
did).
They had a nice IDE with a nice feature set and were quick to support the
new
ones. I think JBuilder was the first IDE to support Java 5.0 (Generics).
Why
can't they support .NET 2.0 as fast as that? Lol. It's not like Sun wasn't
the one
advancing Java (along with the JSRs, etc.). Still, however, they were
always on
top of things, and that's what made JBuilder better than the rest of the
IDEs, which
came mostly from Java AppServer/VM producers (IBM, Sun, Oracle) and were
geared to work best with their specific stack. JBuilder was the swiss army
knife
with the one-size-fits-all screwdriver attachment. I think JBuilder also
had the best
refactoring, outside of maybe IntelliJ IDEA. It was certainly the best for
SWING
development, and Borland developed some pretty good JavaBeans to package
with it.

Eclipse pretty much killed it not because it was free, but because of the
large
community of users behind it that provided this huge flood of plug-ins,
which can-
not be used form other IDEs. Also, more and more technologies are popping
out
from everywhere that JBuilder didn't have access to. You have numerous open
source databases (Derby, etc.), Ajax, Ruby on Railes. Eclipse can even be
used
for PHP development, among the 100+ other programming languages it has
access
to.

Eclipse is today what JBuilder was when it was released.

while the many freely available tools (a) didn't have those limitations,
and
(b) were also capable of being used to write code for commercial UNIX
variants
such as Solaris and AIX with minimal (and in many cases no) alterations.

But arguably not productive enough to make the market viable. We'd have
the
productivity covered, then we move onto improving platform agnostics and
commercial quality config/management tools for the market.

Yes, but on what grounds do you think a tool Kylix would compete with Java
as far as Enterprise solutions were concerned.

I know it's too easy and (not quite fair) to armchair all that happened
with
Kylix, but the bigger point is that Borland simply can't chase the flavor
of
the day that puts attention on a particular market, then abandon that
market
because the flavor of the day wasn't a big seller for it. This manifested
itself in the Kylix product by identifying seemingly simplistic or shallow
market needs (and it was also the most comfortable move for Borland to
make,
because Kylix was conceptually just a port of the existing windows
product).
Now, Borland/CodeGear are in a similar pickle with Delphi -- extending
their
known concept and technology to continue chasing the windows market, in
spite of the writing on the wall.

Well, they went the lazy way in developing Kylix, although with their lack
of
resources compared to other companies, I can understand this. But, the
whole
concept of Kylix was killed when it was delivered in a very poorly-hacked,
unfinished, buggy, put-together way. Because of Kylix, I now know to stay
away from applications which depend on Wine and old Linux distros. I will
stick to something "native" for now on.

They also removed the ability to target Apache in later versions of Delphi
in
the IDE.

To paraphrase the saying, all Borland/CodeGear have in Delphi is a windows
IDE, so any market looks only like some version of a windows IDE market.
The customers are pretty much the same way. Borland/CodeGear has got to
look
at every product move now like Kahn did in the beginning -- take a gamble
with a product, convinced that it is truly innovative, and with little
customer support -- and not continue business as usual. Hopefully Delphi
(or
some descendant) has a place in there somewhere.

I believe Delphi has a place in there, of course, but it's shrinking and as
long as
they continue to deliver a product targetting outdated/deprecated
technologies
it will continue to shrink until they can't fit in there anymore. .NET 1.1?
MySQL
4.1? No PostgreSQL drivers? Probably using an old version of DB2 drivers
also, I haven't checked tho. Delphi is supposed to be the best database
application development tool for Win32 developers, but cannot deliver
sufficient
drivers for current database management systems. For .NET it's behind, and
as
a result loosing third-party support because people can't afford to continue
supporting deprecated packages because a vendor or two has fallen behind
the times.

They also have no 64-Bit compiler, and their C++ toolchain is outrageously
outdated compared to the likes of Visual C++, GCC, and even smaller
offerings
such as Digital Mars. The Open Watcom compiler will soon eclipse
C++Builder,
and that's with retaining it's cross-platform capabilities; and they're
currently
working on Linux support... I think it generates code for the current (or
more
current) processors also, and it's code is better (bigger, but faster).
[This is
important because Delphi uses the Borland C++ back-end IIRC, so you'll
probably
not see a 64-Bit Delphi compiler when they can work on their C++ compiler
and upgrade it to that status].

No support for Managed C++ or C++/CLI for C++ toolchain either.

It almost seems unreal that "Borland" could allow their development tools to
fall
behind, but alas...

- Nate.


.

Relevant Pages

  • Re: A .NET Poll
    ... future versions of .NET will establish a niche market. ... cross-compiling of Kylix /Linux developed application to Windows. ... incomprehensible to me why Borland created such condition for Kylix ... but a developer should not be forced to ...
    (borland.public.delphi.non-technical)
  • Re: Delphi in the enterprise
    ... The more Borland sells their CoreSDP for .NET stuff, the better it is for MS ... teamsystem will be more profitable if the ALM ... need BDS in next 2-3 years to push into that market. ... pushing BDS as the premier IDE in their .NET CoreSDP offering is ...
    (borland.public.delphi.non-technical)
  • Re: A Modest Proposal *
    ... but wasn't adopted even by Borland customers when it was included in the ... but Kylix wasn't marketed or focused on a server-side development ... That's why you don't market to GIMP users, ... Borland/DevCo is very free to innovate. ...
    (borland.public.delphi.non-technical)
  • Re: Embrace and Extend, Borland Style
    ... Although I see your points on development of the IDE, ... Kylix 4, Managed code. ... Isn't this a patent Borland owns? ... Another disadvantage is that if Microsoft sees Borland making some money ...
    (borland.public.delphi.non-technical)
  • Re: to Delphi or not to Delphi?
    ... >> Thanks for the good laugh! ... As if the KCP was supposed to fix the IDE, ... There once even was support from SuSE staff so I could get my Kylix 3 ... I wish Borland did too. ...
    (borland.public.delphi.non-technical)