I know the "purists" would complain, but I would be fine with a Windows IDE
and the ability to produce executible files for Linux, FreeBSD, and Mac
using real VCL with no QT. Seems to me that those three systems have become
very similiar (all a form of Unix) and XWindows is involved. Just give me
a good basic de{*word*81} for each... i.e. turbode{*word*81}. For the price of PC's
today, it's not hard to get a cheap spare system or find an old used system
to run Linux. Since Mac is going to be x86 based soon, just focus on
compiling on the new x86 systems. Do I think it's possible to make money on
that... for sure. Java and .NET are too slow.... I don't care about the
arguements people have. There is no duplication for true control of a
system and true pointers and freeing up memory. Garbage collection sounds
great, but it's undeterminitic. There is a huge embedded systems world
market out there and they run on x86. Go for the basics, and have a good
closed beta test group.
To prove my point, we are working on GUI that works on Fedora using Kylix
C++ Pro right now and it runs on a VIA 400 MHZ x86 with 64 MB RAM and it
works very fast and it has 8 screens doing very interactive graphics that
are user selectable for runtime or configuration. Kylix has been a huge
time saver for developing the GUI. We don't use it for everything.... the
drivers we write are in straight C and use GCC but for the GUI, Kylix has
worked out great.
-Tom
"siegfriedn" <sniedinger@yahoodotcodotuk>wrote in message
Quote
This is my reply post from the delphi non technical group, but have not
had a reply from JK there, so I am posting it here where it is more
appropriate..
John Kaster (Borland) wrote:
>
>We wanted to make money from Kylix. You are welcome to suggest a
>"different business model" that will make money from Kylix.
>
Thank you for the invitation :)
From the positive hype in the early days of Kylix that was clear and the
future looked bright...you came sooo close.. Kylix is still usable today
even with it's based on a very old linux distributions, so the argumnts
that one cannot have a binary only solution on Linux is not very accurate.
I am not really the person to give you advice on a suitable business
model, but these are just some ideas..
If you have limited resources try to open source (perhaps with a dual
license like Trolltech - commercial license for closed commercial
application and free for GPL applications)
- Make all the CLX source available - (it could be abstracted to support
other GUI toolkits like win32, GTK+, wxWidgets, etc to achieve a native
look) - let the community maintain and develop it under Borland
management. (Probably what you are trying to do with the FreeCLX community
project?)
- Create a 'classic' minimalistic IDE as suggested by someone else. Open
source this IDE and release under a suitable license agreement. The core
architechture for this IDE has to be based on pluaggability - according to
Borland standards. (Almost like Eclipse)
How Borland makes money..
-------------------------
- Sell component packs for example the additional components which are in
the Enterprise edition only.
- Sell IDE plugins for example the SDO/ALM intergration stuff, ECO etc..
- Sell an official shrink rapped 'stable' version of Kylix to corporate
customers with all the Borland goodness and support prepackaged. (Like Sun
with Open Office and StarOffice)
Hope you find it useful :)
siegs