Board index » kylix » the end of the Kylix experiment ?

the end of the Kylix experiment ?


2004-02-26 01:34:43 PM
kylix2
g'day everyone,
check out:
www.realsoftware.com/realbasic/
a Visual Basic clone.
note that while the IDE is Win32, it can produce Linux (and Mac) binaries using
GTK 2.0:
www.realsoftware.com/realbasic/rb55screens/l_win_buildsettings.jpg
don't know how good it is but they do have a downloadable demo.
cheers,
Mat
 
 

Re:the end of the Kylix experiment ?

Mat Ballard wrote:
Quote
g'day everyone,

check out:

www.realsoftware.com/realbasic/

a Visual Basic clone.

Gambas is also a good Basic IDE, and you can use it from Linux to write
Linux code. It is free as well.
gambas.sourceforge.net/
 

Re:the end of the Kylix experiment ?

Quote

www.realsoftware.com/realbasic/

a Visual Basic clone.


cheers,



Mat
What has this to do with your subject line??
The IDE does not even run on Linux...
 

{smallsort}

Re:the end of the Kylix experiment ?

Granted, it does come across as a thinly disguised adverti{*word*224}t, I
found it informative that realbasic is finally available for Linux.
However regarding the oddity that the IDE isn't available, I've always
found the mark of an awesome language such as Delphi to be that they
wrote the language and IDE using the very language the IDE is for. What
a wonderful way to test the blue blazes out of your compiler /
Interpreter.
Guess one can draw some conclusions about the portabilty of Real Basic
code between operating systems if the IDE cannot be readily ported.
theo < XXXX@XXXXX.COM >wrote in news: XXXX@XXXXX.COM :
Quote
What has this to do with your subject line??
The IDE does not even run on Linux...

 

Re:the end of the Kylix experiment ?

theo wrote:
Quote

>
>www.realsoftware.com/realbasic/
What has this to do with your subject line??
The IDE does not even run on Linux...
please correct me if my memory has become faulty with old age, but:
1. Kylix promised true cross-platform capability between Linux and Windows;
2. with true RAD capabilities, and,
3. was initally advertised as "VB for Linux".
it seems to have failed for a variety of reasons, including:
a. insufficient sales to cover development costs
b. the "7M tail" of any install, due to an inability to bind directly to either
of the major GUI frameworks on Linux.
c. pissing off the customer base with an over-priced version 1 that was badly
flawed, followed by price reductions leading to virtually giving away version 3
(lousy marketing).
Realbasic have lowered their development costs by _NOT_ producing a Linux GUI
app, but producing the minimal functionality that meets customer requirements:
Linux (and Mac !) binaries. this means that it can survive on a similar revenue
base to Kylix because it has less overheads.
while the lack of a Linux GUI is a slight disadvantage, VMWare and Win4Lin work
sufficiently well that it does not matter that much. bugs that are sufficiently
subtle to surface on Linux but not on Windows will often require time-honored
DEBUGPRINT analysis rather than a nice GUI de{*word*81}. i don't know what the
status with Linux GNU de{*word*81} is, but i guess they have or will get that
working with their binaries.
and no, this is not an ad for them - it is a simple statement that the RealBasic
crowd seem to have delivered what Kylix promised, but never quite got right.
Kylix has a real, more capable competitor in its marketplace.
cheers,
Mat
 

Re:the end of the Kylix experiment ?

Mat Ballard wrote:
Quote
and no, this is not an ad for them - it is a simple statement that the
RealBasic crowd seem to have delivered what Kylix promised, but never
quite got right. Kylix has a real, more capable competitor in its
marketplace.

Requiring Windows to produce the app is not really my idea of "getting it
right." If you need to use Windows to start with, most Delphi apps will run
under Wine just fine, so why not use Delphi? Of course if you need or want
Mac, then it may be more of an option.
Of course Gambas is also a basic option for these two platforms and I do not
need Windows to develop with it. The price is right too, aka FREE.
I like the idea of being able to choose my development platform. Suppose I
do not use Windows and do not want to buy a license for Windows to do my
Linux development? What does Real Basic offer me then?
I am NOT putting down Real Basic. It might be more useful for those running
Windows and simply desiring to make Windows programs and being able to
deploy them to Linux. For this crowd it could work well. But for those who
want their apps to be xplatform, but primarily are targeting and developing
on Linux, I think Real Basic is a tad off in this area. It is kind of like
"Buy Windows so you can develop Linux applications." :)
Perhaps Real Basic will run under Wine. If so, it might be ok.
 

Re:the end of the Kylix experiment ?

Quote
while the lack of a Linux GUI is a slight disadvantage, VMWare and Win4Lin work
sufficiently well that it does not matter that much.
A Linux IDE is not necessary if there is a good remote debug feature,
having the target program run either on a separate Linux PC or in
a(nother) VMWARE client. But being able to debug the target running on
Linux _is_ essential.
-Michael
 

Re:the end of the Kylix experiment ?

"Mat Ballard" < XXXX@XXXXX.COM >wrote in message news:4043c9d2$ XXXX@XXXXX.COM ...
Quote
theo wrote:
>
>>
>>www.realsoftware.com/realbasic/

>What has this to do with your subject line??
>The IDE does not even run on Linux...

please correct me if my memory has become faulty with old age, but:

1. Kylix promised true cross-platform capability between Linux and Windows;
2. with true RAD capabilities, and,
3. was initally advertised as "VB for Linux".

it seems to have failed for a variety of reasons, including:

a. insufficient sales to cover development costs
Probably true.
b. the "7M tail" of any install, due to an inability to bind directly to either
of the major GUI frameworks on Linux.
I would say the 7MB tail was less important than the hoops one
had to go through to attached the tail to a deployed app.
Quote
c. pissing off the customer base with an over-priced version 1 that was badly
flawed, followed by price reductions leading to virtually giving away version 3
(lousy marketing).
From the posts I've read, those pissed off were angered by
Borland's failure to provide bug fixes.
Quote


Realbasic have lowered their development costs by _NOT_ producing a Linux GUI
app, but producing the minimal functionality that meets customer requirements:
Linux (and Mac !) binaries. this means that it can survive on a similar revenue
base to Kylix because it has less overheads.
Which makes it less than compelling .
Quote
and no, this is not an ad for them - it is a simple statement that the RealBasic
crowd seem to have delivered what Kylix promised,
I would disagree. But sales will tell whether they have succeeded at something.
--
Hilton Evans
-----------------------------------------------
ChemPen Chemical Structure Software
www.chempensoftware.com
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (www.grisoft.com).
Version: 6.0.598 / Virus Database: 380 - Release Date: 2/28/04
 

Re:the end of the Kylix experiment ?

Mat Ballard wrote:
Quote
g'day everyone,

check out:

www.realsoftware.com/realbasic/

a Visual Basic clone.

note that while the IDE is Win32, it can produce Linux (and Mac)
binaries using GTK 2.0:

www.realsoftware.com/realbasic/rb55screens/l_win_buildsettings.jpg


don't know how good it is but they do have a downloadable demo.



cheers,



Mat
i don't have problem, java run on linux, mac, windows...
--
Borland rulez pages.infinit.net/borland
 

Re:the end of the Kylix experiment ?

Marc Collin wrote:
Quote
i don't have problem, java run on linux, mac, windows...
{$DEFINE FLAME_BAIT_ON}
don't get me wrong, java is a great language for web apps, but for desktop apps:
1. it does not run, it walks
2. most Java desktop apps i've seen will only run on one platform, and can even
be picky about what what VM and what version.
in theory, Java is great (write once ...) - in practice, it is not a good
solution for desktop apps (... runs somewhere).
cheers,
Mat
ps: i'm currently fighting a losing battle with a very expensive pysch test
program (written in Java) that my wife bought: it will only run on about 50% of
W2k installations (and not the one we need), and it does not run under WXP. it
is even _LESS_ portable than a win16 or win32 app !
 

Re:the end of the Kylix experiment ?

Mat Ballard wrote:
Quote
Marc Collin wrote:

>i don't have problem, java run on linux, mac, windows...

{$DEFINE FLAME_BAIT_ON}

don't get me wrong, java is a great language for web apps, but for desktop
apps:

1. it does not run, it walks
2. most Java desktop apps i've seen will only run on one platform, and can
even be picky about what what VM and what version.

Someone doesn't know how to write a Java application then or they are using
a very old and archaic jre spec (aka 1.1). This was true in the AWT world,
it is not true in either the Swing or SWT world. Do you really believe that
Oracle JDeveloper, JBuilder, Net Beans, etc. is written differently for
each and every platform they run on? the only thing that is different, may
be the script or native wrappers used for the main portion of the program
and the installer (unless they are using a Java based installer). It is
true when certain app servers are included (like Borland's in JBuilder for
instance), that the app server may be platform specific, but the IDE code
is the same.
I have NEVER written a Java 2 GUI app that was not XPlatform, unless it was
intentional (ie making machine specific or OS specific calls). For these
type of apps, I NEVER INTENDED them to be xplatform. I wanted to tie them
directly to the OS for OS specific work.
Even then, most of them would run on most Unixes, if I was using Linux or
Unix to make the OS specific application.
It is true, that the JRE you are targeting is important. You should have
this worked out in your specifications. Most Java 2 jres are backwards
compatible, but not upward compatible (for instance a j2re 1.3.x
application will run under a 1.4.x jre, but a 1.4.x written app will not
always run correctly, on a 1.3.x jre, especially if you are using specific
1.4.x APIs or libs).
If you need to target both (say 1.3.x and 1.4.x), simply use the 1.3.x to
write and compile your code. Then it will work on both jres.
Quote
in theory, Java is great (write once ...) - in practice, it is not a good
solution for desktop apps (... runs somewhere).

When is the last time you used Java? Cannot be recently. Either that, or you
have a very old machine.
Look a few post down. I did a benchmark using Java Swing and using KDevelop
and GNU C++, for simple GUI apps. The difference in runtime and startup
time could be measured in milliseconds. I do not think any user is going to
measure the validity of an application startup or response time in
milliseconds.
If that kind of time difference is important, then I would suggest Assembler
and/or C.
NOTE: Please do not try and compare an IDE to a normal Java application.
JBuilder, for instance, is loading many libraries like Application Servers,
Testing Suites, J2EE and J2SE libraries. I do not know of too many user
applications that will be loading an app server, code testing suites,
JavaDoc readers and writers, UML modelers, and remote de{*word*81}s in an
application. If they do, then it is a developer IDE, not an end user
application. :)
Look at something like JEdit. How fast does it run in comparison to
something like say XEmacs? They will be very close, and these two are
similar in actions, albeit JEdit has a much more extensive Help System.
 

Re:the end of the Kylix experiment ?

"pnichols" < XXXX@XXXXX.COM >wrote
Quote
Read that.. I do not like BASIC language at all, but I will definitely
have
a gander at it, if they offer an eval copy.

The more better tools that become available, the better for all. :)

I agree, though with all the OOP changes, etc. it doesn't really look like
BASIC much anymore at all ...
Just like Borland now refers to OP as Delphi now.
IMHO, Borland really dropped the ball on interpreting their Linux marketing
research data. I think most developers who expressed interest in Linux were
really Windows developers who like to have the "option" to easily port their
applications to Linux ... to keep all their OS options open. A
cross-compiler from Windows that would create a Linux executable or just a
Linux-based compiler to create the executable with source code created on
Windows would probably have been adequate and more popular initially. Linux
developers just aren't interested in paying for tools whereas Windows
developers are. This desire on the part of Windows developers to create
Linux apps is still there but unfortunately Borland now has a bad taste in
their mouth with Linux development...
This is why I find RealBasic so interesting ... if they only can get the
bugs under control ...
 

Re:the end of the Kylix experiment ?

USCode wrote:
Quote
Linux developers just aren't interested in paying for tools whereas
Windows developers are.
That's total bollocks. I'd be most happy to pay for a tool that makes me
more productive.
Quote
This desire on the part of Windows developers to create Linux apps is
still there but unfortunately Borland now has a bad taste in their mouth
with Linux development...
That is because IMHO, they failed to understand the dynamics in the Linux
market and ended up binding themselves to Wine and a UI library of which
they could have known it was in heavy flux. What they /should/ have done is
provide those parts that were and partly still are sorely missing in the
Linux market at that time, a proper RAD UI design tool backed by a freely
available Delphi compiler toolchain and a freely available layer over the
UI lib used.
I think I would speak for many developers in the OSS domain they would be
very happy if Borland would provide them with the tools that would allow
them to develop quicker in combination with the means to redistribute their
code as source compileable by others. This has not happened.
--
Ruurd
 

Re:the end of the Kylix experiment ?

"R.F. Pels" < XXXX@XXXXX.COM >wrote in message
Quote

>Linux developers just aren't interested in paying for tools whereas
>Windows developers are.

That's total bollocks. I'd be most happy to pay for a tool that makes me
more productive.

No its not and using yourself an an example is anecdotal evidence. For
whatever reason they failed, we could argue that point all day, it's clear
they determined that the return wasn't worth the effort. If Linux
developers were scrambling to pay for such a tool then they probably would
have made the effort, but they didn't.
 

Re:the end of the Kylix experiment ?

Quote
>That's total bollocks. I'd be most happy to pay for a tool that makes me
>more productive.
USCode wrote:
No its not and using yourself an an example is anecdotal evidence. For
whatever reason they failed, we could argue that point all day, it's
clear
they determined that the return wasn't worth the effort. If Linux
developers were scrambling to pay for such a tool then they probably would
have made the effort, but they didn't.
Only complete stupid may pay for buggy unfinished product which Kylix is
bright example. Linux developer are not stupids.