Board index » kylix » Re: Borland no longer part of Trolltech

Re: Borland no longer part of Trolltech


2005-05-27 11:29:48 PM
kylix0
Dave Nottage [TeamB] wrote:
Quote

FWIW, I'm producing a desktop product (although it is for one customer
only, but potentially thousands of units) with the current version of
Kylix, and it is not Kylix that makes it scary; it is Linux that does.

How so? If you are designing an application for one client then you
shouldn't have any issue with Linux, specially, if your intended
target is multiple deployment in one organization. You make your
application, test and configure it for one particular Linux
distribution and deploy your application together with that specific
Linux distribution. This approach works very well and often the
end-user in normal day-to-day use of your application doesn't have any
idea (or doesn't care) on what operating system it runs on. Linux is
an excellent choice to host such custom application (for one
organization - multiple deployments ) because it permits to be easily
customized. In situation when you cannot do this then a Linux system
administrator will take care of the multiple deployments, set up and
maintain your application on all involved workstations.
While Kylix application will work well on variety of Linux distros we
cannot say the same about Kylix. The Kylix IDE is based on very old
and heavily hacked wine library. The Kylix IDE is very fragile and
sensitive to environment changes, even if you install it on supported
Linux distribution the Kylix IDE may or may not work correctly.
(That's why one person will tell you it works for me very well other
one will tell you that Kylix is unusable)
The solution is to assemble such development system from selected
components, network it with other Linux boxes and fine tune it to
Kylix needs. Further, it depends from the type of application you
do. For example it is really a nightmare to design a database
application with Kylix. The IDE designer my work fine for simple
database connectivity but if you have more complex database structure
then it suffers form unexpected setting losses, not responding to
changes of components property values, weird behavior, sluggishness
etc.. in other words, it stops to be RAD. Basically, it is necessary
to complement Kylix with external custom made tools to do this type of
development, use third-party components and best to avoid the use of
Kylix dbExpress. Then, there is the persistent problems related to
Kylix incompatibility with current Linux technology, due to lack of
support and updates for Kylix, but I think it is pointless to state
the obvious again and again...
CrossKylix is an excellent way to overcome some of the Kylix IDE
problems, but it is not a substitute or solution for absence of
Borlands support for Kylix and unfortunately will not help much users
of Kylix which do not own Delphi 6/7 (which by the way are removed
from the marked by Borland and Delphi 2005 does not have CLX) You can
go and buy Kylix today, but you will have problems to get Delphi 7, so
how you can use CrossKylix or how you can do the cross-development for
Linux and Windows ?witch is the advertised main advantage of using
Kylix/Delphi? Beside, for some type of applications it is just
necessary to develop it and debug on Linux, so it looks like another
problem to be solve.
Thanks god (and not only) we have the unofficial patches for Kylix,
but there is a limit what they can improve. Borland controls Kylix
IDE and only they can fix it. Kylix users won't be able to do this
because of lack of access to Kylix source code. If the Kylix IDE
cannot be open-sourced for some legal reasons and if Borland doesn't
have resources to do the update, I do not see why the source code for
the IDE couldn't be make available to selected willing developers
under NDA, it surely will not cost Borland anything and we all could
enjoy unofficial Kylix (unKylix) as we are enjoying the unofficial
patches.
juliusz
--
InstallMade - Kylix-specific installer
www.superobject.com/installmade/
www.superobject.com/imoe/download.html
 
 

Re:Re: Borland no longer part of Trolltech

Quote
Dave Nottage [TeamB] wrote:
>
>
>FWIW, it worked fine for me. That is until I started using Delphi 7 and
>CrossKylix, so now the speed of the Kylix IDE on Linux makes no
>difference since I don't (or hardly ever) use it.
>


Kylix 3 works fine for me, although if I were starting a new project I
would go with CrossKylix. So far now you can produce a Linux desktop app
easily that will work great. However as far as I'm concerned that's
only possible thanks to the efforts of Mr. Hausladen and Mr. Kissel.

What happens 3-5 years from now when your customers are demanding an
updated user interface, and those gentlemen have moved on to other projects.

Kylix or CrossKylix really needs a long term commitment from Borland to
be a viable option for me.
Of course the optimal solution would be Borland to step up on this, and I still
hope they'll do.
And while I'm probably not as "futureproof" as a big company, you can rest
assured that CrossKylix and CrossFPC will get updated and maintained for
a LONG period. There is a simple reason why I can't move on to other projects:
The company I own and work for sells server software products and solutions,
and there is a gigantic object pascal codebase. And our enterprise customers
demand Linux server applications. The share of Linux vs Windows product copys
we are selling meanwhile is somewhere near 80/20, while it was the other way
round a few years ago.
Looking at possible migration paths I don't see anything that would come
even close to OP, dbexpress and the whole Delphi infrastructure.
As long as I don't get overrun by a bus or something, I'll make sure that
at least Delphi/Kylix server applications will run on all current Linux
targets. If Borland should decide to drop Kylix, FPC and CrossFPC
will fill the gap.
And in case I'm gone one day, those projects all are open-source and someone
could take over the job.
Simon
 

Re:Re: Borland no longer part of Trolltech

Gesundheit!
"Simon Kissel" < XXXX@XXXXX.COM >schrieb im Newsbeitrag
Quote
Dave Nottage [TeamB] wrote:
>
>
>FWIW, it worked fine for me. That is until I started using Delphi 7 and
>CrossKylix, so now the speed of the Kylix IDE on Linux makes no
>difference since I don't (or hardly ever) use it.
>


Kylix 3 works fine for me, although if I were starting a new project I
would go with CrossKylix. So far now you can produce a Linux desktop app
easily that will work great. However as far as I'm concerned that's
only possible thanks to the efforts of Mr. Hausladen and Mr. Kissel.

What happens 3-5 years from now when your customers are demanding an
updated user interface, and those gentlemen have moved on to other
projects.

Kylix or CrossKylix really needs a long term commitment from Borland to
be a viable option for me.
Of course the optimal solution would be Borland to step up on this, and I
still
hope they'll do.
And while I'm probably not as "futureproof" as a big company, you can rest
assured that CrossKylix and CrossFPC will get updated and maintained for
a LONG period. There is a simple reason why I can't move on to other
projects:
The company I own and work for sells server software products and solutions,
and there is a gigantic object pascal codebase. And our enterprise customers
demand Linux server applications. The share of Linux vs Windows product
copys
we are selling meanwhile is somewhere near 80/20, while it was the other way
round a few years ago.
Looking at possible migration paths I don't see anything that would come
even close to OP, dbexpress and the whole Delphi infrastructure.
As long as I don't get overrun by a bus or something, I'll make sure that
at least Delphi/Kylix server applications will run on all current Linux
targets. If Borland should decide to drop Kylix, FPC and CrossFPC
will fill the gap.
And in case I'm gone one day, those projects all are open-source and someone
could take over the job.
Simon
 

{smallsort}

Re:Re: Borland no longer part of Trolltech

Dave Nottage [TeamB] sighed and wrote::
Quote
Edmund wrote:


>Just a matter of curiosity, why's that? Aside for the conceptual
>differences in how things are executed within the Linux desktop,
>what else?


Lack of knowledge/understanding on the customers part is a huge one.
Yes, I would say that's a major obstacle. The Windows mindset is
so ingrained into the customer, it'd be hard to 'unset' this
concept and have them try a go at Linux.
Quote
Their Windows knowledge far outstrips their knowledge of Linux, and
gaining Linux knowledge is prohibitive due to a number of factors,
including multiple distributions, which can make it confusing and
difficult for a lot of people to gain an understanding of. Multiple
That's another reason why it's getting a little difficult to keep
up in the Linux world. But the issue is, an application developed
on, say Red Hat, shouldn't really have trouble compiling in, say
Mandrake. The problem is library dependency and stuff.
I dream that one day, I'll develop every piece of software under
Linux and run any application under that environment. As it stands,
my programming skills isn't quite up to par to do anything.
Quote

FWIW, it worked fine for me. That is until I started using Delphi 7 and
CrossKylix, so now the speed of the Kylix IDE on Linux makes no
difference since I don't (or hardly ever) use it.
I see. I'll have to check out CrossKylix. And perhaps do a bit
more work on Delphi.
Thanks
Edmund
 

Re:Re: Borland no longer part of Trolltech

Quote
delphi2006 will include C++ personality, perhaps a WXCL could start ?
as VCL alternative...
There is already a Borland wxWindows C++ success story called CBuilderX
www.borland.com/cbuilderx
Since 2 years no further development ... looks like a Kylix clone;-)
Ronald