ANNOUNCE: New KDevelop Design

[Dmitri Semenov]

Hello

 I've just signed up to the mail list.
First of all I would like to explain my background. The reason for this is
that a lot of the same people are going behind and in front of me.
I wrote almost all software for the dos (and some for the win32).  I used
MultiEdit + Watcom + Dos4GW. Do you remember? Not so long ago almost all
games used Watcom+D4GW? I was making marine simulators and marine onboard
systems. But. This summer we all have lost Watcom. Another thing has
happened:  only few companies are interested in DOS extenders and graphics
for the dos. I've started research and found that Linux has grown up from a
child to the man. I do have everything in Linux functionality what I need. I
do not have to use NT, embedded NT, DOS, QNX or other OS to make our
products. I believe the same way will be used by many other developers and
companies in the world. And all of us need some environment which is
familiar to us and powerful.  I spent some time looking for the IDE and have
found that Kdevelop is the very good choice. Of course I can use anything
including mcedit, vi, emac, code crusader. But I do not wish to change my
way to make software. I think may be my modest opinion will be interesting
for the Kdevelop team.

What is good in Kdevelop (KD):
1. Documentation tools integration is perfect. I was dreaming about
something like that for ages
2. Editor and Project Manager are good enough.
3.  I like the KD does not engender a lot of windows like a code crusader.
4. I like class, functions and data navigational systems.

What do I miss in KD?
1. MultiEdit functionality. IMHO MultiEdit style of built macro language and
expansibility was the great. I like intellisense from ME too
3. CVS integration is little bit not clear at first glance (I did not study
this exactly only tried once. sorry if I am wrong).
4. Documentation outside of KDE and QT is not complete yet. C++ and C
reference is a good point to start but it has bugs and lacks many functions
5. Debugger is good enough but is difficult to get arrays and rich assembler
source behind of C or C++. I had everything in turbo debugger. Watcom
debugger was powerful but Borland was much more comfortable.
6. Especially for me is no reason to have something different from C and
C++. But adding good macro language with possibility tune-up will make KD
only better
7. IMHO. The orientation primarily on the KDE only is not a good point.
Software developers like me need a good environment for making software for
wide range of targets (I mean not OS). For example I will use KD for making
targets for the GGI on the X and will replace GGI target for the
semi-embedded systems without X. DO NOT FORGET me and people like me when
discussing functionality. I am afraid that all good (for me) points in KD
will be lost in the KDE integration.

KD is good enough for me even this version. What I wont to see in the future
from my dream IDE.
Let me show:
1. Project manager, class-function-data-files navigation idea from KD are
very good especially with the documentation functionality.
2. Editor is good but defining own macro language for making "everything"
including redesigning of all interfaces is a very important thing to do. The
better somebody will make the system, which can be programmed in its own way
by own tools. That will automatically get off all questions with other
languages support.
3. Integrated Debugger is a good thing but in Unix like environment that is
not so important for good working. Everybody can use any debugger he wants.
4. I would like to see the possibility to switch-off all unnecessary
editors, wizards and so on from my ide.
About KDE and other windowed development I believe that only RAD tools can
be wide spread. There is a good example of Borland C++ Builder and Delphi
tools. But. That's only my IMHO.
Anyway I will be happy to use KD from this moment. Thanks to everyone doing
such the great Job. I believe in the future I will find the possibility to
join the people making free software because it makes them happy.
Best regards, Dmitry Semenov
Marine Electronics Corp. S. Korea


-----Original Message-----
From: owner-kdevelop at barney.cs.uni-potsdam.de
[mailto:owner-kdevelop at barney.cs.uni-potsdam.de]On Behalf Of Sandy Meier
Sent: Saturday, October 02, 1999 12:38
To: kdevelop at barney.cs.uni-potsdam.de
Subject: Re: ANNOUNCE: New KDevelop Design


Hi!

First sorry to all for my late answer and contribution to this discussion. I
hope we will find the best solution for all participants.

>
> c) the new branch starts from scratch with KOM/OP as a base and takes
> needed codeparts from the current branch, such as some classes, the
> parser, treeviews etc.

As far as I know the KDE team is planning to drop parts of the current
CORBA/KOM/OP technology and replace it with something other (canossa).
Nevertheless some parts of KDE2 will use CORBA but it's not recommend to
design
a complete app on this approach. See
http://lists.kde.org/?l=kde-core-devel&m=93870838307236&w=2 for detailed
informations.

> we would have to spend some time to learn about it. Not to speak of the
> plugins for other languages.
>
> If that would make anyone happy, I would agree with the idea of a new
> branch, so we can all take part on this.
Yes I agree, IF we (Catalin's team,kdevelop team and some others) will
implement a new KDevelop, than we should take all part on this.

> - KOM/OP design
see above
> - general User interface: dialogs etc.
> - a dialogeditor team : suggest Java/AWT support, gtk+ support,
very difficult :-( see
http://lists.kde.org/?l=kde-core-devel&m=93871452415241&w=2
> - language team
> - project interfaces team
>
> Let's start discussing it. Do we want a complete restart or not ?

I would prefer to work on the existing codebase, that means an IDE only for
C/C++ and QT/KDE.
Why do we need an IDE for all programming languages? There are already
nice IDE's for Java (netbeans,freebuilder,codeguide) and say Python (IDLE).
And if you want to code for KDE 99% of the users will take C++, or?
I think, if we concentrate our effort on C/C++ we can add an integrated
debugger, dockwidget, MDI , intellisense, plugins and so on in the next
months.
But I doubt that we can do that, if we start from scratch by taking some
reusable code from the current KDevelop. Even with the help of the Catalin's
team.

Ciao!

Sandy

P.S. Maybe I should setup a new mailinglist to reduse traffic on this list?