KDE and Gnome, Qt and GTK [please read this]

6 views
Skip to first unread message

Conrad Sanderson

unread,
Nov 6, 1997, 3:00:00 AM11/6/97
to

This message is a plea to resolve the KDE vs Gnome situation.
Please read it and comment on it.

KDE project: http://www.kde.org
Gnome project: http://www.gnome.org

On Thu, David Finton <dfi...@d.umn.edu> wrote:

: I've been reading the newsgroups and heard all the opinions about
: whether or not people should use a software package based off of
: free
: software, or commercial software. Personally, I like KDE.
: I'm using it
: right now in fact :). In fact I think there should be more
: projects
: like this. Which is why I also support Gnome.

: [...]
: Is KDE the end-all and be-all of WM's? I don't think so.

Both KDE and Gnome are not just Window Managers - they are much more
than that. They integrate a lot of functionality which was
previously unavailable in Linux.

: Will Gnome knock out KDE and be the king? I hope not.
:
: I like having choice. That's why I picked Linux as the OS that
: inhabits my computer. There are lots of choices, and the range
: is expanding all the time. The more the merrier! 8^)

There is a difference between picking an OS and picking a
desktop to use on the OS.

The problem with choosing a desktop is that apps will be written
for either KDE or GNOME, because both use different GUI
libraries and different methods to communicate with the rest
of the system.

So in effect we end up with 2 different systems which do the
same thing. And at the same time instead of having a uniform
system where everybody can develop for it, we get split into 2
camps. Pathetic to say the least. Our development efforts get
diluted by half effectively.

If we kept with one desktop system and didn't like certain parts
of it (for example the Window Manager) we could always write a
new one which was compataible with all of the integration.

KDE or Gnome don't limit our choices - they just make Linux
better, faster, easier and more fun to use. KDE and Gnome
are filling the void in standards that X has given us,
resulting in programs which look and behave different from each
other, and an almost complete lack of integration in the GUI.

*** Please read the 2 sections below *** This situation
has to be resolved.

** section 1:

Let's review the facts about KDE and GNOME:

- there are 2 desktop systems being developed that basically
offer exactly the same thing

- if these 2 systems are complete, then they would be competing
with each other

- during the competition between these systems, apps would appear
on one or the other

- similar apps would appear on both KDE and Gnome, meaning there
is duplicated effort

- KDE is much further ahead than Gnome in development
(ie. it will be finished earlier)

- KDE uses the Qt library

- Gnome uses the GTK library

- Qt is a C++ library

- GTK is a C library

- Qt has better documentation than GTK

- Qt has been around longer than GTK

- Qt is technically "better" than GTK

- GTK is released under the LGPL license.

- Qt is owned by Troll Tech, and it is free to use for free
software development.

- Commercial software using Qt has to pay trolltech for a license.

- GTK can be used in free and commercial software exactly the same.

- The Qt library cannot be modified except by Trolltech. you can
still submit bug reports/suggestions to trolltech.

- a lot of GPL-advocates have a problem with Qt because of its
licensing restrictions (ie. can't modify it directly, commercial
software can't write for KDE without paying Trolltech)

- a lot of Qt-advocates say that the free-software license
is not that restrictive since free-software developers can
use it for free. The commercial fee is fair since it's
commercial software.

- RedHat and Debian stated that they will not distribute Qt on CDs
(but have it in contrib or non-free sections of their
distributions).

- RedHat and Debian will ship GNOME as the default desktop when it
is available.

- Suse says they will distribute KDE as their default desktop
(this is going to occur earlier than RedHat/Debian shipping
Gnome)

- there is an effort to create (from scratch) a free version of
the Qt library ( http://www.fsf.org/prep/tasks_2.html#SEC2 )


Conclusion: a split in the Linux community is going to occur.

(We can of course continue to write apps in the same old way
and not evolve. But people will and do want integration).


** section 2:

So let's review our options with regard to this pathetic situation.

1. Situation stays unresolved. When both systems are complete,
we basically have a KDE vs Gnome software "war".
[ "war" = split ]

2. All KDE development is stopped and existing KDE apps and
components are recycled into Gnome

3. KDE is converted from Qt to GTK

4. Some of the development going into KDE is diverted to
writing a free clone of Qt

5. A seperate project is initiated to write a free Qt clone
[ note: FSF have the free QT clone listed as one of the
highest priority tasks - but is anybody actualy doing
anything about it ? ]

6. TrollTech changes the license of the Qt library to LGPL,
with Trolltech being the primary developers for Qt.

7. KDE and Gnome development stays as it is, and then a freeQT
library appears, released under LGPL. This makes KDE
completely free, undermining the Gnome development.
KDE will have the upper hand in the software "war".

8. Gnome development is stopped and the effort is put into
writing the freeQt clone.

9. Gnome and KDE people actually work _together_ on how
to make programs written for either desktop system
work on each others systems (ie. retain integration).

Notes: if a free Qt clone appears, there will be little incentive
to use the original Trolltech version, hence making the original
Qt basically useless under X11 - Trolltech might as well release
X11 Qt under LGPL now.

Conclusions:

Future prediction if nothing changes: option 1, eventually followed
by option 7.

The easiest option to take would be 9 (if egos can be put aside).

I can see a significant amount of people getting pissed off enough
to take option 5, resulting in option 7.

My recommendation: option 9

Igor Laera

unread,
Nov 6, 1997, 3:00:00 AM11/6/97
to

On 6 Nov 1997 13:14:19 GMT, con...@hive.me.gu.edu.au (Conrad
Sanderson) wrote:

>5. A seperate project is initiated to write a free Qt clone
> [ note: FSF have the free QT clone listed as one of the
> highest priority tasks - but is anybody actualy doing
> anything about it ? ]

>8. Gnome development is stopped and the effort is put into
> writing the freeQt clone.

Is there anybody out there, who appreciated the work of Trolltech ?
They invested much sweat in the linux-development, and you "thank"
them with start of a freeQt ?

Before Trolltech, a unified library wasn't in sight, now two big
projects can't await to get into the sunlight.

From a developer sight of view, instead of writing KDE and
putting (wasted?) manpower into FreeQt, GTK should evolve
more and better. (What about network libs ?)

Yet, this all doesn't looks very attractive to possible commercial
developers...

greets
Igor
---
Attention: remove SPAM from email adress

Rajat Datta

unread,
Nov 6, 1997, 3:00:00 AM11/6/97
to

On Thu, 06 Nov 1997 18:40:41 GMT, Igor Laera <ila...@artefakt.SPAM.com> wrote:
>On 6 Nov 1997 13:14:19 GMT, con...@hive.me.gu.edu.au (Conrad
>Sanderson) wrote:
>
>>5. A seperate project is initiated to write a free Qt clone
>> [ note: FSF have the free QT clone listed as one of the
>> highest priority tasks - but is anybody actualy doing
>> anything about it ? ]
>
>>8. Gnome development is stopped and the effort is put into
>> writing the freeQt clone.
>
>Is there anybody out there, who appreciated the work of Trolltech ?
>They invested much sweat in the linux-development, and you "thank"
>them with start of a freeQt ?
>
>Before Trolltech, a unified library wasn't in sight, now two big
>projects can't await to get into the sunlight.
>
>From a developer sight of view, instead of writing KDE and
>putting (wasted?) manpower into FreeQt, GTK should evolve
>more and better. (What about network libs ?)
>
>Yet, this all doesn't looks very attractive to possible commercial
>developers...
>

I'm afraid this is an example of people bankrupt of ideas. FreeQT
is an attempt to piggyback on the really hard work of design that
Troll Tech has done by those incapable of doing anything besides
copying.

GTK has a lot of promise and a lot of necessary development. There
is room for real innovative work. That's what I had hoped would be
the result of competition in free software. More innovative works
and less of mindless copying.

Troll Tech has done a fabulous job. I've never encountered another
GUI toolkit that is this easy to program to. Also, the hypertext
documentation is fantastically useful. Motif may cost a lot less, but
it's more than many times harder to program to. The challenge for
GTK, or any other toolkit, to match this level of quality is daunting,
but, man, it should be fun.

rajat

Greg Retkowski

unread,
Nov 6, 1997, 3:00:00 AM11/6/97
to

robert havoc pennington wrote:

[Alot of stuff about QT becoming free, KDE converting to GTK, Gnome being
abandoned, a war betweeen Gnome and KDE, blah blah blah... ]

Here's my thought. a QT wrapper for GTK. Then your KDE and Gnome apps are
compatable (it all comes back to GTK). You avoid the entire Gnome vs.
KDE situation, You have a 'FreeQT', you arent re-writing QT from scratch,
and it causes world peace, harmony and the downfall of microsoft. Can we
have a group hug ;)

-- Greg


Vlad 'Borg' Petersen

unread,
Nov 6, 1997, 3:00:00 AM11/6/97
to

Conrad Sanderson wrote:
..........
>
> I didn't mean using different GUI toolkits (we can't win this battle).
> I meant More along the lines of making some sort of common API for
> drag'n'drop, program restart, localisation, preferences (eg. colors
> of menu strips) etc.
........
> This way, both KDE and Gnome can co-exist, with programs written
> for one or the other working with about the same level of integration
> with the desktop.

That would make both libraries/toolkits more bloated.
Let's not go the way of Windows where the users are
wandering through myriads of confusing *dll's. Nowadays,
it is very common for the average Joe X user to run
several applications at a time based on different libraries
which bring the benefits of shared libraries to zero.
In an ideal situation (I realize it will never happen),
all applications _share_ the same library.

> Qt is like a better Motif. Even if "nobody" supports it, it will linger
> on for quite some time, just like Motif is doing.
>
> btw, finally figured out why the X consortium "gave" us those ultra-ugly
> and user un-friendly Athena widgets. It was to make people buy Motif.

Athena was never intended to be a complete toolkit
solution but only a tutorial, an example of how to
create a new toolkit. Unfortunately, some people took
it seriously and started using it for writing real apps.

--
My real email address is: <vladimip at uniserve dot com>
#include <disclaimer.h> | *Good pings come in small packets*
Vancouver, BC | Windows: for IQs smaller than 95
SIGSIG -- signature too long (core dumped)

Jerome Kalifa

unread,
Nov 6, 1997, 3:00:00 AM11/6/97
to

con...@hive.me.gu.edu.au (Conrad Sanderson) writes:

[...]


>
> - if these 2 systems are complete, then they would be competing
> with each other
>

This competition has begun and motivated both parts. Until now, it has
been positive.

> - during the competition between these systems, apps would appear
> on one or the other
>
> - similar apps would appear on both KDE and Gnome, meaning there
> is duplicated effort
>
> - KDE is much further ahead than Gnome in development
> (ie. it will be finished earlier)
>

As a Window Manager, yes. But for most important applications, no :
just compare Gimp and KPhotoshop (haven't seen it yet), and
consider also the Mnemonic project (based on GTK) is of major
importance.

>
> - GTK is a C library
>

...which can be interfaced with C++, Python, and Guile.



> - Qt has better documentation than GTK
>
> - Qt has been around longer than GTK
>

Yes, but its way of development (i.e Troll and nothing else) doesn't
enable Qt to evolve as fast as GTK does.

> - Qt is technically "better" than GTK
>

I don't subscribe to this point of view. Moreover, side projects of
GTK like Guile and Gubi are important.
Let's just say that Qt is more mature than GTK.

[...]


>
> - RedHat and Debian stated that they will not distribute Qt on CDs
> (but have it in contrib or non-free sections of their
> distributions).
>
> - RedHat and Debian will ship GNOME as the default desktop when it
> is available.
>

IMHO these are the key points.



> - Suse says they will distribute KDE as their default desktop
> (this is going to occur earlier than RedHat/Debian shipping
> Gnome)
>

KDE and Suse are both Germans, I think.

> - there is an effort to create (from scratch) a free version of
> the Qt library ( http://www.fsf.org/prep/tasks_2.html#SEC2 )
>

80000 lines of code, exactly like KDE. Is it really more useful than
rewriting KDE with another library?

>
> Conclusion: a split in the Linux community is going to occur.
>

Another proof that Linux is an Unix :).



> (We can of course continue to write apps in the same old way
> and not evolve. But people will and do want integration).
>

Some people do, others don't care. And there are others solutions : A
desktop using WindowMaker or AfterStep + the AfterStep Control Panel +
TkStep + nextaw would have the same advantages as KDE (easy
configurability, integration).
http://www.inf.ufrgs.br/~kojima/


>
> ** section 2:
>
> So let's review our options with regard to this pathetic situation.
>
> 1. Situation stays unresolved. When both systems are complete,
> we basically have a KDE vs Gnome software "war".
> [ "war" = split ]

GNU Emacs and XEmacs, right? Both are feeling fine.

>
> 2. All KDE development is stopped and existing KDE apps and
> components are recycled into Gnome
>

Well, as I'm rather pro-GTK and pro-Gnome, sounds reasonable for me.
Others won't like it.

> 3. KDE is converted from Qt to GTK
>

This is Gnome, except for some details.

> 4. Some of the development going into KDE is diverted to
> writing a free clone of Qt
>

IMHO this is unuseful : too much work, compared as rewriting KDE on
GTK (i.e writing Gnome).

> 5. A seperate project is initiated to write a free Qt clone

Same answer.

> [ note: FSF have the free QT clone listed as one of the
> highest priority tasks - but is anybody actualy doing
> anything about it ? ]
>
> 6. TrollTech changes the license of the Qt library to LGPL,
> with Trolltech being the primary developers for Qt.
>

I don't know if this is easy to do.

> 7. KDE and Gnome development stays as it is, and then a freeQT
> library appears, released under LGPL. This makes KDE
> completely free, undermining the Gnome development.
> KDE will have the upper hand in the software "war".
>

Gnome will surely be mature before a FreeQT library appears.

> 8. Gnome development is stopped and the effort is put into
> writing the freeQt clone.
>

Gimp and Mnemonic are too important to stop the development of a
desktop based on GTK.

> 9. Gnome and KDE people actually work _together_ on how
> to make programs written for either desktop system
> work on each others systems (ie. retain integration).
>

With a different API? It's difficult and it's a big amount of job.

[...]

--
Jerome Kalifa
Centre de Mathematiques Appliquees, Ecole Polytechnique.
91128 Palaiseau Cedex, France. (33)169333981

robert havoc pennington

unread,
Nov 6, 1997, 3:00:00 AM11/6/97
to

con...@hive.me.gu.edu.au (Conrad Sanderson) writes:
>
> - KDE is much further ahead than Gnome in development
> (ie. it will be finished earlier)
>

Probably not by much. KDE only has a year head start. Gnome is building
on the the KDE experience and maybe some KDE code. Gnome is also using
Guile instead of C for some bits, which should go faster. And they
aren't writing everything from scratch, e.g. they're using Midnight
Commander as a filemanager base. So Gnome should catch up to KDE
pretty quickly, I would think. KDE speeds things up maybe half a year,
a year at most. Let's be patient and get it right.



> - Qt has better documentation than GTK
>
> - Qt has been around longer than GTK
>
> - Qt is technically "better" than GTK
>

Keep in mind that Gtk can keep evolving --- features can be added,
etc., like pluggable look and feel. Qt is what it is, or at least
whatever it becomes isn't controllable by the free software community.
So that's a Gtk plus.



> - there is an effort to create (from scratch) a free version of
> the Qt library ( http://www.fsf.org/prep/tasks_2.html#SEC2 )
>

Seems to me that it makes a lot more sense to expend this effort
adding zillions of features to Gtk, making it bug-free, and writing
docs.



> 1. Situation stays unresolved. When both systems are complete,
> we basically have a KDE vs Gnome software "war".
> [ "war" = split ]
>

Pretty likely, unfortunately. It's not like Gnome and KDE have really
different functionality, though they might be slightly different.
And the API's aren't compatible, so it's not like all the free Unices
where apps port between them reasonably well. Bad situation.

> 2. All KDE development is stopped and existing KDE apps and
> components are recycled into Gnome
>

Only real option to avoid having both of them.

> 3. KDE is converted from Qt to GTK
>

Doesn't buy anything. Pointless extra effort compared to option 2.

> 6. TrollTech changes the license of the Qt library to LGPL,
> with Trolltech being the primary developers for Qt.
>

That would be nice, but doesn't seem likely.



> 7. KDE and Gnome development stays as it is, and then a freeQT
> library appears, released under LGPL. This makes KDE
> completely free, undermining the Gnome development.
> KDE will have the upper hand in the software "war".
>

Doubtful. Gnome will have caught up by the time a FreeQt is written,
especially since a lot of KDE development time will have been spent
either writing or debugging freeQt.

> 8. Gnome development is stopped and the effort is put into
> writing the freeQt clone.
>

Requires porting Gimp and current web browser projects. Requires not
using Guile. Abandons possible improvements to Gtk in favor of
starting from scratch to write freeQt. Doesn't have Gtk's
multilanguage capabilities. Takes a good long time before everything
is free and works properly - we're sort of in an uncomfortable limbo
until freeQt is finished. Doesn't have the support of the two major
US distributions.

All are sort of small arguments, but taken together it sounds like KDE
merging into Gnome makes sense.

> 9. Gnome and KDE people actually work _together_ on how
> to make programs written for either desktop system
> work on each others systems (ie. retain integration).
>

I think they're trying, as much as possible, but k* apps will always
require Qt, which is the objection in the first place. So they will
all have to be duplicated with Gtk, or freeQt will have to be written,
and both of those seem harder than just ditching Qt entirely and going
to Gnome for all apps.

> My recommendation: option 9
>

I like 2, but I don't expect it to happen. :)

My prediction is that lots of effort will be wasted on KDE and
eventually Qt will be abandoned, as long as Gnome comes together
within a year or so. But time will tell.

Havoc Pennington

Kenton E. Sinner

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

Rajat Datta (ra...@goteborg.netcom.com) wrote:

...

> I'm afraid this is an example of people bankrupt of ideas. FreeQT
> is an attempt to piggyback on the really hard work of design that
> Troll Tech has done by those incapable of doing anything besides
> copying.

> GTK has a lot of promise and a lot of necessary development. There
> is room for real innovative work. That's what I had hoped would be
> the result of competition in free software. More innovative works
> and less of mindless copying.

It seems that this is exactly what Stallman was ranting against in
his Manifesto. The nature of intellectual property requires that
the wheel be reinvented by everyone in order to do anything useful.
Hence free substitutes for restricted software. Once one free
implementation is created, that is the LAST time that particular wheel will
need to be reinvented. Then people can get on with the fun stuff. :-)

--
"Try being one of *them* for just one year, knowing all the time that
no matter how strong or fast or smart you are, that no matter how many
tricks you learn---you still lose!" -- Mythos

Can't WWW, dont't ask me. :-(

robert havoc pennington

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

Greg Retkowski <gr...@pocketscience.com> writes:
> robert havoc pennington wrote:
>
> [Alot of stuff about QT becoming free, KDE converting to GTK, Gnome being
> abandoned, a war betweeen Gnome and KDE, blah blah blah... ]
>

Hey, don't act like _I_ brought it up. I was just responding to the
summary/request for discussion someone else posted...



> Here's my thought. a QT wrapper for GTK. Then your KDE and Gnome
> apps are compatable (it all comes back to GTK). You avoid the
> entire Gnome vs. KDE situation, You have a 'FreeQT', you arent
> re-writing QT from scratch,

Many people have had that thought, and it's not especially feasible.
Also compatibility goes beyond the GUI toolkit - it's init files, drag and
drop, etc. that really have to be compatible.

> and it causes world peace, harmony and the downfall of microsoft. Can we
> have a group hug ;)
>

That part I'm in favor of. :)

Havoc Pennington

robert havoc pennington

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

con...@hive.me.gu.edu.au (Conrad Sanderson) writes:
> yes, but i don't see KDE people giving up. And a lot of people like Qt.
> There must be some way for GTK and Qt to coexist. Hence the optiion
> below:
>

I don't see them giving up either, unfortunately. I can't say I'd be
too eager to give up after a year of work. But I just can't fathom
what possessed them to use a non-free library.


> I didn't mean using different GUI toolkits (we can't win this battle).
> I meant More along the lines of making some sort of common API for
> drag'n'drop, program restart, localisation, preferences (eg. colors
> of menu strips) etc.
>

I agree that's better than nothing, but it still sucks, because the
real concern here is apps written for Qt. The reason we don't want KDE
is that anything written for Qt has to be (or will be) re-written, or
Qt itself has to be rewritten. We'll have an endless repetition of the
same KDE-Gnome split. Qt = lots of wasted effort. Qt is the problem,
not KDE. KDE is GPL. Problem is that KDE and all apps depend on Qt.

Compatibility is good and I'm sure it will happen, but it's not really
the issue. It doesn't solve the problem created by having both
desktops. I guess it might encourage people to write apps in Gtk,
since they'll still work on KDE, but if everyone writes apps in Gtk
which use a Gnome/KDE common lib why have KDE at all?

Havoc Pennington

David Finton

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

Conrad Sanderson wrote:

> Both KDE and Gnome are not just Window Managers - they are much more
> than that. They integrate a lot of functionality which was
> previously unavailable in Linux.

You are right on this one... Qt and Gnome are both projects that have
long been overdue for the Un*x environment. Most of the previous window
managers have shown that you can use a GUI through Unix. These projects
show that you can use Un*x through a GUI. A great improvemant, if
anything.

> There is a difference between picking an OS and picking a
> desktop to use on the OS.

Depends on what OS you use. In Win95, the desktop *is* the OS. That's
where Linux and its cousins have a distinct edge. The GUI has a much
better chance of developing faster if you have alternatives. Better GUI
= more effective operating system.

> The problem with choosing a desktop is that apps will be written
> for either KDE or GNOME, because both use different GUI
> libraries and different methods to communicate with the rest
> of the system.

This situation is no different that it has been before. In fact, it's
always been the case where several GUI toolkits have been available,
like Motif, Tkl/Tk, and so on. KDE will have its own set of strengths
and weaknesses, and so will Gnome. Linux and its family of apps won't
be hurt by having alternative libraries.

> So in effect we end up with 2 different systems which do the
> same thing. And at the same time instead of having a uniform
> system where everybody can develop for it, we get split into 2
> camps. Pathetic to say the least. Our development efforts get
> diluted by half effectively.

The Gnome philosophy is to fill in the gaps that KDE has left behind.
And I think that when KDE and Gnome becomes more robust, the two will
diverge.

Another thing is that the effort is not "split" in half. Both KDE and
Gnome are cooperating with each other. They're not really in
competition with each other (they're both GNU). Well okay they are in a
friendly competition, but neither project is getting hurt by the other.
Both libraries will fit comfortably on most people's hard drives (as
they do on mine, and my hard drive isn't exactly bursting at the edges
with free space). If people like both libraries, they'll download both
of them.

> If we kept with one desktop system and didn't like certain parts
> of it (for example the Window Manager) we could always write a
> new one which was compataible with all of the integration.
>
> KDE or Gnome don't limit our choices - they just make Linux
> better, faster, easier and more fun to use. KDE and Gnome
> are filling the void in standards that X has given us,
> resulting in programs which look and behave different from each
> other, and an almost complete lack of integration in the GUI.

While I agree that Linux apps need a little bit more consistency, I
don't think that consistency has to outweigh the actual functionality of
the program. We learned that lesson from Win95 (too flashy, runs
crappy).

> *** Please read the 2 sections below *** This situation
> has to be resolved.

<snip>

>
> Conclusion: a split in the Linux community is going to occur.

There have *always* been several parallel project running at once in the
Un*x community. And who said competition is bad for the Linux
community? The better guy will win, or both will prevail.

> (We can of course continue to write apps in the same old way
> and not evolve. But people will and do want integration).

Yes! :)

Liszkowski Joerg

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

In article <slrn664rm9...@hive.me.gu.edu.au>, con...@hive.me.gu.edu.au says...
>

>>> 9. Gnome and KDE people actually work _together_ on how
>>> to make programs written for either desktop system
>>> work on each others systems (ie. retain integration).

>I didn't mean using different GUI toolkits (we can't win this battle).


>I meant More along the lines of making some sort of common API for
>drag'n'drop, program restart, localisation, preferences (eg. colors
>of menu strips) etc.
>

Well A common library for both wold be good, but I fear it won't happen.
But many things can be solfed - or are allready solved of the common API parts:
localisation: both build on 'gettext'
drag'n'drop: do not know - I'll have a try - but I thing both use the code first
implented by offix
preferences: here wold be the bigges differences, but maybe it is possible to write a
preferences manager who would write Gnome and KDE perf.-files
(at least for App. color's and fonts, keyboard shortcuts)
Many other things would be anyway only for the App. internaly - so it needs it's own
pref. manager

Conclusion: There will, hopefuly, a coexistence

Joerg


Simon Kinahan

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

robert havoc pennington wrote:

> I don't see them giving up either, unfortunately. I can't say I'd be
> too eager to give up after a year of work. But I just can't fathom
> what possessed them to use a non-free library.

I sympathise, but I see where they are coming from. Qt is
'nearly free' as far as *most* linux development is concerned.
They clearly reckoned that it was worth it seeing that:

1. Athena looks horrible, and is hard to use.
2. Motif is hard to use, and has bizarre licensing arrangements,
which are beyond my understanding.
3. Gtk (or [insert other free toolkit]) is poorly documented,
unfinished, and, if you like C++ (which many people do for
high level apps), rather C-centric.

Simon

Simon Kinahan

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

David Finton wrote:
> This situation is no different that it has been before. In fact, it's
> always been the case where several GUI toolkits have been available,
> like Motif, Tkl/Tk, and so on. KDE will have its own set of strengths
> and weaknesses, and so will Gnome. Linux and its family of apps won't
> be hurt by having alternative libraries.

Yes. It will (be hurt that is). The point of something like
KDE or Gnome AFAIK is to create a *common* platform for
graphical applications of the kind that OpenStep, or
Windows or MacOS provides, and which the generic X-world
lacks. Generic cooperation between X apps is at a level
MacOS reached in 1985 (ie you can cut and paste text).

To have some apps under Gnome and others under KDE means
that development effort is divided, and you have two sets
of apps which cannot cooperate effectively.

My problem with the Gnome project is that they reasoned that
competing with KDE was better than cloning Qt on the grounds that
the project would be more fun (undoubtedly true), and that they
could correct things they saw as flaws in KDE. They now seem to define
themselves largely as being 'not KDE'. Cloning Qt would have been
a more costructive and cooperative project to undertake.

I feel strongly that this is a situation where we have to hang
together or hang seperately, and this is an idea that doesn't
entirely fit with the free software mindset.

Simon

Mark Jackson

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

: >From a developer sight of view, instead of writing KDE and

: >putting (wasted?) manpower into FreeQt, GTK should evolve
: >more and better. (What about network libs ?)

: I'm afraid this is an example of people bankrupt of ideas. FreeQT


: is an attempt to piggyback on the really hard work of design that
: Troll Tech has done by those incapable of doing anything besides
: copying.

Reminds me of how MS Windows, MS Money, MS Internet Exploder, etc.
got started to *compete* with other products/OSes.

Although in past posts I've favored Qt because of it's maturity, I
can see the drawback to it if it will create a "split".

The major advantage of free SW is that once *anyone* does *anything*,
*everyone* reaps the benefits. Cooperation replaces competition, and
as a result, everyone gains.

On the other hand, since Qt is free for free software uses, plus it
allows the jump into Windoze land, it greatly enhances the value
of Linux to companies. As an example, once the company I work for
currently saw what Qt could do, they are now seriously considering
replacing their Win NT based product line with a Linux based line.
The only reason they are even able to consider such a thing is due
to the existence of Qt.

So, I guess I can see the advantages and drawbacks of both options.

Rajat Datta

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

On 7 Nov 1997 19:48:14 GMT, Mark Jackson <mar...@shell.rmii.com> wrote:
>The major advantage of free SW is that once *anyone* does *anything*,
>*everyone* reaps the benefits. Cooperation replaces competition, and
>as a result, everyone gains.
>

Cooperation replacing competition invariably leads to shoddy products.
Only competing efforts, people driven by their own visions, bring about
really new ideas.

In the context of Linux, new efforts and approaches are to be lauded
and are a sign of health. Otherwise we'll have a pathetic, sad attempt
at a Windows clone.

rajat

jo...@dhh.gt.org

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

Rajat Datta writes:
> Cooperation replacing competition invariably leads to shoddy products.
> Only competing efforts, people driven by their own visions, bring about
> really new ideas.

Cooperation and competition are not mutually exclusive.

John Hasler This posting is in the public domain.
jo...@dhh.gt.org Do with it what you will.
Dancing Horse Hill Make money from it if you can; I don't mind.
Elmwood, Wisconsin Do not send email advertisements to this address.

Conrad Sanderson

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

robert havoc pennington <rhpe...@harper.uchicago.edu> wrote:

>All are sort of small arguments, but taken together it sounds like KDE
>merging into Gnome makes sense.

yes, but i don't see KDE people giving up. And a lot of people like Qt.


There must be some way for GTK and Qt to coexist. Hence the optiion
below:

>> 9. Gnome and KDE people actually work _together_ on how


>> to make programs written for either desktop system
>> work on each others systems (ie. retain integration).
>>
>

>I think they're trying, as much as possible, but k* apps will always
>require Qt, which is the objection in the first place. So they will
>all have to be duplicated with Gtk, or freeQt will have to be written,
>and both of those seem harder than just ditching Qt entirely and going
>to Gnome for all apps.

I didn't mean using different GUI toolkits (we can't win this battle).


I meant More along the lines of making some sort of common API for
drag'n'drop, program restart, localisation, preferences (eg. colors
of menu strips) etc.

What would this involve:

- making a library out of already created Gnome and KDE code
(as much as it's possible).

- the gnome and KDE _base_ libraries are modified to use this
common library

- gnome/KDE programs would still use their respective base libraries
exactly as before (ie. no change).

This way, both KDE and Gnome can co-exist, with programs written
for one or the other working with about the same level of integration
with the desktop.

Looking at all those possible options listed in my previous message,
this is the best way to go. Both KDE and Gnome people have their mind
set on their projects and will not give them up. Making a common
library involves a relatively small amount of work.
Even a very small, minimalist attempt at making ONE function common
(eg. drag'n'drop) is an _extremely_ good start.

>> My recommendation: option 9
>>
>
>I like 2, but I don't expect it to happen. :)
>
>My prediction is that lots of effort will be wasted on KDE and
>eventually Qt will be abandoned, as long as Gnome comes together
>within a year or so. But time will tell.

Qt is like a better Motif. Even if "nobody" supports it, it will linger

tafe...@nospamplease.kde.org

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

Greg Retkowski <gr...@pocketscience.com> writes:
> Here's my thought. a QT wrapper for GTK. Then your KDE and Gnome apps are
> compatable (it all comes back to GTK). You avoid the entire Gnome vs.
> KDE situation, You have a 'FreeQT', you arent re-writing QT from scratch,
> and it causes world peace, harmony and the downfall of microsoft. Can we
> have a group hug ;)

Nice idea, but you should take a look at GTK and Qt header files
before suggesting things that are not practical.

--Stefan

Simon Kinahan

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

David Finton wrote:
> You are right on this one... Qt and Gnome are both projects that have
> long been overdue for the Un*x environment. Most of the previous window
> managers have shown that you can use a GUI through Unix. These projects
> show that you can use Un*x through a GUI. A great improvemant, if
> anything.

...and...

> Depends on what OS you use. In Win95, the desktop *is* the OS. That's
> where Linux and its cousins have a distinct edge. The GUI has a much
> better chance of developing faster if you have alternatives.

I spy a contradiction .....

Simon

Simon Kinahan

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

Conrad Sanderson wrote:
> So let's review our options with regard to this pathetic situation.
>
> 1. Situation stays unresolved. When both systems are complete,
> we basically have a KDE vs Gnome software "war".
> [ "war" = split ]
>
> 7. KDE and Gnome development stays as it is, and then a freeQT
> library appears, released under LGPL. This makes KDE
> completely free, undermining the Gnome development.
> KDE will have the upper hand in the software "war".
> 9. Gnome and KDE people actually work _together_ on how
> to make programs written for either desktop system
> work on each others systems (ie. retain integration).
>
> Future prediction if nothing changes: option 1, eventually followed
> by option 7.
>
> The easiest option to take would be 9 (if egos can be put aside).
>

I'm not at all sure that 7 (the creation of a free Qt) actually solves
anything. It might possibly kill the Gnome project (although there are
niggling other differences in design philosophy), but it doesn't make
use of all the development effort going into Gnome and Gtk apps.

I think the argument over Qt's license is essentially unresolvable. Some
of use find 'nearly free' software an acceptable alternative to truly
free software, some of us don't. It seems to hinge on how important you
think commercial applications are, although paradoxically people who
think they are important tend to favour free development of things like
GUI libraries.

Therefore someone is going to have to give your option 9 some serious
thought. The groups developing KDE and Gnome right now do not seem to be
giving much thought to the issue.

Simon

Stephan Pfab

unread,
Nov 7, 1997, 3:00:00 AM11/7/97
to

Simon Kinahan (sim...@cadence.com) wrote:

[snip]

: My problem with the Gnome project is that they reasoned that


: competing with KDE was better than cloning Qt on the grounds that
: the project would be more fun (undoubtedly true), and that they
: could correct things they saw as flaws in KDE. They now seem to define
: themselves largely as being 'not KDE'. Cloning Qt would have been

^^^^^^^<1>^^^^^^^ ^^^^^^<2>^^^
: a more costructive and cooperative project to undertake.

[snip]

Where did you get this <1> idea from ?
They (gnome/gtk) do a lot of things differently, but not because they want
to be not KDE, but because they want to do it right as they see it.
A quote you will find on the gnome list: "if you don't like it
present a working alternative".
They hatred between gnome and KDE does not come from the coders of both
groups. They actually tend to talk to each other.

Cloning Qt:
You can't be serious. Please have a look at the Qt sources.
The trolls invested an enormous amount of work. A copy
would be : a) hard to write b) no fun at all to write
c) never as good as the original d) would not give the
developer a change to put in some own ideas e) be not
there in the next 1 or 2 years.
I do get the impression that all the people who suggest to
duplicate Qt/KDE never wrote a program longer than a page.
(The motivation for Lesstiff/Wine is a completely different issue.
But ask the programers if they would do it again)

The competition between the two project just will keep
the pace high.

Stephan


Rajat Datta

unread,
Nov 8, 1997, 3:00:00 AM11/8/97
to

On 07 Nov 1997 22:56:12 -0600, jo...@dhh.gt.org <jo...@dhh.gt.org> wrote:
>Rajat Datta writes:
>> Cooperation replacing competition invariably leads to shoddy products.
>> Only competing efforts, people driven by their own visions, bring about
>> really new ideas.
>
>Cooperation and competition are not mutually exclusive.
>

Of course not, but in the context of Linux, there seem to be some who want
to shut down competing efforts of development. The KDE and GNOME project
people are trying to cooperate to the extent possible, while retaining
their own competing efforts at producing a desktop. This is a good thing.

rajat

Brian Knox

unread,
Nov 10, 1997, 3:00:00 AM11/10/97
to

On 7 Nov 1997 23:35:45 GMT, Rajat Datta <ra...@goteborg.netcom.com> wrote:
>On 7 Nov 1997 19:48:14 GMT, Mark Jackson <mar...@shell.rmii.com> wrote:
>>The major advantage of free SW is that once *anyone* does *anything*,
>>*everyone* reaps the benefits. Cooperation replaces competition, and
>>as a result, everyone gains.
>>
>
>Cooperation replacing competition invariably leads to shoddy products.
>Only competing efforts, people driven by their own visions, bring about
>really new ideas.
>
>In the context of Linux, new efforts and approaches are to be lauded
>and are a sign of health. Otherwise we'll have a pathetic, sad attempt
>at a Windows clone.
>
>rajat

Anyone that would claim that Linux is at all related to Windows has
no understanding of the internals of Linux. Linux could never
be a Windows clone with it's current kernel, no matter how similiar
the front end was made to look.

Linux is a completely different beast all together.

Ivan Samuelson

unread,
Nov 11, 1997, 3:00:00 AM11/11/97
to

On 10 Nov 1997 14:14:00 GMT, Brian Knox <bk...@pathfinder.markelcorp.com> wrote:
>>In the context of Linux, new efforts and approaches are to be lauded
>>and are a sign of health. Otherwise we'll have a pathetic, sad attempt
>>at a Windows clone.
>>
>>rajat
>
>Anyone that would claim that Linux is at all related to Windows has
>no understanding of the internals of Linux. Linux could never
>be a Windows clone with it's current kernel, no matter how similiar
>the front end was made to look.
>
>Linux is a completely different beast all together.

And he understands that. He was making the statement that new approaches
for software and development tools should be allowed for Linux. I for one
agree. He in no way stated that Linux is related to Windows. He was NOT
speaking of the internals of Linux compared to Windows. What he was
stating was that he felt that if all we allowed on Linux was software that
MUST be free altogether, then Linux could eventually turn into a pathetic
mish-mash of software that makes it as junky as Windows.

I for one don't completely agree with this view. I've used Linux since
1992 and back then, software for Linux wasn't sparse nor were there
a lot of useful apps like there are today.

For those who dislike Qt and XForms, that's fine. However, some Linux users
I think get too over-zealous stating that Qt and XForms and any apps
developed with them should be boycotted by Linux users.

I welcome these "commercial" or "not totally free" apps. It gives us a more
wide choice/variety of apps to choose from for Linux. As I stated, I started
using Linux when a lot of these apps (Star Office, XForms, Qt, Motif) were
NOT available for Linux. I have all the aforementioned software on my system
and I'm GLAD it's there. It was my choice to get them and use them. Some I
paid for (Motif) others are free for personel use and/or creating apps for free.

Let's just be GLAD that we see such an influx of software for Linux, free and
commercial. I for one am happy. It means that more and more people are beginning
to accept Linux for what it is - a VERY viable platform for business use as
well as personal use.

--
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Ivan Samuelson * Go Colts!
Information Systems Consultant * bol...@netexp.net
Metro Information Services * http://www.netexp.net/~bolski
http://www.MetroIS.com *
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
NOTE: Remove "no.spam." from my e-mail address to reply
DEATH TO ALL SPAM (except in a can)
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Home of the Stoneless Tutorial! http://www.netexp.net/~bolski/tutorial.html
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Chris Waters

unread,
Nov 11, 1997, 3:00:00 AM11/11/97
to

bol...@netexp.net (Ivan Samuelson) writes:

> For those who dislike Qt and XForms, that's fine. However, some Linux users
> I think get too over-zealous stating that Qt and XForms and any apps
> developed with them should be boycotted by Linux users.

Most people, I think, are perfectly happy with commercial apps using
these libraries. It's GPL'd apps using these libraries that cause all
the consteration and fuss, since the license on these libraries is not
compatible with the GPL. Thus, we have the interesting situation
where KDE (for example) violates *its own license*!

Now, of course, in practice, only the copyright holder(s) have the
right to sue over license violations. And it does seem unlikely that
any of the KDE authors would sue a distributor. But! Any
distribution of KDE is AFAICT, illegal, and the distributor *could* be
sued by *any* of the KDE developers.

Personally, I think that RedHat and others who make KDE available at
their sites are taking a pretty questionable risk. And, as for SuSE,
if they're really planning to include it on their CDs, then they're
either insane, or the next best thing.

KDE may be the bees knees, but I ain't going to touch it to find out.

Qt, on the other hand, I'm perfectly willing to use. Just not on
GPL'd projects. (Although, y'know, the idea of making a GPL'd
project, releasing it, then suing all the distributors, just to get
some extra cash coming in....nah. :-) I might release Qt-dependent
software under the BSD license, but personally, I think the BSD
license is incredibly stupid -- if I want to release public domain
software, I'll do it without worrying about some stupid copyright
notice.

Now, there are those who feel that *any* non-free software should be
boycotted. They are extremists, and I don't necessarily agree with
their position, though I sympathize. But *those* people aren't
singling out packages like XForms and Qt.
--
Chris Waters |
cwa...@systems.DHL.COM | pneumonoultra- -osis is too long
xt...@dsp.net (personal) | microscopicsilico- to fit into a single
www.dsp.net/xtifr/ (web) | volcaniconi- standalone haiku.

Vlad 'Borg' Petersen

unread,
Nov 11, 1997, 3:00:00 AM11/11/97
to

Chris Waters wrote:
..........
> > This is _insane_. Whoops! I just got sued by myself because
> > I wrote a GPL program that uses the Solaris OS.
>
> You couldn't sue yourself unless and until you distributed it.

Yes, he did distribute it to himself hence the basis
for a lawsuit. :)

Chris Waters

unread,
Nov 11, 1997, 3:00:00 AM11/11/97
to

con...@hive.me.gu.edu.au (Conrad Sanderson) writes:

> On 11 Nov 1997, Chris Waters <cwa...@systems.dhl.com> wrote:
> >bol...@netexp.net (Ivan Samuelson) writes:

> >> For those who dislike Qt and XForms, that's fine. However, some Linux users
> >> I think get too over-zealous stating that Qt and XForms and any apps
> >> developed with them should be boycotted by Linux users.

> >Most people, I think, are perfectly happy with commercial apps using
> >these libraries. It's GPL'd apps using these libraries that cause all
> >the consteration and fuss, since the license on these libraries is not
> >compatible with the GPL. Thus, we have the interesting situation
> >where KDE (for example) violates *its own license*!

> >Now, of course, in practice, only the copyright holder(s) have the
> >right to sue over license violations. And it does seem unlikely that
> >any of the KDE authors would sue a distributor. But! Any
> >distribution of KDE is AFAICT, illegal, and the distributor *could* be
> >sued by *any* of the KDE developers.

> This is _insane_. Whoops! I just got sued by myself because


> I wrote a GPL program that uses the Solaris OS.

You couldn't sue yourself unless and until you distributed it.

> GPL programs themselves include EXCEPTIONS to the GPL license.
> Have a look at the libc source. Whoops! libc must be illegal!
> better stop using it under Linux... what??? none of my programs work ???

Yes, libc and Motif are granted exceptions under GPL. Qt is not. See
the latest GNU bulletin. http://www.gnu.org/bulletins/bulletins.html

Of course, Troll Tech claims that Qt *is* compatible with the GPL, and
the FSF claims its not. Who is right? The courts would have to
decide. And I'd rather not be involved in a lawsuit, so I plan to err
on the side of discretion, and avoid mixing Qt with GPL'd code. Or
distributing any code which mixes them.

Conrad Sanderson

unread,
Nov 12, 1997, 3:00:00 AM11/12/97
to

On 11 Nov 1997, Chris Waters <cwa...@systems.dhl.com> wrote:
>bol...@netexp.net (Ivan Samuelson) writes:
>
>> For those who dislike Qt and XForms, that's fine. However, some Linux users
>> I think get too over-zealous stating that Qt and XForms and any apps
>> developed with them should be boycotted by Linux users.
>
>Most people, I think, are perfectly happy with commercial apps using
>these libraries. It's GPL'd apps using these libraries that cause all
>the consteration and fuss, since the license on these libraries is not
>compatible with the GPL. Thus, we have the interesting situation
>where KDE (for example) violates *its own license*!
>
>Now, of course, in practice, only the copyright holder(s) have the
>right to sue over license violations. And it does seem unlikely that
>any of the KDE authors would sue a distributor. But! Any
>distribution of KDE is AFAICT, illegal, and the distributor *could* be
>sued by *any* of the KDE developers.

This is _insane_. Whoops! I just got sued by myself because

I wrote a GPL program that uses the Solaris OS. Hmm... it just
happens that it recompiles under Linux. So does that mean my GPL
program is only legal under Linux ??? Some of you guys
are taking the GPL license way too seriously. KDE distribution is not
illegal.

GPL programs themselves include EXCEPTIONS to the GPL license.
Have a look at the libc source. Whoops! libc must be illegal!
better stop using it under Linux... what??? none of my programs work ???

I welcome commercial / non-GPL/ BSD / not-completely-free software
on Linux. GPL is not the be-all and end-all of all things.
Exclusion of all non-GPL stuff (as a lot of people seem to be preaching
here) is just a form of religious fundumentalism.


George

unread,
Nov 12, 1997, 3:00:00 AM11/12/97
to

>I welcome commercial / non-GPL/ BSD / not-completely-free software
>on Linux. GPL is not the be-all and end-all of all things.
>Exclusion of all non-GPL stuff (as a lot of people seem to be preaching
>here) is just a form of religious fundumentalism.

no I'd say people are preaching to keep the "free" distributions free ...
without any commercial software .... redhat wants to say all of teh
tools provided can be used freely for commercial use ... Qt doesn't meet
that ...

if any non-free stuff gets into caldera ... cool ... Caldera isn't a free
distribution so ... if it's a good app .. then great ...

Linux isn't one distribution (thank god)

George

--
------------------------------------------------------------------------------
George Lebl <ji...@5z.com> http://www.5z.com/jirka/
------------------------------------------------------------------------------
WARNING:
Reading this message can affect the dimensionality of your
mind, change the curvature of your spine, cause the growth of hair on
your palms, and make a difference in the outcome of your favorite war.

Richard Moore

unread,
Nov 12, 1997, 3:00:00 AM11/12/97
to

Chris Waters <cwa...@systems.dhl.com> writes:

>
> con...@hive.me.gu.edu.au (Conrad Sanderson) writes:
>
> > GPL programs themselves include EXCEPTIONS to the GPL license.
> > Have a look at the libc source. Whoops! libc must be illegal!
> > better stop using it under Linux... what??? none of my programs work ???
>

> Yes, libc and Motif are granted exceptions under GPL. Qt is not. See
> the latest GNU bulletin. http://www.gnu.org/bulletins/bulletins.html

You seem to think that the GNU bulletin has some legal relevence to the GPL,
it doesn't. Legally the license is the text of the GPL not any specific
interpretation of it. I don't see any mention of Motif etc. in the text of
the license, do you? The GNU bulletin is very hypocritical (in my option)
and RMS has used a different interpretation of how GPL applies in a message
on the KDE mailing lists (at a later date than the bulletin).

As Conrad has said many pieces of GPL code in someway violate the GPL
license, but I seriously doubt anyone will start legal action over it. This
whole discussion is extremely academic and is one I'll not be wasting
anymore time on.

Rich.

>
> Of course, Troll Tech claims that Qt *is* compatible with the GPL, and
> the FSF claims its not. Who is right? The courts would have to
> decide. And I'd rather not be involved in a lawsuit, so I plan to err
> on the side of discretion, and avoid mixing Qt with GPL'd code. Or
> distributing any code which mixes them.
> --
> Chris Waters |
> cwa...@systems.DHL.COM | pneumonoultra- -osis is too long
> xt...@dsp.net (personal) | microscopicsilico- to fit into a single
> www.dsp.net/xtifr/ (web) | volcaniconi- standalone haiku.

--

Reply all
Reply to author
Forward
0 new messages