[Lxde-list] Merge teams for the best. (was: LXDE FORK, was: PCManFM Qt 0.1.0 released)

[Andrej N. G. <an...@re...>]

    Hello!

    I reply to two letters in our mailing lists because they are linked
by the theme but subjects of them weren't so clear on theme.

Alexis López Zubieta has written on Thursday, 25 April, at 12:01:
>I would like to clarify that the reason of this thread is not to fork 
>LXDE or Razor-qt instead we are inviting you and the Razor-qt community 
>to join us in order to make a better product. Both of you are making a 
>lightweight desktop environment (now on LW DE) for those that need it.

>I live in Cuba a third world country, as such we can't buy every year 
>the last computer model for our schools, business and homes so we have 
>to use what we have at maximum, thats why we (the Nova Project) are 
>interested in developing a LW DE. We have many users with Pentium III at 
>700 MHz with 128 MB of RAM those are our stakeholders (the people who is 
>going to use our product). But we have another limitation most of you 
>have an advanced knowledge of computing, those people has none, so all 
>the basic functionalities must be a simple thing to do, we can't ask to 
>such user to open a terminal. As you can see the task that we have isn't 
>simple at all. And this situation is common in every third world 
>country, those who has the major part of the world population. So what 
>we do is important for a lot of people.

>That's why we are so interested in join forces with you and with 
>Razor-Qt. I don't pretend that you to recode everything, my intention is 
>to call you guys realize that you are no making an experimental project 
>or a toy you are making critical systems (for the major part of the 
>world). Thats why I invite you to review what you have done and what you 
>can do now on, you also should think in what we can do working spited 
>and what we can achieve together.

>Andrej N. Gritsenko, Julien Lavergne, PCMan and the rest that think in 
>the integration of the different communities as an opportunity please 
>step in and lets make a formal proposal to the communities, to state our 
>points in a more compressible way. Can I count on you?

    I believe our goals are similar. At least what I always wanted from
DE? It should be:

1) lightweight: it should not be slow in any way on netbooks for example
2) easy to use: I should do anything with just few keypresses (or mouse
clicks for those who loves to hug their rats:)
3) comfortable: it should have some default settings to be nice for new
users without terminal tricks and in the same time let me change every
element of my desktop system if I am advanced user
4) modular: I should have the possibility to construct my desktop from
some elements if I want that

Also I know that accessibility is really thing which is required and it
even more important than any bells and whistles because those people who
need accessibility are more dependent on those things than we are.

You stated above what are computers your people have so let it be the
requirements which we have to support, i.e. LW DE should work fine on
them. I believe that is possible - my main desktop was 450 MHz Celeron
with 128 MB RAM just 4 years ago and it worked fine with KDE 3.5. It
started non momentally of course but I was able to watch movies, surf
internet with Firefox, edit documents, etc. I don't see any reason that
shouldn't be possible today.

And in other letter, Petr Vaněk has written on Thursday, 25 April, at 16:11:
>hi all,
>I'm one of Razor committers.
>
>On 4/24/13 10:59 PM, Andrej N. Gritsenko wrote:
>> This would be just wonderful to have you joining forces. I've started the thread with idea to join
>forces with razor-qt team. In case someone doesn't know, I'm the main
>developer of libfm and pcmanfm at the time being, mainly because PCMan
>has a lot less time and I have much more and I have interest in making
>it. Yes, I'm working with GTK but last time we found out that making it
>compatible with such fast changing GTK3 doesn't worth all the efforts.
>Therefore Qt looks like as viable alternative for another toolkit
>instead of GTK3 and PCMan started his experiments with Qt and it is what
>libfm-qt / pcmanfm-qt are. Yes, I can handle all bugs and feature
>requests for libfm / pcmanfm alone, but what with another LXDE
>components? LXDE team has very few developers. But since razor-qt has
>not too many developers too and they have the same goals (not build
>monstrous integrated complex but rather a desktop toolkit) I think it
>would be beneficial for both teams to join forces. And since you want
>something alike, that sound very promicing. But we should get razor-qt
>voices first I believe. From LXDE side - I and PCMan support the idea,
>some other developers aren't sure yet. Andriy.

>The idea is great for sure. For me, personally, is the Qt way to go
>because of its "ease of development" style for applications. I can
>compare it briefly myself as my patches probably live in Evolution (Gtk
>mail client) and it was quite hard core school of GUI programming ;)
>Since then I live with Qt as mi choice.

    I still write libfm, libfm-gtk, and pcmanfm with glib and gtk base. I
given up the GTK3 support due to reasons I mentioned before - there is
still support for GTK3 up to 3.4 but I hardly will even try to adapt it
to newer versions, even if users will ask me. But I belive GTK2 support
is something we still have to have. At least because some old systems may
(and will) work easier with gtk2. And also because some users may don't
wish to install Qt just because they use exclusively Gtk applications
otherwise.

>I'd like to hear your ideas of cooperation/integration for sure.

    As I already said, together people may do much more than splitted.
Since we decided to give up GTK3 support, we want to have lightweight DE
based on two toolkits - GTK2 for old or small systems (Debian Squeeze on
Celeron II, Raspberry Pi, etc., etc.) and Qt for more modern systems to
allow more flexibility. I do not ask Razor people to start develop GTK
applications and don't ask LXDE people to start develop Qt applications
but everyone can help others such way GTK version will get features it
misses now and Qt version will get features it misses. Everyone in both
camps have own zone of knowledge and working together may make it better.
Of course, everyone will more or less get new knowledge but I don't see
anything bad in that.

>I'm little bit afraid of one thing in potential merging. I think it
>would be more "philosophical" clash of users than real usage affects.
>Current LXDE users might raise their voice with: "down with Qt, we want
>Gtk" because they don't understand (and they don't need to understand of
>course) the easiness of development etc. It can be quite hazard for name
>of LXDE.

    It's what I want - just don't give up Gtk, at least until it will be
abandoned by their creators and major distros. I.e. we have LXDE which is
GTK2 based and Razor-Qt which is Qt based, but both have near the same
number of similar feature components (they are different for now but will
be closer and closer with time). Only developers will know those DE are
close one to other, users will see Gtk one and Qt one so no concerns are
raised.

>On the other side I don't see any point against the move.

>Here are some unordered thoughts about Qt world of Razor:

>- Currently we are using Qt4 with initial preparation for switch to
>Qt5 (which would be quite easy).
>- With Qt5 the Qt libraries are even more modularized tan in Qt4.
>- KDE guys are working on so called "frameworks" = KDE technologies
>and libraries split to *independent* modules mostly (where possible)
>integrated into Qt itself.
>    - the independency should be real, meaning - no cross dependencies
>between frameworks and also no global packages in distributions
>    - which means we could use eg. Solid (hardware info framework -
>batteries, automounting, ...) in Razor instead our own backends because
>Solid is much more tested and it works on almost all platforms while we
>are stuck in udisk/udisk2 only.

    Many Qt applications still use GIO so why not use gvfs then? And some
things may be done via libfm which also uses gio/gvfs. When I prepared
release 1.0.0 I get rid of every memory leak so it should be memory safe
for now, valgrind finds no problems so far.

    Anyway, I'm not going to learn more Qt in nearest future because I
have to bring libfm and pcmanfm to 1.2-beta state (there are huge number
of features to implement for 1.2 series, I want to clear FR tracker as
much as possible).

    Cheers!
    Andriy.