KDE-CVS-Digest for March 19, 2004

In this week's KDE CVS-Digest:
KDE integrates Mono with C# bindings.
A PHP debugger integrated into Quanta.
Work continues on eGroupware / Kontact integration.
Kopete rewrites the Jabber plugin.
Plus, a new tool for monitoring application usage.

Dot Categories: 

Comments

by ASDF (not verified)

One problem I found with the window manager settings is that I tell it to open windows centered yet for non-KDE applications this setting doesn't apply. Is this a bug that can be fixed?

by Lubos Lunak (not verified)

Possibly, if you complain about it at the right place (hint: http://bugs.kde.org) with the right amount of information (hint: which applications).

by angry_mike (not verified)

Good to see KRecipes is progressing nicely. I am using the old version right now and it is excellent but has so much potential to awesome. Great work on the program and keep it up guys!

by Erik Pedersen (not verified)

It really belong s in kdeextragear, not in kdenonbeta

by Anonymous (not verified)

Tell this its author. Perhaps he still considers as unstable considerung the 0.4.1 version number.

by uga (not verified)

Hehehe, thanks. 0.5 will be our first Beta, really :) All versions up to 0.4.1 have been accordingly tagged as alphas.

We're trying to make into the app. the most important features for this new version, and after that begin making it more stable, improve code, useability, accessibility... there's a long way before 1.0 :)))

by ac (not verified)

There have been news about some projects planning/implementing NMM support.
(NMM: http://www.networkmultimedia.org)

Although NMM is not a kde-related project right now, it could be very
useful for the Linux desktop.

That's why I think, if kde plans to use NMM, kde should, for the success of
the Linux desktop, encourage NMM to become hosted on fd.o

No kde-only networkmultimedia please!

And while this is done, please get rid of atrs in the same time.

by Spy Hunter (not verified)

GStreamer is already hosted at freedesktop.org. I think GStreamer is a better project. I'm not sure what rules fd.o has about hosting similar projects, but you may have a hard time convincing them to host NMM when GStreamer is already there. Also, I don't want a "KDE uses NMM and Gnome uses GStreamer" situation to happen. It's already decided that Gnome will use GStreamer, so if KDE decides to use NMM, NMM better be a WHOLE LOT better than GStreamer to justify the incompatibility.

by Oliver Bausinger (not verified)

The fd.o guys don't have a problem with hosting different projects.
But the biggest problem with NMM is that it is not a real community project
but more of a research project.

by Nick Shafff (not verified)

and research is very bad?

another idea - incorporate interesting things (algorithms?) of NMM into GStreamer

by anon (not verified)

> and research is very bad?

No, it's just that:

- it may or may not be evolved into a community project, rather than a research project founded by a university
- it may not be atuned for desktop usage as much as a non-research project might

by Datschge (not verified)

Did you already take a look at it? My impression of NMM not really being "attuned for desktop usage" is imo actually a huge feature considering that it already allows you to share and transport multimedia data between different machines in a network, eg. manage all music and video data on your computer and stream it to your stereo/TV with respective adapted interfaces, the ideal way I'd want to setup TV/music set-top boxes in a home network.

As for the worry that a research project might not evolve into a community project you must be kidding considering the many FOSS projects which directly and indirectly started as research projects.

by anon (not verified)

> Did you already take a look at it? My impression of NMM not really being "attuned for desktop usage" is imo actually a huge feature considering that it already allows you to share and transport multimedia data between different machines in a network, eg. manage all music and video data on your computer and stream it to your stereo/TV with respective adapted interfaces, the ideal way I'd want to setup TV/music set-top boxes in a home network.

That's very cool, but still not a good example for desktop usage (not saying that NMM isn't, just saying that example isn't) :)

> As for the worry that a research project might not evolve into a community projet y
ou must be kidding considering the many FOSS projects which directly and indirectly started as research projects.

Of course, that's why I said "may or may not"... for every research project that morphs into a community project, there are two that don't.

by David (not verified)

In the potential for desktops to do a huge amount more, and reach out to many different devices, I'd say it is very significant for the desktop. Look at the popularity of MythTV.

by David Johnson (not verified)

The software on Freedeskto.org should be research/reference software. It is a site to promote desktop interoperability, and not to be another sourceforge. Xserver is there because Xserver is research. Just because there are a lot of people trying to make it another sourceforge doesn't mean we have to participate in the destruction of a good idea.

by Stephen Douglas (not verified)

Surely it should be the other way around? Just because GStreamer is used in Gnome doesn't mean it's ideal for us, the choice of multimedia backend should stand on its own merits.

by Tom Chance (not verified)

Not at all, and for two reasons:

1) From the users point of view, one single multimedia framework being shared by all desktop applications means less packages to install and manage, and less multimedia frameworks to configure

2) From the programmer's point of view, it means more work; if KDE goes for NMM or MAS, it will also have to support Gstreamer so those who use Gstreamer for most of their other multimedia applications don't have to manage two underlying frameworks

arts is such a mess not only because it's buggy, but because it's yet another thing to configure on top of the kernel modules and (possibly) ALSA. If we go and adopt yet another new standard, apart from freedesktop.org, we make things yet more hellish for programmers and users alike.

by Nick (not verified)

If KDE starts using GStreamer I will probably switch to Gnome. Whats the point of KDE if it only wraps up Gnome technology? Just another layer that introduces bugs and has to play catch up with the real thing.

by anon (not verified)

gstreamer != gnome technology

PS: we'd gladly have you switch to gnome, I sure don't want people like you in the kde community.

by Nick (not verified)

gstreamer == gnome technology

PS: The kde developers seem to have lost the power to create. Instead they write lame wrappers around ugly C-libraries. Is this KDE spirit? if yes, then I am out, probably with many others.

by Max Howell (not verified)

GStreamer is used by Gnome, but it is not a Gnome technology. If it was specific to Gnome would KDE even be able to consider it?

by anon (not verified)

Creating a multimedia framework from scratch would take at least three years to stablize and more than probably 50k (more realistically, 100k+) LOC.. are you willing to code such a beast?

by Ryan (not verified)

Ah, good old Not Invented Here Syndrome. Killed many a company that spent years trying to re-implement something they could have had for free if they had just swallowed their silly pride and joined a common cause.

It seems like your comment is based on nothing but immature hubris and some misguided asthetic about the ugliness of C.

The part you don't seem to get is that, _if_NMM_becomes_a_mature_project_, and it is really superior enough to justify the cost of changing everything over, then KDE *and* GNOME will both switch to NMM. Its pointless to fracture the desktop community (that extends beyond *your* pet desktop, and the sandbox politics that fringe users like yourself seem to develop) for some perceived benefit of a product that isn't even useful yet.

When NMM ships code that blows the socks off GStreamer, *then* its time to consider switching.

Cheers,
Ryan

by Allan S. (not verified)

No GNOME wouldnt change to NMM, since it is writen in good KDE-like C++. To much politics in the GNOME camp to accept that (did they change to aRts back when it was superior or use DCOP?).

by Spy Hunter (not verified)

aRts has never been superior to anything, and it's no surprise GNOME never adopted it. As for DCOP, by the time it was clear DCOP was a success, GNOME already had too much infrastructure built up around CORBA and Bonobo to switch.

by anon (not verified)

arts was way superior to esd back in 2000 :)

by Nick (not verified)

I think Stefan Westerfeld tried to reach a hand out to Gnome by using glib in arts. It seemed however as if the Gnome people chopped it off. What do you learn from it?

No, arts used more and more glib in order to become part of CSL.. CSL was never implemented fully before people saw it's design flaws.

NMM and gstreamer both have much better modular architectures.

by Nick (not verified)

Ok, so why not base the whole desktop on wine or lesstif? Come on just swallow your pride!

It seems to me like you never have had to deal with ugly legacy C code! It is a huge pain to work with and it is certainly not what I want to develop in my free time!

I like(d) KDE because it broke with the old shit, because it is(was) something new and fascinating. That is the motivation of most developers I know from experience.

btw, you can always go to #gstreamer and experience some love for KDE from the gstreamer developers.

> Ok, so why not base the whole desktop on wine or lesstif? Come on just swallow your pride!

We already depend on many C-based libs, when it makes not to duplicate their functionality. Do you really want to drop usage of libjpeg, libgif, libpng, cups, libsmb, libxml2, libxslt, openssl, mpeglib, libtiff, libmng, freetype, libogg/libvorbis, imlib, alsa, libsane, libgphoto, libdb, xlib/xrender/xvideo, zlib, libmal, libldap, libpcre, libidn, libart, libaudiofile, xinelib, and probably more C-based libraries.

> It seems to me like you never have had to deal with ugly legacy C code! It is a huge pain to work with and it is certainly not what I want to develop in my free time!

Are you even developing a multimedia app in KDE, or are you just a troll? I don't know any "nick" in kde-multimedia.

>> btw, you can always go to #gstreamer and experience some love
>> for KDE from the gstreamer developers.

--

> Are you even developing a multimedia app in KDE, or are you just a troll?
> I don't know any "nick" in kde-multimedia.

I'd assume the latter since the GStreamer developers stated on
several occasions (some of them more than a year ago(?) when
the topic of choosing a MM framework for KDE 4 came up)
that they want to work with the KDE team, too.

by Nick (not verified)

They also hate it and stated on several occasions that they want it dead in the water.

You can surely give a proof for your accusations,
or I'll call you a would-be troll...

To throw in some info myself:

About GStreamer commitment to compatibility
through KDE release cycles (on IRC one year ago):
http://lists.kde.org/?l=kde-multimedia&m=104614380715889&w=2

General: Here's a pointer to the discussions one year ago:
http://lists.kde.org/?l=kde-multimedia&m=104633304511264&w=2

by Nick (not verified)

"Are you even developing a multimedia app in KDE, or are you just a troll? I don't know any "nick" in kde-multimedia."

No I don't develop KDE-MM stuff, Its people like you that scare developers away.

And about the libraries you mentioned, they are pretty old. They are well known through insecurity news though, which shows their quality.

You certainly wouldn't start a project in C anymore these days, but thats exactly what gstreamer is. We programmers call that LBR software (legacy before release). I don't think KDE should sacrifice sanity for the sake of being compatible with the overhyped gstreamer project that favours the technology of gnome that was founded and exists to destroy KDE (Ximian people are on the gnome board of directors, mind you). There is no real value in gstreamer. The plugins come from other sources anyway and 95% of the users dont even want a framework like arts, nmm or gstreamer, because they just want to listen music or watch a dvd.

by David (not verified)

Funny. Try telling that to the Gnome people then.

"When NMM ships code that blows the socks off GStreamer, *then* its time to consider switching."

It already does, and it seems to be pretty well structured. It would be silly not to use it for the sake of what some people think is unity.

by Rayiner Hashem (not verified)

You are welcome to do that, but it must be noted that GStreamer isn't a GNOME technology. It is written in C, but if we're so obsessive about getting rid of every trace of C code on the system, we might as well fork the kernel and write "KLinux"

by Maynard (not verified)

Amen Amen I say to thee.

I use GNOME and have k3b installed on my system. Although I do not use it much. I figured until space was a concern, it doesn't matter that I have the whole of kdebase installed too.

I support GStreamer because it is very forward thinking. It may actually be a good thing that GStreamer is done in C rather than C++. It not like there are no bindings for C++. There are some in teh KDE CVS already anyway.

The biggest complaint I heard was that about glib. well, I do not know if installing glib is a bad thing in itself too.

by anon (not verified)

> The biggest complaint I heard was that about glib. well, I do not know if installing glib is a bad thing in itself too.

That's not a problem, since arts uses it too. However, I think the MM folks prefer something that is more hackable. One of the reasons arts wasn't well maintained once Stefan Westerfeld became too busy with real life was that it's code was full glib-isms that nobody wanted to learn :)

by Maynard (not verified)

Well, it is a shame really. I suppose in a way that is the difference between free software and proprietary. If someone won't learn to use a library because it full of glib-sims, then it really ca't be helped.

Gstreamer has a lot goin for it. Besides being quite portable (written in C), it is very light. It was successfully used in embedded applications. It is really futuristic and useful NOW, not later. NMM will just be an extra library for codec makers to support, or devlopers to develop for. Not to say a decision has already been made which to support by third parties,

by David (not verified)

"I use GNOME and have k3b installed on my system. Although I do not use it much. I figured until space was a concern, it doesn't matter that I have the whole of kdebase installed too."

Try looking at the space equivalent software would take up on Windows. Having Gnome and KDE installed is small in comparison. Makes you wonder what it all does....

by Nilesh Bansal (not verified)

Whats wrong in using GNOME technology. Infact this is what OSS is all about, that no one needs to redo already done work. If their technology is good we must use it.

GNOME and KDE need to go together. Multimedia server should be gnome and kde independent project and both desktops need to accept same standards.

Linux desktop should not be some GNOME vs KDE fight but 2 desktops co-existing having some common standards so that users can switch between the two easily.

There is nothing wrong with it, in fact KDE uses many libraries that originate in GNOME, such as libxml2, libart, etc. gtkhtml in turn was a port of kde 1.x's khtmlw and is used in GNOME. I'm sad that there isn't as much reuse of KDE-originated code in GNOME as the reverse as the GNOME devs still don't seem to want C++ :(

by David (not verified)

The choice of multimedia framework has to be good, and has to have potential to be extended far further in the future, as Linux and freedesktops get ported to many different devices and environments. NMM looks extremely promising in this regard. You can't compromise on quality for the sake of 'one this' and 'one that'. The question is, would everyone else be willing to compromise for this?

"if KDE goes for NMM or MAS, it will also have to support Gstreamer so those who use Gstreamer for most of their other multimedia applications don't have to manage two underlying frameworks"

Why? Would Gnome or any other environment support NMM? Stop making it sound as if GStreamer is the only option, because it isn't.

"If we go and adopt yet another new standard, apart from freedesktop.org, we make things yet more hellish for programmers and users alike."

Excuse me? The last I looked GStreamer was most certainly not promoted as a standard on freedesktop.

by anon (not verified)

I think gstreamer would already been accepted as the choice for KDE 4 if it weren't for the fact it extenstively used glib and gobject. There is nothing wrong with those two libs (no political problems), but technically, there might be. arts is also based on glib, and there were only a few people willing to hack it between KDE 2.0 and KDE 3.2, hence it's buggyness.

On the other hand, the gstreamer development team seems _very_ supportive and eager to fix any KDE related bugs in gstreamer.

by Coma (not verified)

arts is also based on glib, and there were only a few people willing to hack it between KDE 2.0 and KDE 3.2, hence it's buggyness.

Good thing is GSTreamer is already adopted by Gnome, and hosted on freedesktop. That means that even if KDE developers don't take much time coding on it, the project will still be supported by other projects.

by Allan S. (not verified)

What does that have to do with anything? GStreamer's inclusion on freedesktop.org was _exactly_ not an endorse, although a deliberate attempt by the GNOME camp to spread general confusion and misinformation.

by Christian Schaller (not verified)

As one of the people actually involved in the move of GStreamer to fd.o let me explain a couple of things.

a) being hosted on freedesktop is not an endorsement or a proof of standarization. If you read the guidelines for fd.o hosting you see that any project which wants to be a cross desktop technology can be hosted there. Just like there are many projects in GNOME CVS and in KDE CVS that are not 'official' part of those desktop, so will there be in freedesktop CVS not (still) officially standards. You will notice two X implementations hosted there atm for instance.

b) It was partly a technical decision since we where getting fed up with sourceforge having so many issues

c) Sending out a clear signal that we are serious about crossdesktopness was an important part of the decision. We never did anything IMHO opinion to spread confusion and misinformation. In fact if you read my recent GStreamer article on OSNEWS you would see that I clearly explain the fd.o move with wanting to send out such a signal in the hope of increasing the chance of KDE inclusion.
d) Nowhere have any GStreamer developer claimed that being on fd.o is the standard. Only place I ever seen that argument is in fact on kde sites and lists, and the normally by people who are not involved with GStreamer but who use/develop KDE and want or don't want gstreamer in KDE. So I think you should look at bit closer to home before pointing fingers and casting blame.

can't kde have some kind of KAudioObject which all kde-apps use, which in turn just uses the user-selected audio system (nmm, gstreamer, jack, alsa, oss, arts, esd etc.)?? Would that introduce much overhead/latency?

Does this question disclose that I have no clue? :)

by anon (not verified)

the problem is that some of those, like gstreamer already have that mechanism t o plug into different sound servers.