MAY
27
2006

First Day KDE 4 Multimedia Meeting

In the rainy Netherlands, eighteen KDE hackers have been working in the Annahoeve on Multimedia for the fourth incarnation of KDE. This report outlines the meeting topics, and the results of interesting presentations and explains how KDE developers outbid each others marshmallow records.

The meeting officially started at noon, but the organization and Matthias, Martin, Alexandre, Florian and Gábor were already at the Annahoeve at 9:30. We had things up and running rather quickly and some smalltalk went on. Marshmallows were put on the table and Adriaan told everybody they could earn a marshmallow if they were able to stick 10 pieces into their mouth. Sebastian got close.

The afternoon started with a brief keynote, followed by presentations about Phonon and amaroK. Matthias Kretz talked about the basic design goals of Phonon (see his slides). It should provide the multimedia support that 80% of the applications need in a simple way; the other 20% are simply not its focus. For example, if an application needs precise control at sample level, it is better off talking to a media framework directly. Generally speaking, most applications just need some basic things like playing or streaming video and audio. Matthias further explained how effects worked, and what to consider when building a back-end, and demonstrated some some code examples.

Max Howell talked about moving to amaroK 2 and the Windows port that will come with that. The amaroK hackers are really excited about amaroK 2 and they really want to move on and work on it. A core redesign is needed to prevent regressions from coming up in new releases. This weekend will be used to plan large parts of this redesign.

Max also pointed out some issues with packaging, mostly due to packagers making weird choices when building amaroK packages like depending on MySQL. These things should be solved by making packaging amaroK easier and by more communication with the packagers.

Other topics were Last.fm and Audioscrobbler integration. Last.fm offers more services that amaroK doesn't yet make use of, it can be used to give and share arbitrary labels and other information about songs - the amaroK developers want to add support for this. Last.fm also offers personal radio streams, with for example your favorite music, or music listened to by people with a similar taste. Integration in amaroK will allow you to skip songs in the stream, and Last.fm even learns from that.

Usability has always been an important focus in amaroK, but it's hard to get consensus about certain issues. Still the plan is to do some serious work for amaroK 2 with the help of the two usability experts available at the multimedia meeting. Being able to meet face to face makes usability discussions much easier - as you're less likely to step on someone's toe if you can see them...

After a great dinner, the hacking continued. Lively discussions and silently hacking were intermingled, and the resulting code was committed.

All in all, we came to an important conclusion:
The important thing in Free Software is not the free beer; it is the Free bar which ensures the beer will STAY free.

Comments

I'm heartly and vibing waiting for amaroK 2 =D
Keep up the free bar!


By etherea` at Sat, 2006/05/27 - 5:00am

Yeah, especially since I might want a free beer or two!


By Someone at Sat, 2006/05/27 - 5:00am

Until Amarok is not a part of core kde multimedia package I will not use it.


By Artem S. Tashkinov at Sat, 2006/05/27 - 5:00am

It isn't part of the core kde multimedia package. So cheers.


By Ian Monroe at Sat, 2006/05/27 - 5:00am

amaroK will never be part of the core kde multimedia package. It is in extragear/multimedia because the developers don't want to follow the KDE release cycle.
I don't see any reason why that would be bad. You get bug fixes fast, and the number of new features between releases is not overwhelming you (like in some core KDE apps).


By Firetech at Sat, 2006/05/27 - 5:00am

I wish Kopete had done things that way, and stayed out of kdenetwork. Sometimes things change too fast with the protocols it uses, and full KDE releases don't come often enough.


By jason at Fri, 2006/06/02 - 5:00am

When Amarok becomes a part of the core kde multimedia package I will stop using it (and KDE).
KDE should become a Desktop Environment (hence the name KDE) and not a monster package consisting of a desktop environment and a gazillion of additional applications.

If it's too hard to install third party applications then this problem has to be resolved, but bloating KDE is no solution.


By max at Sat, 2006/05/27 - 5:00am

I totally agree with you. Furthermore, I really hate to do an apt-get install kde and have installed for example kpilot, kstart, kalzium and a hundrer of other apps that I dont want. I am no saying that this apps sucks, they are great, but I dont want them. It would be grate to do an apt-get install kde and have only the desktop environment... then, any other app that one would like could be installed manually... Even all the games are installed, that are not utils o functionality apps... or the education apps... these are apps that most of us dont use...


By Erne at Sat, 2006/05/27 - 5:00am

> It would be grate to do an apt-get install kde and have
> only the desktop environment...

Try Kubuntu. You can install and deinstall every single application via apt. E.g. I kept three games I like and removed the rest without any problems.


By Ascay at Sat, 2006/05/27 - 5:00am

I don't know what distro you are using, but in all likelihood, that "kde" is a metapackage that is meant to do exactly what it is doing: install ALL OF KDE.

You probably can instal lpiecemeal if you want (of course, then you can complain that you must choose between hundredds and hundreds of packages).

And after all... what's the problem with installing all of KDE? What does it cost you? 1GB of disk? So, about .35 euros? If it takes you 10 minutes to sort out what packages you want, you are already losing money.


By Roberto Alsina at Sat, 2006/05/27 - 5:00am

Or maybe you have to figure out how your package manager works, kde is a meta package, so is kdebase, but a much smaller one that just "includes the nucleus of KDE, namely the minimal package set necessary to run KDE as a desktop environment", http://packages.debian.org/unstable/kde/kdebase
http://packages.debian.org/unstable/kde/kde shows what the kde meta package consists of.


By Peppelorum at Sun, 2006/05/28 - 5:00am

Sorry but I do not share your point of view.
Since I am starting to use Linux (SUSE 8.0) I found KDE a very good graphic desktop manager, even the many additionals ( Kword and Co.). I am still using it, fascinating from the development, improvements, functionality and stability reached.
I am not aware if amaroK 2.0 will become a part of the KDE core or not, if the functionality, stability, improvements and eye-candy of all this system will remain better than today, then welcome. By the way I am a happy user of amarok 1.4

I can only say keep this good work, showing all the world that the main improvements for the IT are comming from the open source comuniy.

Juan Ehrenhaus


By Juan Ehrenhaus at Sat, 2006/05/27 - 5:00am

Interesting to see the entity known as superstoned on this forum is a real live KDE contributor! =)


By KDE User at Sat, 2006/05/27 - 5:00am

ehm... well, at your service ;-)


By superstoned at Sat, 2006/05/27 - 5:00am

When will KDE4 be out? The project has been too quiet about this! Is it hard to port KDE3 to Qt4? Why is this taking so long? It will be bad if Qt5 is out before KDE4 is done.


By curious at Sat, 2006/05/27 - 5:00am

The first public preview release (aka alpha) is scheduled for October, 2006. It's been announced on .kde recently.

> Is it hard to port KDE3 to Qt4?

It's not that really hard, the real problem is that KDE4 will bring many new truly exciting technologies meant to ease and improve your desktop experience (not to forget about OpenGL based desktop). Just look through KDE4 goals at kde's wiki: http://wiki.kde.org/KDE+4+Goals

> It will be bad if Qt5 is out before KDE4 is done.

Be sure Qt5 will come out in the near future. I foresee Qt5 will not come before the end of 2006.


By Artem S. Tashkinov at Sat, 2006/05/27 - 5:00am

Any Software engineer will tell you that doing two things at once is a bad idea.

It was suggested that we first port KDE-3.5.x to Qt4 and release KDE-4.0.0.

Unfortunately, those that wanted to do two things at once and make large changes to KDE for KDE-4.0.0 won the argument so that is what we are trying to do.

I still fear that this will cause large problems. However, I hope to be proven wrong.


By James Richard Tyrer at Sat, 2006/05/27 - 5:00am

if you want KDE 3.5 to be binary compattible to the other KDE 3.x releases (and yes, that's what KDE is commited to do) you can't port to Qt 4. And KDE 4.1 needs to be binary compattible to 4.0, if not - developers can't use it and trust their work will be worth anything after a year or 2.... So you can't refactor big things after KDE 4.0.

What you say might be right - but it is not like KDE has a choice - if they want to keep developers on board...


By superstoned at Sat, 2006/05/27 - 5:00am

Perhaps come clarification is needed.

What I, and others, suggest is that KDE-3.5.x be ported to QT-4 thus creating KDE-4.0.0. This could also include fixes for known problems in KDELibs which would break binary compatibility if fixed since this is the only time when they can be fixed. But, the radical new changes should be delayed till the port was completed.

Then after this was up and running, that new features that are added to create KDE-4.1.0.


By James Richard Tyrer at Sun, 2006/05/28 - 5:00am

Dear JRT, what would users gain from having a KDE 4.0 which basically is a KDE 3.5? Not much. Even worse, independant KDE apps would not run due to binary incompatibility.

And if you have a close look at the transitions happening you will find that developing of KDE 4 is indeed done as two steps in series, first porting to Qt 4/KDElibs 4 and then developing new features/frameworks. The only thing that is left out is a release of the ported state, as the value/cost relationship is way too low.

So please stop making uninformed claims (and foremost stop doing so as self claimed engineer, for sake of your fellow engineers' reputation)...


By Else at Sat, 2006/05/27 - 5:00am

even if they port kde3.5 to qt4, kde should be able to take advantage of the optimizations and other new 'under the hood' features of qt4

kde on qt4 may be faster, leaner and consume less memory

so, there are somethings one will get just by porting kde to qt4


By raj at Sun, 2006/05/28 - 5:00am

James, as you point to "Any Software Engineer" - I can point to one or more practicing software engineers who have seen the transition between KDE 1->2 and 2->3, and can vouch for the ability of the KDE team to do two things at once (a jump in Qt and a jump in KDE functionality). Are you going to withdraw your statement now, or add some new qualifications on what sort of software engineers you mean?


By taj at Sun, 2006/05/28 - 5:00am

No, I will not withdraw my opinion because it is my opinion and Free Software is about Free Speech.

There was a disagreement about this and I don't think that others on the side I am supporting have changed their opinions. Time will tell if our concerns are valid or if we were simply being too cautious.

What I am saying is that it isn't a good idea to do two things at once when modifying software (in this case porting to a new tool kit and making major changes to the software), and I hope that there aren't any problems and delays.


By James Richard Tyrer at Sun, 2006/05/28 - 5:00am

Fair enough re: free speech, my primary purpose was to mention that your statement is factually incorrect. You are certainly breaking no laws by continuing to stand by it.

There will almost certainly be problems and delays, and as a Software Engineer that shouldn't unduly cause you concern, as the alternative is worse. The next time the KDE cycle allows for BC breakage, and significant framework redesign always causes BC breakage, it will almost be time for a Qt5-induced BC breakage. If the cycle is changed to allow BC breakage in mid-stream, that will only allow another period of problems and delays.

Sum total period of problems and delays will almost certainly be higher than if it is simply done now, especially as Qt 4 provides so many opportunities to get the redesign done while porting (MVC, for example).


By taj at Mon, 2006/05/29 - 5:00am

Your "any software engineer" argument has been bunked several times before by others and I will do it again. I'm a software engineer and I think that what is going on right now is a good idea. So there.

What your suggestion means is that KDE-4.1 would be incompatible with KDE-4.0 which is a really Bad Idea(tm). Only the linux kernel folks get away from breaking everything between minor releases (I have no idea why).

Do you really think that any developer/distributor would touch a release that is incompatible with the previous release and superceded by an incompatible release just after a few months? Now any software engineer will tell you that THAT is a bad idea. KDE-4.0 would end up being a non-release while everybody is waiting for the "real" release of KDE-4.1 which integrates all the new stuff. Plus it would create a really annoying mess.


By atte at Mon, 2006/05/29 - 5:00am

> Only the linux kernel folks get away from breaking everything between minor releases (I have no idea why).

And the GCC folks.


By Gábor Lehel at Mon, 2006/05/29 - 5:00am

I fear for KDE4 because now, some important parts of it are just vaporware.
Look at plasma, there isn't at this point a plan for the new interface in KDE4, just a wish to change the desktop for a mor emodern concept.
I went to a presentation in FISL 7 where Aaron Sigo showed just a icon the had menus when you put the mouse over it with options like open, delete, etc. Besides the concept being failed in my option (I belive it should open a menu for it, not over the icon) is is so few concepts until now.
Probally the interface of plasma will require a lot of coding, but I don't see plasma being done soon, so when october arrives we'll probally see only a kwin running without kicker.

My point? KDE probally will be out only on second half of 2007, I belive by the end of the year, so don't keep holding your breath until then.
PS: There is nothing wrong with it, maybe with a large extra time, developers can fix long standing problems on popular KDE problems (kopete could use a lot of help and speed improved) or start adding new and cool features.

I have faith KDE4 will be a great thing, but I don't have faith it will arrive soon enought.


By Iuri Fiedoruk at Sun, 2006/05/28 - 5:00am

> Usability has always been an important focus in amaroK, but it's
> hard to get consensus about certain issues.

Please bury these custom icons, tell your artists to complete the standard icon sets and keep consistency at a high level.


By Anonymous Coward at Sat, 2006/05/27 - 5:00am

Those custom Icons are cool.
If your brain is overstrained by those good looking icons you can always switch to the default icons.

But please don't try to deprive other people of those nice icons.
They make amarok look much better and yes - in my opinion they even improve usability. Usability does *not* mean that every program has to look as boring as kedit.


By max at Sat, 2006/05/27 - 5:00am

let's fight about this, max... i know where you are now... :D


By superstoned at Sat, 2006/05/27 - 5:00am

Speaking of custom icons, when i installed 1.4 beta, the custom icons were default; and when i installed 1.4 final, the kde icons were default.
Did the amaroK team change this, or should i 'blame' the packager?


By rinse at Sun, 2006/05/28 - 5:00am

Changed temporarily because the theme is still incomplete. We hope to get it completed for 1.4.1.


By Mark Kretschmann at Sun, 2006/05/28 - 5:00am

Perhaps the icons are "way cool" but to be a KDE application the developers should (also) supply a set of icons in the current default style (CrystalSVG for KDE 3.x.y) and a set of HiColor (generic|unthemed)icons.

The current KDE icon theme should be used by default if running with the KDE desktop.

Other than the above, I see nothing wrong with also providing custom icons which the user can use if they want.


By James Richard Tyrer at Sat, 2006/05/27 - 5:00am

> The current KDE icon theme should be used by default if running with the KDE desktop.

++


By Carlo at Sun, 2006/05/28 - 5:00am

A clarification on the subject of custom icon themes and programmers.

We have been trying for the last six or seven months to find someone who are able to create these icons for us, please do! The squad are all coders (or in my case someone who is unable artistically to create CrystalSVG style icons because my own style is so different from that), and have really tried hard to find artists.

The real problem is that this is no small task - the current (and incomplete i might add) list of icons we're missing can be found on our wiki: http://amarok.kde.org/amarokwiki/index.php/Artist_team:Icons

At K3M, however, Thomas Zander and i managed to finally reach an understanding on the subject - He has offered to function as an intermediate between the amaroK team and the KDE Artists team. Missing currently is for the icon list to be completed (with good descriptions for each icon, so that the artists can actually figure out what the icons are to be used for) - i have stepped forward to create (or rather finalise) this list, and hope to be able to do this within the week.


By Dan Leinir Turt... at Tue, 2006/05/30 - 5:00am

I'd like to know how are KDE 4 multimedia apps going to deal with restricted media formats. The situation right now as I see it is the following:

- If you haven't got installed the correct media codec, then when you play the multimedia file say in amarok or kaffeine, it fails and the media player tries to go to the next file.
- If you have a list of for example mp3 files, then amarok just keeps showing 2 or 3 OSD per second, and consuming quite a lot of cpu.
- In order to install the correct media codec, you need to take care of your self. Depending on your distro, there's always multiple different ways to solve the problem, but something certain is that it will be a troublesome experience.

We cannot fix the problem by just saying "that's a problem that each distribution should overtake by itself". Instead this is the solution I suggest:

Just copy what Firefox do with multimedia plugins like Flash! When a file couldn't be opened, give a OSD notice to the user saying something similar to "This song is in MP3 format, but it couldn't be played. Please CLICK HERE install support for MP3."

After that, a popup window shows up, which by using (for example) khotnewstuff downloads a list of the apps available for each distro that resolve the dependency attached with a short description and whatever is needed to each one. For example if there's one MP3 codec which probably infringes some USA patents, warning the user "If you live in USA, please don't install this codec if you don't want to have codecs, use the Linspire codec instead". Also you can give a short description of why the codec was not installed "MP3 is a restricted media format and that's why Ubuntu Dapper Linux does not come with support for it".

If installation ends successfully, ideally the MP3 will play without neededing to restart amarok (just like after installing flash plugin in firefox ;-).

This "multimedia support installer" could be generally available to multiple application: konqueror (for i.e. flash support), kaffeine, amarok..


By Eduardo Robles ... at Sat, 2006/05/27 - 5:00am

Perhaps we need an API for this task :

- define interfaces to add new software, and let the distributors implement them. After that all Apps will be able to install software a defined way.

Its like Phonon , define the API , and then implement backends

the interface should :
- install by keyword
- install deps
- and so on


By ch at Sat, 2006/05/27 - 5:00am

Perhaps we need an API for this task :

- define interfaces to add new software, and let the distributors implement them. After that all Apps will be able to install software a defined way.

Its like Phonon , define the API , and then implement backends

the interface should :
- install by keyword
- install deps
- and so on


By ch at Sat, 2006/05/27 - 5:00am

read the report this evening/night from today :D


By superstoned at Sat, 2006/05/27 - 5:00am

Repeat with me:
security, security, security.
/home mounted with noexec.
security, security, security.

The Only True Way (tm) of installing codecs (AND browser plugins!!) is thru your distro's package manager! Anything else is a threat! Seriously.

So, if you use Dapper and you try to play an MP3, just show the following message:

For Dapper to play MP3s, you should go to Adept, enable multiverse, and install the package gstreamer0.10-plugins-ugly-multiverse (possibly with links for opening adept, or dcop instructions to do so -- requiring the user to give his sudoing password, of course.)

HTH


By Humberto Massa at Sat, 2006/05/27 - 5:00am

Yeah, joe blow users that just want to play some music will really love that. Besides, everyone knows that the "one true way" is only to install things you've personally reviewed the source of and compiled yourself. Everything else forces reliance on untrusted 3rd parties. Besides, you forgot that /var, /tmp, and /root (at a minimum) must also be mounted noexec.


By bah at Sun, 2006/05/28 - 5:00am

Banshee, the next best iTunes clone for Linux. Yay!


By Max Howell at Sat, 2006/05/27 - 5:00am

amaroK as a full-blown audio player is certainly superior to banshee, however banshee still has great daap support which is something that amaroK still lacks. Really nice to see that there's a Google SoC project on this though; hoping it will all go through. :)


By apokryphos at Sat, 2006/05/27 - 5:00am

Maybe, but AmaroK is even better than iTunes :-)
But thanks god, Amarok runs perfectly under Gnome, too.


By max at Sat, 2006/05/27 - 5:00am

1. Podcasting is better under iTunes.

2. Indeed, who cares about toolkits and libraries.


By And at Sat, 2006/05/27 - 5:00am

Um, if you could please explain why and how itunes is superior with Podcasting, i would be pleased to improve this support. Have you used 1.4? we have increased the usability and feature set of podcasting dramatically.


By sebr at Sat, 2006/05/27 - 5:00am

That's what we all want. Really.


By Max Howell at Sat, 2006/05/27 - 5:00am

Banshee devs are sell-outs to the Free Software cause. By releasing Banshee under the extremely weak MIT X11 License, they are welcoming DRM-crippled plugins so that the entertainment cartel can handcuff users and rip them off.

No thanks!


By AC at Sat, 2006/05/27 - 5:00am

Who cares, everybody can still use Amarok or Songbird or lTunes.

But indeed, it is a weakness and a risk of Banshee.


By And at Sat, 2006/05/27 - 5:00am

Pages