Monthly Archives: January 2006

Something new in KDE 4: Phonon

There some things which are in the focus of the news, and some which are not – one of these things which hasn’t been in the focus for a quite long time is KDEMM. KDEMM ought to be the new multimedia api, and should, together with a backend, replace aRts.

Some days ago it was announced thta KDE now gets a new, better name: Phonon.
That’s nice because it shows that the developer really think about how they want to sell their program. And it shows that we are getting rid of the “K” in the most application names :-)

So what is Phonon now? As I wrote, it should replace the actual system in a way – that’s quite nice because although aRts was bleeding edge when it was released there hasn’t been any real development in the last years, and the project is now officially dead. In the meanwhile the rest of the world didn’t stop development, and the requirements for multimedia on the desktop developed and enhanced, too.

To give you an idea:

  • In former times there was a strong need for a soundserver to mix different audio streams from different sources – now, with Alsa since kernel 2.6, this need is not as strong as in former times because the most computers which are running KDE are running a Linux with 2.6 kernel (only the BSDs and older Linux have no Alsa).
  • In former times we had only some audio sources and normally one sound card. Today we have probably more hardware soundcards, several audio streams and these streams in and out (music + calling in from kopete/jingle, for example).
  • Today we have video also – if we really want to make a multimedia api we have to deal with Video too – and then we have to deal and manage these different streams, too, probably guiding them through different tunnels to other devices.

And there were other problems with aRts, too. Who ever tried to wrap skype over a normal built-in soundcard knows what I am talking about.

So why don’t just switch to gstreamer, you may ask? Well, the main reason is: it is hard for kde-developers which are used to Qt object model, to develop something with a glib api. If you want to attract developers, give them easy-to-use apis. Sure, you could write a gstreamer-only api to give the developers a qt-gstreamer api, but then you could rise other questions: Is gstreamer really the highest which can ever be reached? What is if there is something new, and gstremaer is not able to follow? Or what is if gstreamer does not want to work for some machines well enough, and no one wants to fix that soon because there is something more important for them? And aren’t their other projects which are very promising? Have a look at NMM for example. Why should KDE fix to one specific backend, banish all other interesting projects and get back into subjection where it wants to escape, when you in any case have to develop an api? KDE should use, and KDE does use this opportunity! Thanks a lot to vir.

If you still don’t believe me, check this discussion.

And for me? Well, I’m no developer, but I can try to convince the people what the “right way” is – and I can try to keep asking if everything really gets fully integrated. Just think of a full integrated, totally jabber-video and jabber-audio capable Kopete working together with kontact which has all your phone numbers (synchronized from your mobile phone, of course). The music comes from amarok, the video from something-yet-to-develop (or an advanced and integrated kaffeine), and everything – and that is important – just works, with the backend you want. Yes, I’m a dreamer, but that’s necessary sometimes :-)

RPM rollback

I have to admit that I thought I am quite good at knowing which possibilites and capabilities all the package manager and software manager have – but I missed something very important: rpm, and because it can be used with yum as well, Fedora Core has the ability to rollback rpm packages, and that in a very nice way:

This post show how it works, and it is not much.
And it’s even easier to use:

1. To configure yum to save rollback information, add the line tsflags=repackage to /etc/yum.conf.

2. To configure command-line rpm to do the same thing, add the line %_repackage_all_erasures 1 to /etc/rpm/macros.

3. Install, erase, and update packages to your heart’s content, using pup, pirut, yumex, yum, rpm, and the yum automatic update service.

If/when you want to rollback to a previous state, perform an rpm update with the –rollback option followed by a date/time specifier. Some examples: rpm -Uhv –rollback ‘9:00 am’, rpm -Uhv –rollback ‘4 hours ago’, rpm -Uhv –rollback ‘december 25′.

I searched a little bit around about this feature and found eventually this discussion. Looks to me that it is not really sure if this feature will stay in rpm, and if it is really stable. These news are bot so good…

I would like to have such a feature in Linux, or at least in Fedora Core, because it is something “other” operating systems have, too. And it can be very useful if you use test-software or 3rd party software, too.
Just imagine that you have a system where you can use a overbloated gui to “turn back time” for your software versions – that sounds wired, but sometimes useful :-)

Phishing and Spam

Well, I use the internet, I have an e-mail address, and so it is very certain that I get spam – quite a lot of it, something around 80 e-mails a day.

What to do? Well, I use kmail, so the choice is quiet easy: I installed Spamassassin, and went through the very comfortable spam-filter-wizard of KMail. After that, Spamassassin identified the most spam mails, and all I had to do was checking once a day if there was a false positiv.

So long, so good – but still not perfect: I had some phoshing e-mails annoying me again and again. Not that i would click on them, but they were automatically sorted into my e-bay or paypal directories, and I didn’t want it to. So one chance was to train spamassassin against phishing e-mails – but that would be very hard because for example ebay-phishing is very similar to real e-bay mails despite one changed address.
Another problem is that spamassassin works quite slow – it need some time for scanning a e-mail, and KMail is blocked at this time, which is a problem of KMail.

So what to do?

For the second problem I decided to switch over to another spam filter. Although Fedora Core does not come along with another KMail-supported filter, I found a bogofilter package somewhere and decided to give it a try – and it works very fast, much fast than spamassassin. We will see if bogofilter gets slower the better I train it, but from the first experiences it is almost untrained much fast than an untrained Spamassassin, so I’m quite confident about that..

What what to do against Phishing? Luck helped: I installed ClamAV some days ago just to have something if I need it for friends when they send me something. And because I wanted to know if KMail was able to communicate with ClamAV I configured it to scan all my incomming mails.
Some hours later I realized that there was something in the “Virus” directory in KMail – I checked, and found a phishing e-mail without any attachment. That confused me and I scanned the e-mail manually with ClamAV. And see:

$ clamscan Paypal\ Flagged\ Account
Paypal Flagged Account: HTML.Phishing.Bank-159 FOUND

———– SCAN SUMMARY ———–
Known viruses: 42498
Engine version: 0.88
Scanned directories: 0
Scanned files: 1
Infected files: 1
Data scanned: 0.00 MB
Time: 2.108 sec (0 m 2 s)

That’s something I really appreciate! It helps to keep an overview about everything going on in my mail system.
But one note: that can only work if you do not have filters which react to mails comming from “ebay” or “paypal” and stop filtering them after detecting. In this case you have to uncheck the “stop filtering” box somewhere in the filter configuration. Or you could put the virus filter rule on top of all mails. It’s left to you, as you want :-)

MTP – the second

Some months ago I wanted to have a MP3/Ogg player – actually more a Ogg player because almost all my music is Ogg (btw.: I know that the correct term for the audio codec is Vorbis, not Ogg-Vorbis or Ogg, but the file format is ogg, and this is a blog, so I don’t care so strict :-p). I heard lots about the very good Ogg support of iRiver, and decided to buy one of their players. It was in the time when the new T30 came out, and I bought one of the first available – and was deeply disappointed. They used the new mtp-protocol, designed by Microsoft and some others, and at this time only usable with Windows Media Player 10, which requires Windows XP.

I had the hope that iRiver probably would release a new firmware to use the device as an mass storage system, but they answered with no. And there were no help available in the Open Source world because these drivers were just released – I was one of the first people who got one…

Later on I decided to send iRiver to hell, but found satisfaction with a new (btw.: much better) device, an U2 from iAudio.
I’m using this U2 since then and I’m completely satisfied with it. And I started to inform myself a bit more about iAudio and realized that they build quite good players and that these players are mostly working as USB mass storage and have Ogg support – what do you want more?
Actually, I was so pleased that I bought later on a X5 for my sister – a very cool and nice player which has a hard drive and works like a charm – and after all I know technical much superior to the famous iPods.

Time passed by, and I eventually forgot about mtp and this stuff – until today!

Today I was working in the German Wikipedia again, and continued my some days ago started work to polish up all articles about, ogg, vorbis, theora etc. While working their I came pass a site which listed several devices which were able to play Ogg. I saw the T30 there again, and decided to give google a quick look if there is something new on the mtp/Linux front – and their is!

Thanks to these form posts I read about the new development version of gphoto2:

Basic MTP support for iRiver mp3 players using ptp driver T10, T20, T30, M415

gphoto2 now comes with mtp support! Ok, built especially for the iRiver devices, but it’s a very good start, and as it is a protocol-“standard” (Microsoft way of standard, not a real standard) it should work on other devices, too.

I am very happy to read this – especially because Linux now has a support which several other Windows versions don’t have (Windows 2k, Windows XP with older Media Players, and so on). Ok, it’s still a development version, but I hope they will release it soon – and if not, whatfor do we normally use Linux distributions? They can provide packages with extra packages!

Yes, on days like this I really like the Open Source community, the development processes, and that it really works! :-)

Google Talk – open federation

I just had a look into google blog and found the new information that google is opening it’s jabber servers for other servers. That means that everyone who has a jabber address can talk to everyone on google talk – pretty nice!

So Google is using open standards, respects them and is extending them if it makes sense – something other companies don’t make at all. And it’s one of the reasons why I still like google (from a technical point of view, not from a censorship point of view), although they could start publishing their software for Linux to be more liked by me ;-)

Let’s see how fast and well the kopete group will implement the jingle plugin, and how well it will interact with KDE. I like to test around :-)