Entries from October 2004 ↓

Wim, Johan and me meet up with some cool and crazy hackers in Barcelona last night trying to help them get Flumotion going so they could stream the Piksel conference in Norway today with it. They had a really cool office in Raval which was part office part hackers cave :) Got a system called Pure Data demonstrated which where quite cool. Kinda like gst-editor on steroids. Had a great time compiling software, drinking, lots of ‘tobaco’ being smoked and in the end debating world problems.

Bilboed reported good progress on Pivitivi
today. Happy to see a GStreamer and Gnonlin based Non Linear editor finally nearing a first release. I have promised to help with advocating Pivitivi and grow a vibrant development community around it. Hopefully we succeed :)

Been playing my music over the last days using Amarok. Its using GStreamer and Markey have been doing some great work both hacking GStreamer and Amarok to get it to where it is today. If you are a fan of xmms style music players this is definetly a step up from xmms. The inline visualizations for instance are very nice; and the fade to silense on exit is another cute touch. Not sure how much I agree with the UI choices done in the playlist setup, but with the huge amount of music players out there, variety is good. No point in having 10 players who behave identical :)

Got the contract with David Vignoni signed today so he will start creating new artwork for Fluendo starting monday. Hopefully we will have a Flumotion release it not to many weeks looking stellar :) Having worked with David a lot before, since he is the maker of the Nuvola icons which I put into gnome-themes-extras, I am really looking forward to working with him on Fluendo related graphics. Might also make getting Nuvola in g-t-e up to date easier :)

Weekend aproaching now. Going to the cinema today with Johan Dahlin to once again try to see collateral. Then afterwards I probably try to hook up with Wim to experience some more of Barcelona’s nightlife.
On thuesday thomasvs will be thankfully be back. He way to central to our operations to not have on hand here. Already have a long list of items I need to discuss with him.

Have been working a lot on the Flumotion manual over the last few days (flumotion-doc in svn for those interested). Those into this will know that when you encode a stream or file with many formats which are variable bitrate formats you can choose to encode in a specific bitrate (more predictable for streaming) or encode towards a specific quality level which keeps the image looking better but the stream could at times grow in size which of course often is problematic if you are streaming over the web.

Anyway you set this by typing in a number either for the fixed bitrate your want or the quality you want. But of course people without a lot of experience in the field would have no idea what bitrate or quality number they want for their theora video (myself included). So I figured trying to relate the quality and bitrate settings to stuff like DVD or VHS in the manual might be an idea. Well easier said than done, first of all there are no objective ways to measure this and secondly few if any people seemed to have a clear idea on what values would relate to what qualities. That you have to add resolution into the equation too doesn’t exactly make it easier. I found one test online comparing sound quality between codecs on various bitrates, but it only convinced me that my plan was much easier to come up with than execute.

So instead I ended up explaining the difference between setting the bitrate and setting the quality and suggest the person streaming do some tests to find a level that fit their demands :)

jdahlin have been doing some cool work lately on Flumotion getting it to autodetect the available devices on each worker. Which means that if you install flumotion workers on 2-3 machines each with different kind of video/audio input devices and then bring up the flumotion wizard it will give you a selection of devices based on what is available on the machine of the worker in question :) nifty!

Also started to see some really good progress on some work we contracted out and sent out a second contract for our second contract asignment. Since it is part of my job to handle them it feels very good when things seems to work out smoothly. Soon starting some work where we are the ones being contracted so I guess we will know how it feels at both sides of the table :)

New release of the server out yesterday. Glad to see that the issues reported where either known or small. Which means it will be easier to move ahead instead of being ‘stuck’ where we currently are we the server in order to fix basic issues. Thomas and Johan have been doing some fantastic work on the server so far.

e8johan: The answer to your question is both yes and no. Yes, sure it is a plus, but on the other hand there have been and is lot of other options for easier OO in GNOME using C++, Python and Java for instance. So concluding that Mono developments main attractiveness is due to easier OO is probably much to simplistic. I think it is more that people feel that with Mono the platform is actually taking a big step forward; ie providing more than just some extra programming language gloss. That Miguel is a good advocate probably doesn’t hurt Mono either :)

That said I think the position of Python as a programming language in the GNOME community is understimated. Considering that Red Hat seems to use it for most of their stuff, Ubuntu is using it for their stuff, my own company Fluendo is using it for our stuff, and of course a lot of independent developers are using it, I think that Python is probably the most widely deployed development language for GNOME apart from C.

Pointing back to the mono discussion I guess Miguel would be fast to point out that you can easily tie inn Python stuff to Mono using Iron Python. Although afaik no of the main commerical or independent users of python are doing so yet :)

Just read Jonathan Schwartz blog entry defending the settlement of the Kodad lawsuit. It was sad reading. I guess it would be impossible for him to say that the settlement was the stupidest thing Sun have ever done, McNeally would probably have fired his ass for it. But no matter what angle Jonathan tries to put on it Sun showed less backbone and insight in this case than Microsoft has done with Eolas or IBM have with SCO. And if he thinks people in Java community will praise him for giving in to Kodak I think he is wrong. What Jonathan fails to see is that giving in to Kodak (or anyone else) is wrong cause it leads to the same problems as giving in to terrorists and blackmailers; it will spur more such actions. Maybe Sun have got Kodak of the back of Java developers for the time being (of course if I where Kodak I would really start looking around for other patents to bring up knowing that Sun is ready to cough up huge amounts of money no matter what), but they have also made sure that thousands of other people with dubious patents have seen and learned that taking Sun to court over them will get Sun to hand over huge amounts of money even if they, as Schwartz himself admit, think the claims and patents are bogus. And the natural extension of that is of course that if a deep pocketed giant to Sun are prepared to pay great amounts of money to settle groundless claims then surely smaller players (like the companies and people in the Java development community) would be even easier pray wouldn ‘t they?

So Jonathan you haven’t saved anyone, you just brought hell one step closer to our door. Thanks…..

Spent the weekend back in Oslo. It is incredible how much more you enjoy the Oslo fall when you know that you will only be exposed to it for a short while; that you have a warm refuge further south to retreat too again.

Managed to get some stuff I wanted like a fan for the bedroom ceiling. There is a little funny that I was easily able to find that home in Norway, yet have been unable to find it in Barcelona as the need for such a thing is ‘a tad’ bigger in Barcelona.

While I was resting on my laurels home in Norway the rest of the team where hard at work preparing for a release of our streaming server in Barcelona. Things are starting to look pretty nice and I was even able to make a few screenshots 1,2,3,4,5
of the wizards and a test videoscreen running :)

The plan is to make a release this evening so that people interested can start playing with it. We probably keep this first release relativly lowkey since it is surely to contain some bad bugs. Then we do a new release in a week or two which we announce more loudly.

Discovered a new GStreamer application today called Buzztard. It is a application like FastTracker or SoundTracker for those of you in the know about these kind of applications. One interesting comment that is both a bit tragic and rather amusing at the same time is the comment about the application which they are trying to duplicate the functionality of, a windows app called Buzz. It turns out this application Buzz is not being developed anymore since the developer making it lost his source code…..kinda says it all why open source is the way to go for small garage software projects.

Another day of reading through contracts, updating myself on the technology issues involved and more. Discovered Bastien felt I overstated his intentions in my latest gnomedesktop.org posting :). Sorry Bastien.

Anyway Wim and Ronald is continuing their crusade to fix playback issues. An issue where people using newer version of gnome-vfs didn’t manage to view the Fluendo stream using Totem turned out to be a server bug. Since we haven’t seen the bug with any other software I am guessing the issue is so common that most applications have a workaround for it. Which of cause makes the issue harder to spot for server developers and thus many pieces of server software have the bug, which in turn pressures applications developers to add a workaround for it. And you have an evil circle going. Anyway the server is now fixed and the stream streams beautifully even with gnome-vfs 2.8 (with no work-around :).

Planing on reading the RTP book we have at the office. While will not be doing any of the implemenation work myself I felt I needed to know more as RTP and RTSP of course are very important protocols in our business. So Colin Perking, I don’t know who you are, but I do hope your technical writing is understandable also for mere mortals :)

We also got a book at the office called ‘Streaming Media Bible’ not long ago. I had hoped it would be a good book to learn some basics of streaming for the less technically inclined at the office. The book has some usefull parts, and I guess it will be a better book for Alexia than me, but I have to said I was disapointed on the book. One thing I was hoping to find for instance was some in-depth discussion on the various streaming protocls and their pro and cons. Think I found one small paragraph making some broad statements so far.

Started reading Harry Potter at home before going to sleep. Did half of book 3 last evening and I guess I finish it today or tommorow.

Also returning to Oslo on Friday for a short weekend visit. Will try and arrange for some more of my stuff to be transported down to Barcelona. Will also be nice to get away from appartment fixing, work, and spanish for a few days :)

It is incredible how much small issues can cause big pain. I have NVIDIA card on my machine and have been trying for some time to get a program that uses some NVIDIA specific OpenGL stuff and the Cg tools. Problem is that this program needs the NVidia version of glext.h to compile. This have given me a series of headaches. First of all it turns out this header are not installed by the NVidia installer script unless you run it with –opengl-header as a command line option. Secondly since I was using the installer this header file kept being overwritten when I got security updates etc. of Mesa.

Ok so tired of playing that game I installed the nvidia rpms today form Livna.org. Turns out they are built without the –opengl-headers option so the files are missing in there too :) Filed a bugzilla report to livna.org so hopefully there will be a new package soon with the needed header files.

sometimes it is a bit painfull living on the edge :)

Pushed new releases of GStreamer, GStreamer Plugins and Totem today. Hopefully good enough for Fedora 3 already :)

Read The DaVinci Code yesterday. On many levels I find it to be as good as the reviews leads you to belive. For instance he has clearly researched the background for his topic very well and weave his story very well together with the real world. In regards to raising important spritual questions I think his book is a a masterpiece. The problem in my point is that he seems to have skimped somewhat on the character development and motivations in the book. For instance I am still not sure why a woman who loves his grandfather very much, who has been raised by him and due to that probably share his rather liberal values should be so morally shocked when learning that her grandfather participates in sexual rites that she refuse to speak to him for ten years. The reaction seem to extreme and her anger to longlasting considering who she is and the nature of the ‘transgression’.

The trigging event of the book also seems a bit weird considering the conclusion of the book. Considering that fear of loss of knowledge is the starting point then a conclusion which basically says there was no real risk of knowledge being lost is kinda weird, especially when the person at the begining should have known this.

Also having the crock behave crocked at a point in time when behaving non-crocked would rather obviously give him increased chance of success towards his goal is also a bit weird, although the author do try (unconvincingly) to justify his behaviour somewhat.

Another point never explained is that if your family is in continues mortal danger until the secret you guard becomes publicly known then you have a rather strong motivation to reveal it. The rather cavalier attitude of ‘grandma’ in regards to the urgency of publicising it seems a bit out of place due to this.