Archive for the ‘Software’ Category

Using MoNav – Eat your own dogfood

Sunday, March 27th, 2011

I’m just back from a family meeting in Heiligenstadt (Upper Franconia), the region where parts of my family originate from. It’s a nice town with some restaurants providing very good traditional food at a reasonable price. It’s sited in a valley which provides nice surroundings, and the interior of the medieval protestant church was rather surprising.

Of course I used MoNav for navigating there. The route it calculated was “just the right one”, and the track logging I recently hacked clearly raises its value.

What I’m missing most is speech output. Actually I already know how to do audio output on the N900, which is quite simple thanks to Qt. But a lot of backend work is required to create usable output. As CEST started today and I can smell the spring, it is unlikely that I will add this feature during the next weeks. Except we get some cold and rainy days, of course :) .

MoNav Maps updated

Monday, March 21st, 2011
openclipart.org, public domain

openclipart.org, public domain

I’ve modified the script for MoNav map generation and ran it anew. monav.openstreetmap.de now hosts packages for car, bike, and pedestrian routing. The files are compatible with MoNav 0.3 as built from SVN. The files are not compatible with MoNav 0.2. The page also contains some basic installation instructions.

I’ve also uploaded a couple of installers. A deb for the Nokia N900 is provided as well as debs for Ubuntu i386 and amd64. A binary only tar.gz is also available.

I’ve also created a first rudimentary Win-Installer using the Nullsoft installer and Nsisqssg. Unfortunately the binary crashes during startup, so this installer is not available. I’m rarely using Redmond-OS, so in case someone else with some C++/Qt4 knowledge wants to debug the issue I wouldn’t object :) .

Please note that the binaries provided have been built from SVN trunk.
The main purpose of providing the files is to show what is possible with the collaboratively collected data of openstreetmap.org. The software possibly will harm your machines and thus is not intended to be used by end users, but by brave mappers keen on testing bleeding edge software.

Recent MoNav Additions

Tuesday, March 15th, 2011
Displaying a tracklog using openstreetmap.org data (CC-by-SA)

Displaying a tracklog using openstreetmap.org data (CC-by-SA)

Since the last post, I spent a couple of hours writing code for MoNav. The follwing list provides an overview over the lastest additions of mine:

  • There are three map widgets in MoNav, the main map, the place chooser map, and the street choser map. As there was no central instance to control the zoom level, some nasty effects occured in case the user used the various zoom controls (F-keys, mouse wheel, zoom bar, zoom buttons). The zoom levels now are stored persistently for each individual map, and managed by one central method, setZoom(int). The code now provides reasonable default zoom values instead of using the maximum possible zoom at each startup.
  • On the N900, the backlight now remains switched on as long as there is GPS data. Of course the user can still switch off the backlight using the hardware switch.
  • MoNav centered the map only in routing mode. In the map view mode, the GPS position left the viewport without the map following its position. As a cyclist, it may well be that you don’t want to use routing, but want the map to follow the GPS position anyway. There is no preference option for this new feature. The map link to the GPS position gets lost as soon as the user pans the map. It gets set during application startup, when loading another map data set, the user selects “Goto Departure”, or the user selects “Goto Gps-Location”.
  • In routing mode, the auto rotation now is optional. In case you are used to maps which always show north up, this option is for you.
  • MoNav got GPS updates by the GPS subsystem every 5 seconds or so. Though the documentation of QGeoPositionInfoSource states that a default object gets updates each time there are any, calling setUpdateInterval(1000) clearly boosts the GPS performance of MoNav.
  • The user can now switch track logging on and off via the preferences.
  • The tracklog files created by MoNav now are proper UTF-8 encoded files, which JOSM no longer denies to consume.
  • MoNav now is capable of rendering the tracklog in red colour (see screenshot), which is a much sought-after feature for pedestrians, hikers and cyclists.

MoNav Tracklog Rendering

Friday, March 11th, 2011
MoNav Tracklog Rendering (map material openstreetmap.org, CC-by-SA)

MoNav Tracklog Rendering (map material openstreetmap.org, CC-by-SA)

As the openstreetmap addict I am, I want to log any trip I do. Recently I told MoNav how to do it, but the tracks did not appear on-screen.

Meanwhile the tracklog gets written in GPX format natively (support for track segments is provided), and MoNav reads the current GPX tracklog during startup. The tracklog files are named according to the current date (e.g. MoNav Track 2011-03-10.gpx). In case one quits MoNav right before 12PM and fires it up again right after 12AM, the track will appear empty due to the new log file created. Each time the tracklog gets flushed (current default value is every 60 seconds), MoNav creates a backup file (e.g. MoNav Track 2011-03-10-bck.gpx), just in case the battery drained while the file is open for writing.

A couple of minutes ago, I did a short test trip around the station in Karlsruhe, which is witnessed by the above screenshot. Obviously the GPS receiver of the N900 was not that accurate, which is due to the fact that I had it in my pocket, not mounted to the handlebar.

I’ve already seen a couple of bugs, e.g. the track painting is lagging behind the actual GPS position by a couple of seconds. I hope I can iron out such issues during the next couple of days, as the biking season will hopefully start »real soon now«™.

The code is not in SVN yet, as some other patches still are pending for being approved. I’ll keep you posted as soon it is in SVN.

Automated MoNav Map Preprocessing

Thursday, March 3rd, 2011
European Map (openclipart.org, public domain)

European Map (openclipart.org, public domain)

The most exiting routing application is pretty useless without any map material. So is MoNav. Well, wait, you can convert some data from openstreetmap.org at your own, can’t you?

No, many people can’t. In case you are used to a Linux box, you may have good luck, as you have all the tools required at your diposal. But many potential users of MoNav might not be capable of compiling the application and map material by themselves.

So I spent some evening hours to set up a simple script which can automate the preprocessing. As STRATO generously is sponsoring three servers to be used by the OSM community, a few CPU cycles are available to do the hard processing work. Though I have no confirmation yet whether the machines can be used to convert MoNav maps regularly, I already converted some minor mapsets. Do not expect huge sets (like the whole of europe) soon, as this would hit the machine quite heavily.

The download page looks quite ugly (CSS experts wanted), as it gets automatically created by the abovementioned script. But anyway, it provides the possibility to download preprocessed maps for MoNav (currently only car routing maps are offered). Hope you enjoy.

Tweaking MoNav’s Vector Renderer

Sunday, February 27th, 2011
Rastatt in Monav (data by osm and contributors, CC-by-SA)

Rastatt in Monav (data by osm and contributors, CC-by-SA)

I could not resist the temptation to tweak the vector rendering code of MoNav.

(more…)

Vector Rendering in Monav SVN

Wednesday, February 23rd, 2011
Speyer Vector Rendering (Data by openstreetmap.org and its contributors, CC_by-SA)

Speyer Vector Rendering (Data by openstreetmap.org and its contributors, CC_by-SA)

Thanks to the work of James Hollingshead, MoNav SVN contains the first incarnation of a vector renderer which aims to be a high performance renderer which creates the required map tiles “on the fly”. The current SVN code is a bit awkward to compile and use, hence this posting. The info provided here is replicated by the OSM wiki. The following bash commands have been executed using Ubuntu Lucid 10.04.1 LTS. You will at least need Qt 4.5 dev packages, and probably some additional packages before MoNav will compile properly.

(more…)

MoNav tracklogging patch finished

Monday, November 1st, 2010
Animal Footprints (openclipart.org, public domain)

Animal Footprints (openclipart.org, public domain)

My favourite OSM based routing application, MoNav, lacks tracklogging. Thus I sat down this weekend and wrote some code to accomplish this.

The current code collects the data from the positioning subsystem. After a configurable period of time (default is 60 seconds) it flushes this data to a simple CSV logfile. While writing the data, a copy of the file is kept to avoid data loss in case the battery drains at the very same moment. The folder for the logfile is configurable, while the filename gets automatically created according to the current date and time. When closing MoNav, the logfile gets converted to a more versatile GPX track file.

I tested the code during an six hour trip to the black forest, and it seems to work just fine. I hope my code will make it into the main MoNav repository sooner or later.

While I usually do mapping by foot or bike, I mapped steep inclining ways in Obertal by car today. Many of those ways end at isolated dwellings or farm houses, or morph into graveled forest tracks. That’s the area of interest, sited in the Murg Valley, Black Forest:

A very nice trip under sunny and warm weather conditions, untypical for november. I’d like to get more of those, at least until christmas :) .

Hacking MoNav (Dreams come true)

Monday, October 25th, 2010
Travel-compass (openclipart.org, public domain)

Travel-compass (openclipart.org, public domain)

Prelude

As I started mapping and writing software for openstreetmap back in 2006, I did this mainly as I noticed that my Garmin eTrex Venture Cx for 500€ was incapable of doing proper bike routing. Map data was only available commercially, far too expensive for private use.

These days, we had no slippy map or any other maps at all. If we wanted to have maps, we grabbed a portion of data via JOSM, converted it into svg using Osmarender and looked at it in Inkscape. Soon we began to convert the data into the Garmin format using cgpsmapper, and we struggled a lot with unordered segments. We checked the maps in Qlandkarte and spent many hours to order the segments causing troubles. We did all of this as we knew that openstreetmap was extraordinary and revolutionary.

(more…)

Usability painted on a wall

Sunday, September 19th, 2010

In Wissembourg, I came across a building of Glad, a company which sells firewood. I’m sorry for the bad picture. I just stopped the bike and took the picture ignoring all principles I know of. But anyway, here it is:

Mouse pointer on a wall

Mouse pointer on a wall

What I find amazing is the simple but clear message of the installation: »I sell firewood, and you’ll find information both in french and german languages by visiting my web presence.«. In case your job includes coping with user interface design: Can you remember the last time you found such an intuitive metaphor?

Compiling and running MoNav

Sunday, August 15th, 2010
MoNav displaying a route in Karlsruhe

MoNav displaying a route in Karlsruhe

Today was a cold and rainy sunday, so I grabbed the occasion to play with various software which can make use of our collaboratively collected Openstreetmap data. One major gap in the toolchain is the lack of an end user friendly, working navigation software for mobile devices. There are several approaches, but none of them seems to be capable to replace an average user’s Garmin or TomTom device. Most are hard to compile, install, configure and use. And even if you manage to get it up and running, it is likely that you miss some features which you knew on your proprietary devices.

(more…)

Openstreetmap Routing News

Monday, July 19th, 2010
Beach Trip (openclipart.org, Gerald_G, public domain)

Beach Trip (openclipart.org, Gerald_G, public domain)

Most of you might already have heard about it, but anyway, here are two additions to the open source routing pool.

Vodaphone released most of the code of their discontinued Wayfinder navigation product into the world wild web. It seems the code needs maintenance before it is useful. Let’s see whether some enthusiasts are taking over.

The far more interesting announcement, though, has been posted right at the beginning of this year’s State of the Map conference. The Open Source Routing Machine (OSRM) is a »high-perfomance routing backend«. Check out the demo provided via Geofabrik. It’s just some C code. After compilation, you’ll get a handful of binaries, including the usual tool to convert OSM XML data into some binary file format and a web server which handles the requests and delivers the resulting route.

The speed is just amazing. However, after reading some of the source code, it currently appears to obey the following limitations. What I recall:

  • The result gets shipped in KML format. There’s nothing wrong with that, but I guess a GPX format was more versatile. I’m pretty sure, though, that KML has been chosen with reason.
  • AFAIR there’s only the web server’s output, no file output.
  • The code can only cope with a single tag per element, e.g. using the highway tag. It currently is not possible to select items by multiple tags, e.g. highway=track, tracktype=grade1.
  • Most options are hard coded, so it’s not possible to play with the options by editing a configuration file.

Anyway, it’s the initial release, and it looks very promising. I’m pretty curious how it will be accepted by the community, how it will develop over time and whether there will be collaboration or competition between OSRM and Routino.

First Offline Bike Routing on the N900 using Routino

Monday, June 14th, 2010
Biker pictogram (openclipart.org, public domain)

Biker pictogram (openclipart.org, public domain)

As a long time contributor of the openstreetmap project, I always dreamt about using our collaboratively collected data not only for creating maps, but also for (offline) routing. Actually, several routers exist, but none of them seemed to be really capable of doing offline bike routing.

Appearently I missed one bit. This weekend, I tried Andrew M. Bishop’s Routino. It’s a set of command line tools to convert osm XML files into a binary format and to use this data for routing. Additionally some helper applications are provided.

Surprisingly, the code cross-compiled for Maemo 5 without any further interaction (as advertised), just by typing make. Hmpf. Hmpf. Frankly, I was very exited.

The application is highly configurable via some XML files. Thus it supports routing using configurable profiles as well as shortest and quickest routes while supporting via points. It can output the route in various formats. I’ve converted a small set of osm data and tested the results for bike routes and was quite impressed.

I wrote some experimental code using PyGtk to make using it a bit more convenient on the N900:

Routino Frontend

Routino Frontend

It can get the start position from the location system of the N900. Unfortunately I struggled with DBus and was not successful yet getting the viewport coordinates from Mappero for convenient graphical destination selection, but I hope to get this working during the next days. Currently, the destination can be copied over using the clipboard.

Here’s a route calculated by Routino and loaded into Mappero:

Routino Route loaded into Mappero

Routino Route loaded into Mappero

It’s far from being convenient to use or user friendly. But for me a dream comes true: Offline bike routing on my mobile computer. Thanks Andrew, you made my weekend!

OS-Install-o-mania Parts 1, 2 and 3

Monday, May 31st, 2010
Unintentionally trashing the Vista installation on a Dell computer

Unintentionally trashing the Vista installation on a Dell computer

One decade back, I was installing operating systems quite often. Win 98 was broken every now and then, and Linux was developing very quickly.

Meanwhile I’m very reluctant in updating operating systems. It just takes too much time to set up the system, to reinstall all required applications and to do all the configuration work. Or the other way around: It’s just a waste of time.

However, there are circumstances where your machines need some healing hands.

#1: N900 Upgrade

This week, a much anticipated update for the Nokia N900 was finally released. I was eagerly waiting for it, as I hoped that it would erase »The Big Bug«. The update was advertised as an »over-the-air«-update, installable via a wireless network connection. Unfortunately, the device didn’t offer the update on my machine. After waiting for a couple of days, I got suspicious and searched the web for the reasons.

Obviously the device performs some checks first, and in case certain circumstances occur, it denies to offer the update. The reason might be that I enabled an additional repository and installed (only) a couple of extra software packages. Further, the N900′s operating system is not installed to the 32GB flash memory. Instead, it lives in some wee small extra partition of about 200MB. I’ve never understood why it is designed this way and why the space is that limited. Unfortunately, the Debian-like operating system stores each package in a folder in this small partition, further reducing its size.

So I ssh’ed into the phone and tried to install the update using apt-get. It told me that it wants to download about 100MB of packages and that there was not enough space left. I created a folder on the 32GB flash drive and symlinked to it. Now the update packages were downloaded and the update went fine – halfways. I saw some warnings during the update, and finally it stopped.

So I did it the hard – by reflashing the device. I downloaded the current firmware image and a GUI tool for the Mac. Unfortunately this tool failed to complete the flashing process so the device was bricked. I tried again using my Linux box and everything went fine now. As all personal data lives on the 32GB storage which was not touched by the flashing process, I just needed to reinstall some additional software packages (mainly those for mapping) and was back in business.

#2: Reinstalling Vista from a recovery CD

The backlight of the display on my Dell notebook was broken (it appeared as a defective contact, as the backlight switched off every now and then). I’ve a service contract and hoped they would replace it as soon as possbile (»next business day«). Surprisingly, there was some (bi-daily) mail exchange where they told me what to do – one bit of information per message.

First they asked me to install some BIOS and driver updates, though I wrote in my initial message that I’m running Linux on the machine and that everything looks like a hardware defect. I wanted to do them the favour and booted Vista from the (seldom used) NTFS partition. After downloading and installing the driver updates, Vista prefered to show a blue screen at each attempt to boot it. Doh.

Dell further asked me to run a diagnostics tool before booting the operating system. This led me nowhere, as the backlight of the display always switched off while the tool was running.

Next they offered to catch the device and to repair it off-site. The device would be unavailable for up to two weeks. Hey guys, there was a reason why I spent the extra bucks for the on-site service.

Finally, after almost two weeks, a technician dropped by and exchanged the display. Heck, I didn’t remember how bright the display once was :) .

I do not really rely on the Vista installation, but I boot it every now and then to compile and test some code of my own. Due to some file system errors, problems with KDE after upgrading Ubuntu, the broken Vista installation and the fact, that the 15 GB partition was too small for Vista (yes, Vista needs at least 10GB for its installation, 2GB for a swap and 2GB for a hibernation file), I decided to set up the complete machine anew.

I backed up my home directory and used the Kubuntu installation disk to repartition the hard drive.

The Vista installation went halfway straightforward thanks to the Dell recovery CD. But later on I got stuck as there was no driver installed for the wireless network adapter. I tried to install it from an additional CD that came with the device, but to no avail. I finally downloaded some drivers from the Dell page and got it working, but this still was a challenge as it was not obvious which file to download and install. This was surprising as I passed the service tag of the device to the online support system.

Finally Vista insisted to download and install plenty of huge updates. I read »Usability Engineering kompakt« meanwhile :) .

#3: Installing Kubuntu 10.04 LTS

This one was simple. Insert the CD, enter some data, select the right partitions for installation and let the installer alone for a couple of minutes. After booting the system, plug in the USB hard drive containing the backup of your home directory and copy the files back to the hard drive. Log in and feel comfortable.

Conclusion

It was good luck that it was a rainy sunday. I really hate those bread- and senseless installation orgies. I hope the machine will now run until it gets replaced.

Adieu KDE – peu à peu

Sunday, May 30th, 2010
Office utils (openclipart.org, public domain)

Office utils (openclipart.org, public domain)

I’ve been a KDE user for a very long time – actually since I use Linux on my machines. I remember the day when I wiped Win 95 and installed SuSE Linux (do not recall the version) back in 1998. I sat in front of the computer, staring at the KDE filemanager. And I thought: »And that’s the operating system better than Win?!?«. Meanwhile I know better, and KDE matured over the years. I helped with a bunch of applications, writing manuals, filing bug reports, and finally started to write my own code to get some tools I missed.

KDE 3.5 was a very usable and stable system, while KMail still is an excellent mail user agent.

However, the switch to KDE 4 was a major break, introducing lots of rewritten code and new design concepts. While I appreciated all the ongoing changes, I left my system untouched, still running KDE 3.5.

I upgraded as my distro was outdated and the repositories went offline. This way I got KDE 4.3. Well, it was running and ready to use, but there were lots of issues which did remind me of the days when struggling with Win 95.

One source of eternal joy is the mess around KDE PIM, Akonadi, Strigi, Nepomuk, Virtuoso-t and the like.

Virtuoso-t CPU usage, draining the battery

Virtuoso-t CPU usage, draining the battery

Kontact, the PIM application embedding mail, contacts, feeds and the like, started to rely on Akonadi, which in turn requires an SQL database server for operation. Unfortunately it didn’t work out of the box, and after following some docs on the net and trying out several meds to cure the problem, I finally gave up and decided that I will never run an SQL server just to manage some personal data. So I have to close the annoying Akonadi error message dialog each time I log in.

Akonadi error message after each login

Akonadi error message after each login

I understand the design concepts behind the PIM suite and am impressed of what the KDE community tries to achieve. But at first I need a working system where I can perform basic tasks. Further, I want to create backups just by copying the data of my home directory to another hard drive. I will not start to write specialized scripts which will backup and restore some additional data of a database server. If I wanted to, I’d install an Exchange server and enjoy the journey.

One of the cool things I always liked in KDE was the fish-protocol which allows you to edit files transparently directly on a server over an ssh connection. Unfortunately this was one of the features broken in KDE 4.3.

Meanwhile I upgraded my distro and got KDE 4.4. The fish protocol is operational again (which is great). But the problems with Akonadi remained. I still cannot access my contacts from within Kontact, as it complains about problems with the Akonadi server.

Impossible to access contacts

Impossible to access contacts

Additionally, it seems that there still is absolutely no possibility to sync any of the PIM data with my cell phone.

The file indexer continues to keep my CPU and notebook fans busy (hi nepomukservicestub). I saw it eating one of my cores completely, but also saw it at 70% CPU usage.

Nepomukservicestub easting 49% CPU

Nepomukservicestub easting 49% CPU

Dolphin, the file manager, also shows some severe bugs. Today I wanted to change the file permissions of an (admittedly huge) backup file available on an external USB hard drive. I rightclicked and chose to open the properties dialog. The result was a frozen file manager which I killed manually. After two further attempts I used the command line to set the file’s properties. In case you expect that all this fan blowing at least makes sense because you’ll easily find files afterwards, you might be wrong.

Nepomuk search failed

Nepomuk search failed

To me KDE 4 appears as an overloaded desktop environment with too many applications, too many features and too many bugs, which in turn has problems performing basic tasks like managing files, PIM data and doing some sync to a cell phone. I really appreciate the decision to turn the computer into a social desktop to manage one’s internet activity. But I wonder whether it was better to keep the rewrite of the desktop separate of the breaks and changes of the PIM applications. The propellerheads behind KDE were brave and decided that they better do it the hard way. KDE probably will be a cool desktop in a couple of years again.

But meanwhile I need a replacement, and less surprisingly I’m considering to give another desktop a try. Besides Gnome and Evolution, there are other window managers and PIM applications. If you’re a KDE switcher and want to share some experience, please drop me a line.

Mappers heading towards Wissembourg

Sunday, May 23rd, 2010
La Maison du Sel en Wissembourg

La Maison du Sel en Wissembourg

Finally the rainy days are over. The warm and sunny weather conditions allowed for the first mapping trip to Wissembourg by bike – the biking season finally has started.

I mapped a lot of details along the way. Some short breaks every now and then and finally a nice supper at the destination took time. Luckily »Deutsche Bahn« (a german railway company) provided public transport back home.

The location services of the N900 are pretty good, but there are also some weak points:

  • The N900 uses assisted GPS (AGPS) to get its position really quickly. If however no callular net is available, it needs ages (up to twenty minutes) to get a GPS fix. I’ve also seen such a behaviour on one of Garmin’s premium GPSrs, but only every now and then. The N900 does it each time it is outside the network of my provider. I guess the behaviour could by improved by saving an almanac to the device, but appearently the N900 doesn’t do it.
  • The first fixes the GPS subsystem delivers to the requesting application are sometimes really far (and I mean far) away from the actual position. This makes it difficult to check the trip’s length at home. I’ve never seen such a behaviour on the Garmins I used, and I consider this being a software bug.
  • Modern vehicles use screens with a vapour deposited metal layer. This hinders the GPS signals flowing into the vehicle. If the Garmin’s have a fix outside the vehicle, they are usually (but not always) capable to keep the fix after entering the car or train. The N900 today had difficulties, but as the WBT I also used today did not really do much better, I guess there were really bad GPS conditions in the train.
  • The big bug requires a second device which is more reliable for collecting tracklogs. Both Maep as well as Mapper loose their tracks in case you must hard reset the device.

Here’s how the tracklog recorded in the train looks like. Note that it should follow the rails precisely:

The device is an almost perfect gadget anyway. All I want to say is: If you reliably want to collect the tracklogs of your trips, just get an additional logger. The N900 is an exccellent general purpose device. It is not a specialized GPS receiver and logger.

Osmarender rules changed to render forest parcel names

Sunday, May 9th, 2010
Forest parcel name rendered in osmarender, zoom 17

Forest parcel name rendered in osmarender, zoom 17

At crossings of tracks in the woods, you often can find up to four name plates displaying the names of the adjacent wooden parcels. I usually placed a node tagged as place=locality and added the names. Of course this leads to a situation where the names will overlap each other, and after collecting some of those plates, it was desirable to draw the parcels as areas with the name rendered at the center, not as individual nodes at the vertices. I currently add a polygon tagged as landuse=forest.

While Mapnik renders the names of natural and landuse features in zoom levels 14 through 17, Osmarender only rendered them visibly at zoom level 17. It rendered natural names at zoom levels 16 and 15, but subsequently the caption was hidden by the rather huge surrounding wood polygon. landuse was missing in those levels.

I’ve thus adjusted the render rules so the names of natural and landuse features also get displayed at zoom levels 16 and 15. I hope this won’t mess up the map in densely populated areas. If so, please drop me a line and I’ll change the additional k="landuse" v="*" to something like k="landuse" v="forest|wood".

As soon as the changes are slurped by the tiles@home clients, I’ll start to convert the locality-nodes to something like amenity=nameplate or similar to finally clean up the map.

Mapping Hiking Routes

Saturday, May 1st, 2010
Hikingmap Route

Hikingmap Route

While curing some cold, I mapped a hiking route to make it appear on the Hiking and Bridlemap.
(more…)

LAC 2010 conference

Friday, April 30th, 2010
Audio Icon (openclipart.org, public domain)

Audio Icon (openclipart.org, public domain)

Though I’ll not attend, I’d like to advertise this year’s Linux Audio Conference, “the conference about Open Source Software for music and sound” which takes place from May 1-4 2010 in Utrecht, the Netherlands. Hopefully the programme wets your appetite.

Wiping the bugs in favour of new ones

Thursday, April 29th, 2010
Hürschkäfer (Lucanus Cervus)

Hürschkäfer (Lucanus Cervus)

Being a Linux user for about 12 years now, I’ve also used KDE since version 1. I also used version 2, and KDE 3 finally provided a comfortable user interface for a Linux driven computer.

However, I avoided to install version 4 due to the heavy changes to the underlying system. I needed to install KDE 4.3 anyway as soon as the lifetime of my operating system was over. And frankly, 4.3 still had some minor but rather annoying bugs. I don’t want to complain. I’m a member of the global open source developer’s village, and my very own code is far from being bug free.

Meanwhile KDE 4.4 is out, and Ubuntu 10.04 (aka Lucid Lynx) was released today. Operating system upgrades usually break some things, so I’m usually reluctant to install them. But today, I could not resist.

So I’m running a backup right now, and will run the upgrade immediately after it. The weekend will be a rainy one, so I’ll have some time to fix potential issues. Wish me luck :) .