Cambus.net

Installing Go on Mac OS X

2012-02-06T17:42:00+01:00

Installing Go on a Mac OS X system is a pretty easy process. In order to do so, you will need to have Xcode and Mercurial installed; it is possible to download precompiled binary packages from the official Mercurial site, but as I have Homebrew installed on this system, that’s what I’ve been using instead.

In your home directory, type the following command to clone the Go repository :

hg clone -u release https://go.googlecode.com/hg/ go

We then compile everything by invoking the following commands :

cd go/src
./all.bash

On my system, the whole operation is quite fast and takes approximately three minutes to complete (compilation + all unit tests).

real    3m10.309s
user    2m48.050s
sys     0m41.652s

Finally, you need to set a few environment variables in your .bash_profile file :

export GOROOT=$HOME/go
export GOOS=darwin
export GOARCH=amd64
export GOBIN=$GOROOT/bin
export PATH=$PATH:$GOBIN

Congratulations, Go is now installed on your system and you’re ready to start working with it!

RIPE Atlas : an Internet measurement network

2012-02-02T17:34:00+01:00

I’ve been taking part in the RIPE Atlas community since a few weeks already, a project which aims to build the biggest Internet measurement network ever. By doing so, I’m now hosting a small probe (built around a Lantronix XPort Pro module) sent by the RIPE NCC which is doing round-the-clock reporting.

This is a picture of my setup, the probe is connected to an USB power adapter :

And here is a closer picture of the probe :

Basically, the probe is measuring round-trip time (RTT) to the each of the 13 root servers and also does ”anycast instance discovery” measurements, which is precisely why I find this project particularly interesting. Put simply, it allows to know which instance of a given root server you end up querying; for example, when I query the K-root server, I end up connecting to the DENIC operated instance located in Frankfurt (Germany). The RIPE NCC produce publicly available maps using the collected metrics, which allow gravitational radius visualizations for each server instances. For reference, here is list of all root servers instances.

Unexpectedly, the probe is also a nice way to monitor my Internet connection, which according to this graph is pretty stable :

OpenDNS Stickers

2012-01-30T21:29:00+01:00

Earlier last week, I received the OpenDNS stickers I requested through their dedicated e-mail account at the end of December. It was a nice suprise, as I was just coming back from work and had totally forgot about this. They are some nice pieces of DNS related collectables, and I couldn’t resist posting a picture here :)

Even if I’m not using their public DNS servers on a permanent basis, they sure proved useful along the years and saved me more than once in some desperate situations. I should mention that I also appreciate their efforts in raising DNS awareness to the general public. So thank you OpenDNS!

PS : They now have a dedicated page for the stickers, and you can request some here.

Hurricane Electric IPv6 Certification

2012-01-18T12:55:00+01:00

I recently completed Hurricane Electric IPv6 certification and reached the “Sage” level. This was a fun and educational process and I would recommend it to anyone interested about the topic, as it’s a really nice way to start getting familiar with IPv6.

Also, I should mention certified IPv6 Sages are entitled to receive a free IPv6 themed t-shirt :)

ANSi Web Browser concept

2012-01-03T17:00:00+01:00

This is an attempt at building a text mode Web browser interface. It has built in tabs, back and forward buttons, an home button, an address bar with favicon display, bookmark icon, and country flag. It also comes with a few default extensions : a pagerank plugin, a currency converter, a weather plugin, and a music player as well. What else could you ask for? The possibilities are infinite.

By the way, in case you wonder if such a browser makes sense in the social media era, the answer is YES : you can of course share the pages on various social media sites. The icons used come from the ANSi Social Media Icon Set (CC-BY-SA 3.0 licensed).

The rationale behind this design is quite simple : text mode is the undisputed best way to display information on a computer screen. There is no other option.

You can download this mockup as an ansi file, or converted in PNG (with iCE colors turned ON) for easier viewing.

Happy New Year 2012!

Compiling Apache 1.3.x on modern Linux distributions

2011-11-10T22:37:00+01:00

For a lot of reasons, some people still want to be able to run Apache 1.3 on modern Linux distributions, and unfortunately for them, it doesn’t compile out of the box anymore. The encountered errors are, however, easy to fix and you will learn how to do so in this tutorial.

We start by downloading and unpacking sources of the latest 1.3.x version, which happens to be 1.3.42 :

wget http://archive.apache.org/dist/httpd/apache_1.3.42.tar.gz
tar xvfz apache_1.3.42.tar.gz

We then run the configure script :

cd apache_1.3.42
./configure

Most likely, this will fail and this error message will get displayed :

+ Warning: Your ‘echo’ command is slightly broken.
+ It interprets escape sequences per default. We already
+ tried ‘echo -E’ but had no real success. If errors occur
+ please set the SEO variable in ‘configure’ manually to
+ the required ‘echo’ options, i.e. those which force your
+ ‘echo’ to not interpret escape sequences per default.

+ NOTE: You may also need to edit the shell invoked by
+       ‘configure’. Some shells (e.g. dash) have a
+       faulty echo builtin.
+ using installation path layout: Apache (config.layout)

This happens because in most modern distributions, sh is just an alias to dash, and dash interprets escape sequences which are therefore not echoed as they should.

The workaround is to simply use bash instead of sh to run the script :

bash ./configure

Now, let’s compile everything :

make

After compiling a few files, the process will halt with this error :

gcc -c  -I../os/unix -I../include   -DLINUX=22 -DHAVE_SET_DUMPABLE -DUSE_HSREGEX -DNO_DL_NEEDED `../apaci` htpasswd.c
htpasswd.c:101:12: error: conflicting types for ‘getline’
/usr/include/stdio.h:671:20: note: previous declaration of ‘getline’ was here
make[2]: *** [htpasswd.o] Error 1

The internal Apache getline function is conflicting with the getline function from the standard I/O library. We can simply fix this by renaming the getline function, which is defined and called in htdigest.c, htpasswd.c, and logresolve.c :

sed -i ‘s/getline/apache_getline/’ src/support/htdigest.c
sed -i ‘s/getline/apache_getline/’ src/support/htpasswd.c
sed -i ‘s/getline/apache_getline/’ src/support/logresolve.c

Let’s run make again, and everything should finally compile without problem :

make

Please note, however, that Apache 1.3 branch reached end of life and is therefore depreceated and not supported anymore. It means you will not be getting any bugfixes or security updates and should anything happen, you will have to patch the source files yourself.

Setting up a Node.js development environment with NPM and Cloud9 IDE installed locally

2011-11-05T17:45:00+01:00

In this tutorial, you will learn how to setup a complete Node.js development environment, including NPM (the Node Package Manager) and Cloud9 IDE to edit, run, and debug Node programs. The following installation instructions have been sucessfully tested on Debian 6.0 “Squeeze” and on Ubuntu 11.10 “Oneiric Ocelot”, but they should also work properly on previous versions as well.

We will install everything in the user’s home directory : this way, there is no need for any root access to the machine (provided the required packages are, of course, already installed), and each user can manage its own version of Node.

We start by installing (as root) required packages to build Node and later fetch NPM and Cloud9 :

apt-get install build-essential g++ libssl-dev curl git

We create a ‘local’ directory within our home directory, and modify our .bashrc file to add it permanently to the $PATH variable :

mkdir ~/local
echo ‘export PATH=$HOME/local/bin:$PATH’ >> ~/.bashrc
source ~/.bashrc

We fetch, unpack, configure, compile, and install Node. At the time of writing, the latest stable version is 0.4.12, so we simply set the temporary $NODE_VERSION variable accordingly :

export NODE_VERSION=’0.4.12’
wget http://nodejs.org/dist/node-v$NODE_VERSION.tar.gz
tar xvfz node-v$NODE_VERSION.tar.gz
cd node-v$NODE_VERSION
./configure –prefix=~/local
make install
cd ~

We then fetch and install NPM :

curl http://npmjs.org/install.sh | sh

And finally, we install Cloud9 IDE via git :

git clone git://github.com/ajaxorg/cloud9.git

If everything went fine, congratulations, you can now launch Cloud9 IDE and type your first program. First launch can take some time, as the shellscript will need to get some required dependencies prior to run the program.

~/cloud9/bin/cloud9.sh

Note that the Cloud9 launcher accepts parameters allowing to specify IP and port to listen on, workspace directory, and so on and so far. Below is a list of all available options :

Show this help message
    –help
Load the configuration from a config file. Overrides command-line options. Default: 
    -c, –config 
Run child processes with a specific group. Default: 
    -g, –group 
Run child processes as a specific user. Default: 
    -u, –user 
Activate debug-mode. Default: 
    -d, –debug
Define an action to execute after the Cloud9 server is started. Default: 
    -a, –action 
IP address where Cloud9 will serve from. Default: <127.0.0.1>
    -l, –ip 
Port number where Cloud9 will serve from. Default: <3000>
    -p, –port 
Path to the workspace that will be loaded in Cloud9 (may be relative or absolute). Default: <.>
    -w, –workspace

Now, point your browser to the following URL : http://127.0.0.1:3000

Create a new file (Select File, New, then JavaScript file) and paste this code (this is the Hello World HTTP server example from the Node.js website) :

var http = require('http');

http.createServer(function (req, res) {
  res.writeHead(200, {'Content-Type': 'text/plain'});
  res.end('Hello World\n');
}).listen(1337, "127.0.0.1");

console.log('Server running at http://127.0.0.1:1337/');

Run it :

And check the result in your Web browser :

Congratulations, you just ran your first Node program : see how easy and straightforward it was?

You now have everything ready in order to start developing Node.js applications! Should you need any information or resources to get you started, you can find useful links in the NodeCloud directory.

SciTE : a lightweight text and code editor

2011-09-18T20:30:00+02:00

SciTE is a lightweight, fast, and crystal clear text and code editor for those who love minimalism and efficiency. It is, however, a fully featured and highly customizable editor supporting tabbed files, code folding, and language specific syntax highlighting. It is one of my editor of choice for small projects : actually, I’m typing this article using it.

SciTE feels oldschool, a reminiscing gem from a software era when efficiency prevailed over complexity and bloating : in fact it feels so good that I decided to use FVWM (which is, for the record, the first Window Manager I ever used back in early 1996) to take a screenshot.

Some interesting facts and features :

It is built upon Scintilla (a source code editing component), and is extendable through Lua scripts.
Cross-platform (Good to note : the Windows version is just one single .EXE file!).
Interesting selection features : SciTE allows rectangular block selection by holding the CTRL key (ALT key under Windows) while defining the selected area.
Supports syntax highlighting and code folding for a large choice of languages.
Fully customizable through a text based configuration file.

Teletext in Germany - Part III

2011-08-17T14:17:00+02:00

This is the third part of the teletext in Germany serie, containing screens from the following channels : rbb Fernsehen, RTL, RTL II, Sat. 1, Sport1, Super RTL, SWR Fernsehen, Tele 5, VIVA, VOX, WDR Fernsehen, and ZDF.

First part : http://www.cambus.net/teletext-in-germany-part-i/

Second part : http://www.cambus.net/teletext-in-germany-part-ii/

rbb Fernsehen :

RTL :

RTL II :

Sat. 1 :

Sport1 :

Super RTL :

SWR Fernsehen :

Tele 5 :

VIVA :

VOX :

WDR Fernsehen :

ZDF :

Teletext in Germany - Part II

2011-08-17T14:17:00+02:00

This is the second part of the teletext in Germany serie, containing screens from the following channels : Kabel eins, MDR Fernsehen, N24, NDR Fernsehen, Nickelodeon, n-tv, Phoenix, ProSieben, and QVC.

First part : http://www.cambus.net/teletext-in-germany-part-i/

Third part : http://www.cambus.net/teletext-in-germany-part-iii/

Kabel eins :

MDR Fernsehen :

N24 :

NDR Fernsehen :

Nickelodeon :

n-tv :

Phoenix :

ProSieben :

QVC :

Teletext in Germany - Part I

2011-08-17T14:17:00+02:00

Faster than expected, here comes the second article of my World of teletext serie : it is dedicated to german teletext this time, and contain 110 screens and one video.

Unfortunately, I have no idea about the actual state of teletext in Germany, as the country totally switched to digital television through a process completed in 2009. However, analog cable and satellite broadcasts will remain available until 2012, and this is actually how I’ve been able to capture these screens from France in June 2011 : using an analog satellite decoder with antenna pointed to Astra (19.2°E).

This serie contains screens from the following channels : 3sat, ARTE, Bayerisches Fernsehen, Channel 21, Comedy Central, Das Erste, DMAX, Eurosport, hr-fernsehen, Kabel eins, MDR Fernsehen, N24, NDR Fernsehen, Nickelodeon, n-tv, Phoenix, ProSieben, QVC, rbb Fernsehen, RTL, RTL II, Sat. 1, Sport1, Super RTL, SWR Fernsehen, Tele 5, VIVA, VOX, WDR Fernsehen, and ZDF.

Overall, Germany was a very interesting country to cover, lots of lovely screens with nice pictures, and spotless design giving access to clear information. Don’t miss the video at the end of this article featuring a skyline with a gorgeous blinking TV tower!

Due to the huge number of pictures, I decided to split it in three parts to allow easier viewing, follow these links to reach the next parts :

Second part : http://www.cambus.net/teletext-in-germany-part-ii/

Third part : http://www.cambus.net/teletext-in-germany-part-iii/

3sat :

ARTE :

Bayerisches Fernsehen :

Channel 21 :

Comedy Central :

Das Erste :

DMAX :

Eurosport :

hr-fernsehen :

Teletext in France

2011-08-16T17:09:00+02:00

Here is the first article of my (hopefully prosperous!) World of teletext serie! As a text mode lover, I’m very sad to see this service being shut down as we are going through digital television transition, and therefore, wanted to keep track of this important part of text mode art history.

In my native country, France, analog television broadcasting is being gradually phased out : a process which started in February 2010 and will be going on through the end of November 2011. Midi-Pyrénées (the region I was born and where these pictures were taken during my last visit there), is one of the very last place to still have access to analog television and thus teletext, but transmission will unfortunately definitely cease on November 8th 2011).

I managed to capture screens from the following channels : TF1, France 2, France 3, France 5, Arte, and M6. Unfortunately I couldn’t capture anything from Canal+ as their teletext service is already totally shut down in the whole country since the end of 2010. Please note that the France 5 serie was taken during daytime, as they share the 5th channel with Arte (broadcasting during nighttime).

For those interested about technical details, and especially how to take pictures of a TV screen without seeing the vertical blanking signal artefact (this will obviously only work with fixed images), here are some recommendations :

Use a tripod : this is absolutely mandatory as it will not only allow you to get clear pictures, but also to maintain cohesion between all pictures in a serie.
Take pictures at night in order to avoid light reflections on the screen.
Enable “Night Mode” on your camera, and disable automatic flash.
Use a very low ISO Speed, I used a value of 100 to take these pictures.

That’s all for now, expect a serie about teletext in Germany pretty soon as I just have to go through all the pictures I took, but I can already tell you : it’s gonna be massive!

TF1 :

France 2 :

France 3 :

France 5 :

ARTE :

M6 :

From WordPress to Octopress

2011-08-11T17:09:00+02:00

As some of you may already have noticed, this site is now powered by Octopress. I made the switch from WordPress two weeks ago now and as I tweeted at that time, I couldn’t be happier of my choice. I still totally agree with this statement : it makes perfect sense for slow paced blogs like this one, and clearly, I’m not looking back. Now that I’m used to Markdown Syntax I realize how painful it was to write articles in TinyMCE, as no matter how hard I tried, it always added some garbage into the markup : something detail oriented people like myself can hardly cope with. By the way, as a matter of fact, the whole WYSIWYG bullshit is one of the biggest lies in computer history.

Also, thanks to Disqus, the comment spam problem inherent to WordPress totally disappeared : ever wondered why all huge WP sites like TechCrunch use an alternative comment system? Of course doing so alleviate database load and allows easier scaling, but moreover, it prevents comment spam to inflate tables and seriously damage query performance.

Here are the key points about Octopress :

Generate static files : definitely lightweight and easy to host
Built on top of Jekyll, so easily extendable and customizable
Integrated Twitter and Google+ sharing buttons, Twitter feed integration
Integrated code embedding plugins supporting syntax highlighting
Beautiful and modern looking out of the box, mobile friendly layout (see below!)
Responsive layout : no need to resize images before posting!

So once again, thanks to Brandon Mathis for Octopress, it’s exactly what I was looking for!

ANSi and ASCii themes for Chrome

2011-08-04T23:53:00+02:00

I’m pleased to announce the release of my minimalist ANSi and ASCii based themes for Chrome and Chromium. I kept everything simple on purpose, as the default theme is very clean, nice, and efficient : these two themes just aim to give a discreet text mode feeling to everyday life browsing, nothing more, nothing less.

The color differences from the original theme can be seen in the status and download bars as well as in the detached bookmarks bar (where I used color #07 (light gray) from the ANSI / EGA palette) and in the text links where color #08 (dark gray) was used.

ChromANSi screenshot :

ChromASCii screenshot :

Downloads : ChromANSi and ChromASCii

20 Years of Linux : From Typewriters to Computers

2011-07-08T12:56:00+02:00

This video is my contribution to the Linux Foundation 2011 video contest : Celebrating 20 years of Linux.

It was created using stop motion technique, and consists of 332 pictures taken with flash enabled in order to add some random artefacts and create grain. The typewriter being featured is an heavyweight Olympia Super Deluxe manufactured in the late fifties, which belonged to my grandmother at the time she was teaching touch typing : amazingly, it still works almost perfectly after almost fifty years!

Moreover, I only used open source software in order to render it, namely :

ImageMagick : used to resize each frame to 1280x720 and convert it to grayscale
eSpeak : a text to speech synthetizer, used to render the soundtrack from a text file
LAME : used to convert the soundtrack to MP3 prior to encoding the video
FFmpeg : used to join everything together and encode the video

For those having troubles to decipher eSpeak output, the text reads as follow :

On August 26th 1991, nearly 20 years ago, a Computer Science student by the name of Linus Torvalds posted this message on Usenet. He was writing a new operating system from scratch, which would later become known as Linux. At the time, a lot of people did not own personal computers and typewriters were not yet uncommon. During the next 20 years, what started as a one man effort became a massive community project involving dozens of thousands people spread all across the globe. Linux has become a modern and robust operating system running on a wide range of hardware and architectures. It is ubiquitous and widely used on servers, clusters, embedded devices, smartphones, and tablets. 2011 marks the 20th anniversary of Linux, and given the fast pace of its development, the best is yet to come! Linux, from typewriters to computers, 1991 to 2011 and beyond!

That’s it! So happy 20th birthday Linux, and to many more to come!

ANSi Social Media Icon Set

2011-06-09T00:09:00+02:00

I’m very happy to announce the release of my ANSi Social Media Icon Set, licensed under the Creative Commons CC-BY-SA 3.0 license and containing icons for the following services : Facebook, Twitter, LinkedIn, Google +1, Digg, Delicious, StumbleUpon, Reddit.

Each icon has been drawn in ANSi with a resolution of 3 lines by 6 columns : pixel equivalent would be 6x6 as even if they are only 3 lines high, I could of course take advantage of the “Upper half block” and “Lower half block” extended ASCii characters (respectively chars #223 and #220, please refer to my Ascii Table if you’re unfamiliar with the ASCii charset). Once converted in PNG, each icon is rendered into a 48x48 bitmap file, which is quite a standard size for social icons nowadays. One interesting point to note is that given the minimalist range of colors used, file sizes are EXTREMELY small : as icons are saved in 1-bit, 2-bit or 4-bit colormaps, their sizes range from 156 to 206 bytes.

Drawing in such a small resolution sure proved to be challenging, and I had to trade accuracy for aesthetics (e.g. the StumbleUpon icon lacks the U of the original). Surpisingly, the 16 colors (EGA palette) limitation was not a problem at all (thanks to iCE colors, allowing to use 16 foreground colors instead of 8), which either means the palette had been really well chosen back in the days, or that icon designers tend to stick to similar colors.

That’s it. Share and enjoy the set and if you use it somewhere I would love to hear about it, please leave a comment.

Download links : ansi-social-media-icon-set.tar.gz or ansi-social-media-icon-set.zip

P.S. : Google guys, if you read this, I would love to draw an ANSi doodle someday!

Node.js ANSI Logo

2011-05-21T00:55:00+02:00

I’ve been reading a lot of stuff about JavaScript lately, and on Node.js in particular, which is such a great project and on top of that, seems to fit all the requirements of an upcoming site I will start working on pretty soon. Meanwhile, it inspired me to draw a Node ANSI logo, which you can download as an ansi file, or converted in PNG for easier viewing.

I’m also working on a small site gathering Node.js resources which I expect to launch during the following weeks.

GoogleCL : Command Line Tools for Google Services

2010-11-22T00:00:00+01:00

GoogleCL, which stands for Google Command Line, is a collection of tools allowing users to access the Google Data APIs. They are particularly useful for those who wish to accomplish automated tasks such as posting news articles on Blogger, scheduling appointments on Google Calendar, adding and editing documents on Google Docs, and so on and so far. On top of that, it also offer the ability to upload images and videos on Picasa and Youtube respectively.

GoogleCL is written in Python (requires version 2.5 or superior) and depends on the gdata Python client library (version 1.2.4 and up, preferably 2.0.x). At the time I’m writing this article, the latest versions of these tools are googlecl-0.9.11 and gdata-2.0.13, and you can follow these instructions in order to install them :

wget http://gdata-python-client.googlecode.com/files/gdata-2.0.13.tar.gz
tar xvfz gdata-2.0.13.tar.gz
cd gdata-2.0.13
sudo python setup.py install

wget http://googlecl.googlecode.com/files/googlecl-0.9.11.tar.gz
tar xvfz googlecl-0.9.11.tar.gz
cd googlecl-0.9.11
sudo python setup.py install

Once everything is setup, you can use the tool to access the following Google services : Blogger, Calendar, Contacts, Docs, Picasa and Youtube.

Here are some examples showing how to achieve some basic tasks :

google blogger post –title "GoogleCL" "GoogleCL is a fantastic tool!"
google calendar add "Job interview for the COBOL programmer position!"
google contacts add "Bobby, bobby@example.com"
google docs edit "Todo List"
google picasa create "Travel Pictures" ~/pictures/sealand/*.jpg
google youtube post –category Education googlecl.avi

However, you might get this error message when trying to perform some of the tasks mentioned above :

Traceback (most recent call last):
File "/usr/bin/google", line 812, in
main()
File "/usr/bin/google", line 798, in main
run_once(options, args)
File "/usr/bin/google", line 542, in run_once
googlecl.write_devkey(options.devkey)
File "/usr/lib/python2.5/site-packages/googlecl/__init__.py", line 546, in write_devkey
with open(key_path, ‘w’) as key_file:
IOError: [Errno 2] No such file or directory: ‘/home/foo/.local/share/googlecl/yt_devkey’

In order to fix that, one needs to create the directory in which GoogleCL can store the autorization credentials :

mkdir -p ~/.local/share/googlecl

And voilà, you’re now able to use GoogleCL : pretty straightforward, isn’t it? Personally, I mainly use it for uploading video content on YouTube, as it allows me to do so using a VPS providing far better uplink speed than my home connection : definitely a great time saving tool I’m glad I came across.

GoogleCL project page : http://code.google.com/p/googlecl/

Decoding Youtube HTTP Error 500 Message

2010-10-26T17:29:00+02:00

I was quite delighted this morning when I discovered Youtube’s custom HTTP 500 server error message for the first time :

**500 Internal Server Error** Sorry, something went wrong. A team of highly trained monkeys has been dispatched to deal with this situation. If you see them, show them this information :

This message was followed by informations encoded in modified base64 for URL applications (base64url encoding), which differs from original base64 by the fact that the padding character ‘=’ is dismissed and the ‘+’ and ‘/’ characters gets respectively replaced by ‘-’ and ‘_’ characters. Assuming the given information data has been saved in a file called error.txt, using sed, it is easy to convert the base64url encoded data into original base64 :

sed ‘s/-/\+/g;s/\_/\//g’ < error.txt > error-base64.txt

Then we can decode the base64 error message :

base64 -d error-base64.txt > result.txt

Now, what’s contained in result.txt you may wonder? Sadly, it just looks like random data, and using the file command to try to determine the file type seems to confirm this statement :

file result.txt
result.txt: data

Sadly, contrary to my expectations, there doesn’t seem to be any hidden message… Screw you, highly trained monkeys!

Zotonic : The Erlang CMS

2010-09-22T01:08:00+02:00

Nowadays, a PHP based CMS is the de-facto standard if you want to publish online content, and it’s easy to understand why when one checks web hosting offers around : it’s the most widely available option and the easiest way to deploy web solutions. However, when site performance becomes a crucial issue, one might want to explore alternative options. Lately, I’ve been starting looking towards Python web frameworks (and more specifically Django), but there are of course plenty of others options as well.

Why not give a try to Zotonic, an open source CMS written in Erlang? Erlang you say? Yes, Erlang, a programming langage created by Ericsson (they make mobile phones as well) and designed to create distributed, fault-tolerant, and real-time applications. It means that of course, Zotonic benefits from the langage features mentioned above, making it a very interesting solution for high-load websites.

Here are the keys features :

Speed : up to 10 times faster than a PHP based CMS
Template language : based on Django templating system
User interface : clean, easy to use, and powerful back-end
Built upon proven technologies : PostgreSQL database and jQuery JavaScript Library

Some screenshots of the front-end and a screencast showing the administration interface :

In conclusion, Zotonic is a very promising CMS even if it is still a young project and as such, lacks maturity and a wide user community. However, I encourage you to give it a try, just for the sake of experimenting with new solutions : I’m convinced Zotonic popularity will rise over time.

Project site : http://zotonic.com