looking east

Yesterday I met Basile, my master thesis supervisor, a very informative and productive meeting, indeed. Having a competent and motivated advisor is probably the most important thing when writing a thesis. So I guess I am lucky: Thank you, Basile!

After having spent the last few weeks combing through the literature and installing Linux distributions, I felt like I eventually needed some more information about what my thesis should be like. And yesterday, I finally learned what my thesis is going to be about. Being a computer scientist, working on a thesis at the faculty of “Lettres” is anything but easy. There is a completely different way of observing and seeing things awaiting me out there. Anyway, so what is my thesis going to be about? Here a (tentative) list:

• Introduction: General introduction into the topic, description of the methodology and related works, definition of the study area
• Linux and China:
  • History of Unix/Linux/Operating Systems/Computing/Open Source
  • History of Redflag Linux
  • Description of Redflag Linux (might be at least partly technical), Localisation techniques, special features, comparison to other Linux distributions
• Meeting a/several developer(s):
  • This is what probably scares me most: the “sociological” part of the thesis. It will mainly consist of an accurate (or even minute) description of how the developer works, where he (or she? well, I guess it will be a male, so let’s stick to “he”, if I’m lucky to find a “she” I’ll come back and update this post) works, how he communicates with the community, how he programs, which tools he employs etc.
• Analysis and Conclusion: depends on what I discover…

“Think of all the activities that must be carried out for any work of art to appear as it finally does. For a symphony orchestra to give a concert, for instance, instruments must have been invented, manufactured, and maintained, a notation must have been devised and music composed using that notation, people must have learned to play the notated notes on the instruments, times and places for rehearsal must have been provided, ads for the concert must have been placed, publicity must have been arranged and tickets sold, and an audience capable of listening to an in some way understanding and responding to the performance must have been recruited.”

Howard S. Becker, Art Worlds (1982), p. 2.

One of the main concern of sociology of science and technology is to show that each product (with knowledge being the product of science) is the result of social cooperation and collaboration at each step of the production process. Understanding the production process in detail is thus crucial in order to retrace which (social) decisions have been taken at which point and how they have influenced the current state of the product. Applying Howard S. Becker’s description about the symphony orchestra concert to Redflag or Asianux will require a more thorough understanding of the activities which need to be carried out to build a Linux distribution with support for Asian languages.

Redflag and Asianux are both Linux distributions, that is a composition of software built from the Linux kernel and a set of additional packages. These out-of-the-box applications allow users to easily setup and install Linux without going through the pain of compiling, installing and configuring all components from source, including the necessary drivers for a particular hardware. Information about how to build your own Linux distribution can be found here:

• Linux from Scratch, a set of step-by-step instructions for building your own custom Linux system, entirely from source: http://www.linuxfromscratch.org
• How to build your own Linux distribution, a IBM article on the topic: http://www-128.ibm.com/developerworks/library/os-lfs/index.html
• ROCK Linux, a Linux Distribution Build Kit: http://www.rocklinux.org

Rereading Howard S. Becker’s Tricks of the Trade on sociological research, I started working on a mind map to illustrate which aspects one might want to consider when studying Redflag Linux. Of course, studying Redflag Linux can mean many things and includes a whole range of possible study areas. One could, for example, study it from a computer scientists’ perspective, maybe analysing the Redflag Linux distribution from an architectural point of view, showing how it is composed of different elements such as a Linux kernel, a GUI and a set of other applications. We could also, for instance, study how localisation and encoding is handled by the system. This kind of analysis would be made from a purely technological point of view, thus applying the rules of traditional technological research.

The research and its outcome will therefore all depend on the methodology applied as well as the definition of the study area. What I will try to apply for my thesis, is an approach similar to the one defined by the sociology of science which Wikipedia defines as follows (useful links to related work can be found here):

“Sociology of science is the subfield of sociology that deals with the practice of science. Generally speaking, the sociology of science involves the study of science as a social activity, especially dealing with the social conditions and effects of science, and with the social structures and processes of scientific activity.”

The idea of my thesis would thus be to show how Redflag Linux became what it is today. In order to make such an analysis, all involved actors and objects need to be taken into account as they might have influenced the development process. My mind map which is still in a very rudimentary state, tries to identify these elements:

However, there are still some important problems in defining my topic. My main obstacle might be language. My Chinese is by far not good enough to read and understand the instructions on the Redflag website, left alone the forum posts and blog entries. I will need to decide what to do about this. Then, another problem seems to be the Redflag development process itself as I haven’t found any evidence of a source code or a open source community. I know, this again might be related to the language problem, but it might as well be that Redflag is developed in a different way than other open source projects. What worries me here is not that the Redflag development model is non-standard, but rather that, if this is really the case, it will be very difficult to find more information about the development process. However, as the Asianux source code is available online and as Redflag claims to be based on Asianux, analysing the Asianux development process might be an alternative.

Intel has announced this year that Midinux, an Redflag-based Linux distribution, will be featuring its Mobile Internet Devices (MID) and UMPC’s (Ultra-Mobile PC’s). Intel is thus planning to embrace an open source model for the implementation of it’s operating system. The choice of using an Asian Linux distribution might be tactically smart: Asia (especially Japan and South Korea) is known to adopt new technological innovations much faster than other parts of the world. Additionally, featuring language support for Chinese, Japanese, Korean, and English will be a great advantage in the MID market. However, competitors are not that far behind: Google has announced few days ago that they will be releasing a first version of the Android SDK on November 12th. Android is a complete mobile phone software stack that includes everything a manufacturer or operator needs to build a mobile phone. Additionally, unlike Midinux so far, Android will be made available as open source via the Apache v2 license.

Unfortunately, I couldn’t find any available source code or downloadable ISO image so far.

See some comments on Midinux here:

• Midinux 2.0 on a Q1Ultra by Clayne Robison (Intel)
• Intel to unveil “mobile internet devices” — Red Flag Linux with iPhone-style interface by Ashton Mills
• moblin.org an umbrella open source project focused on the development of Linux for Intel-based devices
• Red Flag MIDINUX on greghaspants2.wordpress.com

Startup:

Login screen:

Redflag offers a Windows-like Desktop environment, featuring items like “My Documents” and “My Computer” as well as a “Control Panel”:

Oh, and I could configure Redflag to accept my swiss-german keyboard settings with support:

While randomly browsing through the available settings, I discovered an interesting error message:

Unfortunately, Redflag Linux does not allow to change the language once the system is installed. I therefore decided to reinstall a Chinese version in a different virtual machine (see HowTo).

Here are some screenshots:

The following HowTo has been tested on Ubuntu 7.10 using VMWare Server 1.0.3 and Redflag Linux Desktop 6.0.

Downloads

Start with downloading the following components.

Prerequisites:

• Install build-essential (sudo aptitude install build-essential)
• Install the xinetd server (sudo aptitude install xinetd)

VMWare Server:

• VMWare server from http://www.vmware.com/download/server/
• As the canonical commercial repository is not yet available for Ubuntu 7.10, you need the VMWare installation patch from http://knihovny.cvut.cz/ftp/pub/vmware/vmware-any-any-update113.tar.gz

Redflag Linux:

• Redflag Linux 6.0 from ftp://220.194.60.89/pub/redflag/dt6/rf6/RedFlag-dt6-sys.iso
• You might also want to get the Redflag 6.0 Tools disc from ftp://220.194.60.89/pub/redflag/dt6/rf6/RedFlag-dt6-Tool.iso

Installation is easiest when you create a bootable CD image from the Redflag iso files.

VMWare Installation

The following description can also be found in the Ubuntu tutorials.

• Extract all the archives to a directory (tar -zxvf VMware-server* ; tar -zxvf vmware*)
• Run sudo ./vmware-install.pl
• Select all the default options *EXCEPT* do not compile the modules at this point. (Do you want this program to try to build the vmmon module for your system? NO)
• Apply the patch by running sudo ./runme.pl to build the vmmon module (answer YES to compile the proper modules).

At this point, you are ready to launch the VMWare server by typing vmware. Probably, vmware will answer with a message saying that VMWare server is not properly configured. DON’T run the config script as indicated in the message, but remove the folder /etc/vmware/not-configured (sudo rm -rf /etc/vmware/not-configured). Then run vmware.

Redflag Installation in a virtual machine

As I couldn’t find a predefined virtual machine for Redflag, I had to create one by myself. The installation on a external hard disk failed twice, so I decided to first install it on my main hard disk and then copy the virtual machine to the external one:

• Start vmware
• Choose ‘create a new virtual machine’
• Choose the operating system that will run in the machine (I chose RedHat Linux because it’s somehow related to Redflag, but I guess any Linux option should work).
• Memory: Minimal memory requirement is 256M, I created a 500M virtual machine to be on the safe side.
• Disk: Although Redflag Linux should be able to run on a disk of size 5GB, my first installation was disrupted with a size of 10GB because the installation process was running out of space. With 15GB everything ran smoothly.
• Complete the creation of the new virtual machine.

Now you can start installing Redflag on the newly created virtual machine:

• Launch the virtual machine by choosing ‘power on this virtual machine’.
• Type Esc while the machine is loading and choose to boot from CD.
• The CD will then start booting and the installation process begins.

Possible problem:

• Anaconda fails to load. An error message saying exec of anaconda failed: Permission deniedinstall exited abnormally is displayed. This is a known issue with RedHat (which uses anaconda as well), see here.
  • try to launch the CD with the option linux allowcddma
  • if it doesn’t work (didn’t for me either), try linux ide=nodma which worked for me.

I am currently installing Redflag Linux on my Ubuntu 7.10 machine with VMWare. The anaconda loading problem was solved when I typed ‘linux ide=nodma’ on the boot screen. The issue seems to be related to DMA and some Intel chipsets as described here. Unfortunately, this fix makes the installation really slow… so, I am still waiting (since 72 minutes) for it to complete.

For the meantime, I’ll append some installation screenshots.

Installing in graphical mode:

Redflag promises (I’ll probably comment them some other time):

• I just discovered the Redflag forum here: http://www.redflag-linux.com/phorumredflag/phorumredflag
• And I discovered some links to other Linux forums and help pages: http://www.redflag-linux.com/phorumredflag
• Finally, I found an entry by someone who had the same problems installing Redflag in VMWare (although using different versions) as I have: http://www.redflag-linux.com/phorumredflag/phorumredflag/read.php?15,171004

Unfortunately, it’s all in Chinese. I have suspected that of course, but it still comes kind of like a shock :).

I have finally successfully installed a running VMWare server on my Ubuntu machine. The following steps should have then allowed me to install Redflag Linux as a virtual machine:

• Download the most recent version 6.0 of Redflag Linux from ftp://220.194.60.89/pub/redflag/dt6/rf6/RedFlag-dt6-sys.iso (which of course took ages)
• Burn the iso-image to a cd
• Configure a new virtual machine with 500 mb memory and 20 giga hd-space on my external harddisk (I don’t want it to run on my main harddisk)
• Launch the new virtual machine and install Redflag from CD

What really happend:

• The virtual machine started successfully and the cd launched the installation process

• The installation was halted, because anaconda could not execute

A similar problem seems to occur when installing Redhat, so I might be able to find some help on the internet. Redhat’s Bugzilla has a bug entry for this issue, but the suggested booting with ‘linux allowcddma’ doesn’t work in my case. Thus, the question arises: where can I find a forum to discuss problems with Redflag?