Web Log _

Distributing a Distro (The Indian Story)

Date: 15 November 2005.

In India we have several magazines actively supporting Linux. The support comes through various means, namely by publishing articles on applications, explaining OSS programming principles, through simple walkthroughs and by distributing Open Source software and Linux distros.

Though these magazines do a commendable job in supporting Linux, I find severe gaps and limitations in the way they provide the content. In the current write-up I am dealing with only one aspect of the scenario; i.e. "Distributing a Linux Distro".

Since long we are trying to project the more user friendly face of Linux to our computing user base. The objective is much obvious, to increase the interest of people in this new and open source software pool. (Here the use of the word 'project' is deliberate. If you screw-up something in Linux it will cost a hard time to fix. Any veteran Linux user will agree on this.) The best thing about Linux is that many companies have tried to put-together bits and pieces of software from this OSS pool to create friendly distributions for people to use. But coming to the real usability of the distro, there are many things which are kept incomplete in the free versions. The causes are many- The commercial viability of a paid version comes first and next comes policy decisions, patent issues, copyrights and other legal stuff. Since the release of Fedora project we consistently highlight this issue in different forums and LUG meets. Still I find the support provided by the "Distro Distributing Magazines" oblivious of our discussions and goals.

The first thing I had highlighted earlier is that both "Fedora" editions and "Mandr(ake/iva) Community" editions can be considered no better than beta grade software. It's like a test bed for their commercial offerings which are released later on. For that matter, Mand(ake/riva) initially did not follow "Fedora" route but it in course it became a compulsion consequent to the infamous LG CD Rom problem which plagued Mandrake users.
The second thing is that the software provided in the distro's is in no way complete in the Indian Context. Most of the Mag'CD users come from the student community who want to learn and contribute and people who 'seriously' want to 'fiddle' with Linux. I have used fiddle here because in my view if you are investing on FLOSS technology you would never want the platform to show up bugs during your critical days. If you have good downloads either you will go the 'deb' way (Ubuntu, debian ...) or would take to 'Emerge' this or that (Gentoo). Otherwise if you have to purchase a distro then sky is the limit. Returning to the Mag'CD users, they are either trapped by a thin pipe to the net or they are too skeptical to download OSS stuff just to experiment with. The Mag'CD users can be broadly classified into a hypothetical group set.

1. First comes the broadest group, the group of people who are still testing OSS for their desktop.
2. Second comes the group of people learning different parts of this almost unlimited technology.
3. Third comes the group of people doing productivity work with OSS tools.
4. Fourth comes the group of people in commercial application of OSS tools.

For the first group, the main considerations are: working multimedia stuff and network and internet connectivity along with desktop productivity applications. Usually it is found that Mag'CDs suffer from glaring omissions like absance of full fledged media player playing MP3, WMA and DVDs. Here it should be remembered that Linux plays the widest available media formats. While the Mag'CD would certainly contain the players but due to licensing and other proprietary issues usually lack the able parts of the back-end stuff. Either it is the issue of a stripped xine, missing Gstreamer plugins or wholescale non distribution of Mplayer.

People dual-booting Win/Lin would certainly be happy with Thunderbird mail client along-with plugins and SunBird iCal software which is available for both the platforms. People connecting to internet from Linux would certainly need a dial-up software. I have seen several dial-up software but none matches the configurability and finish of Kppp. However it remains the reality that still several distros have preferred their specific dialers to Kppp even going to the extent of locking the functionality of Kppp through a crippled configuration. This first group of users never dig enough to clean up the mess. If it works right-away it is well and fine and they will use it but if it does-not then they leave it and won't touch it again.

In the desktop productivity software set, OpenOffice is guaranteed to have been provided with. Now comes the disparity problem with Win, while in Linux the Beta of ver2 which generates .odt files is usually found, the much prevalent ver1 software in Win is unable to read it. It is because Mag'CDs have kept their distance from distributing Win versions of OOffice2. The first group of users would be facing a quite ridiculous problem for such a version conflict in distributed software. Are we here to keep complacent till some Win-only magazine come forward to carry the OOffice2 for our dual-booting users?

NOTE: OpenOffice2 is already released. Here is a button for you to click.

In respect to the second group, the sky is the limit. So nothing can be expressed in particular. For people in this category content from both the preceding and following section are applicable.

Coming to the third group of people, they mostly do programming and other content creation work. They are either serious learners in the field or regular users. The people of this group are expected to have familiarity with the OSS tools and their applications. For such people out of box support is not expected. However some common elements should at least find their place. For programmers the omission of proprietary but freely available stuff like compiler kits such as Java Development Kit or programming documentations can be quite nagging. The size of JDK, Kylix, Eclipse, NetBeans, PHP documentation, Java Documentation is such that if a moderate bandwidth is not present, it is quite difficult to procure these developer tools from internet as these packages run in huge megabyte sizes. Even if some stuff is already provided, their complementary parts will be missing, for example distributing Eclipse without CDT, graphical designer, PHP plugin and other freely available plugins or missing development libraries. This scenario can never be generalised because the development tools are so myriad and the developer community is so scattered. However a notional comprehensive set can be created any way. Presently KDevelop and Anjuta are too limited to start with.

The fourth set of people who are expected to either evaluate or deploy commercial applications would be needing third party stuff which they are prepared to procure themselves so there is no discussion on this group. They are expected to be veterans in the trade and they are expected to be able to procure their necessity.

Among the above four groups of people the Mag'CD is not able to satisfy even a single group. But this scenario can certainly change and change for better if the Mag'CD leadership show some foresight.
After the distribution of a non commercial distro, the complementary package for the distro need be supplied. This can be done category-wise considering the size of the available media. The productive stuff specifically targeted to the recently distributed distro should take preference over the fancier stuff distributed as source. Magazine articles on preferred mode of installation and first-run of the application should complement to the distribution. When an application is distributed, the additional supportive plugins and extensions should be provided along with so that the user can take a wholesome view of the application. For example if JDK is distributed, the documentation should accompany and if Firefox/Thunderbird is distributed, the most useful extensions should be included and if Eclipse is distributed then CDT, PHP plugin etc should find their place too.

Distributing Distro after Distro won't help promote the Linux adoption. First the Mag'CD leadership should focus on making the distro complete by distributing additional software targeting the user-base.

- - - - - This Article is under GNU FDL with title as FC text. Contact Author for any queries.