| Inside Lineox Linux |
(Feb. 11, 2004)
DesktopLinux.com interviews Raimo Koski, the founder of Finland-based Lineox Linux. Offering a Linux distribution based on Red Hat's software, Lineox reduces the cost of Linux by eliminating many of the production fees associated with boxed sets and bundled support programs. Koski explains Lineox's unique approach and how customers can benefit from regional support.
Q: Briefly, what is the story behind Lineox? You began in Finland and explain your value proposition as cost savings attributable to VAT, that is repeatable in other counties with similar duties. Can you tell us how this came about?
Actually it is our Finnish book publisher, IT Press, who did most of the product design. Book publishing is much more established industry than software industry and they do the same as e.g. car industry, every bolt and nut counts. Card board box costs, jevel box for disks costs, space for big boxes costs, DVD is cheaper than 7 CD-ROM's, VAT for books is lower than for software (Luxembourg has proportionally the biggest difference, 3% vs. 13%, Sweden by numbers, 6% vs. 25%).
So in Finland we already have a publisher and we hope to get a major publisher for our English installation manual also. Despite LIFF, you usually need an installation manual, if you don't have an another PC with which you can browse LIFF. We will use on-demand printing house at first and maybe try publishing ourselves in Europe, but big publishers have much more effective sales networks.
Using a book publisher is an old trick. Red Hat used it and then Mandrake rose to being the most sold distribution in many European countries thanks to MacMillan already with their second version.
Q: Many companies are now bundling support with the cost of the Linux OS. Your approach would seem to work best with an IT department that is fluent in Linux or looking to hire their own system integrator. Is this who is using Lineox today?
Our previous distributions published by IT Press in Finland have been sold as books and book buyers don't usually expect technical support from the publisher or authors. Lineox is now trying to expand to support, but we feel that support should be separated from products. Take Finland for example. Here the nearest Red Had or Suse office is in Germany and I would be much surprised if anyone there speaks Finnish. So, the best support option would be some local company or consultant combined perhaps with web or email based support and why not the much advertised community support. Anyway, the customer should have the choice.
In comparison to the other leading Linux commercial distro's, are there any other ways you are unique?
Maybe it is the way of thinking like a lazy person. How to produce maximum effect with as little work as possible and do it as simply as possible. Examples of this are taking freely available RH source packages and just compiling them, making scattered 900 MB of documentation files much more usable by just indexing them and creating a menu structure.
Many distribution makers want to do everything by themselves and their way. I have followed Red Hat most closely and found many examples where modifications, which I felt were unneccessary, actually broke things. The original RH 7.3 kernel had backported zlib code which broke zisofs and RH kernels have always been a horrible patchwork, grub still has splash image patch which breaks menucolor commands and RH hasn't patched the documentation to reflect that. I believe that Mandrake is much worse in that respect, but I haven't followed it recently.
Q: You have decided to use Red Hat as the basis of your distribution. Europe has traditionally been more dominated by SuSE. What advantages does this afford Lineox? Why Red Hat?
SuSE's installer is not OSS and I believe that SuSE usage is much lower than their sales figures suggest because they don't offer easy download options. My belief is that Red Hat usage is much higher than SuSE usage in Europe, but sales might be lower.
RHEL offered the best opportunity for us to offer at low cost something that otherwise would be very expensive. We just needed to compile the sources after doing some modifications. Actually it was not that easy, but I have made many modified, Red Hat based distributions before. Additionally, the recent policy changes by Red Hat created an opportunity we felt was just right for us. I would compare that opportunity to Red Hat's decision not to include KDE in RH 5.x versions, which Mandrake then used to become very fast one of the big distributions. (That was much more visible here in Europe because KDE was much more widely localized.)
Q: Another differentiator you have is the focus on documentation. What sets your documentation apart from other distributions?
I must note that it is not a fixation. I just had this idea how to make the vast amount of available free documentation more usable and available. I have ideas how to make it even better, but that would involve so much work, that I think I will work next on some other ideas instead.
Speaking about docs, one important reason for choosing RHEL instead of Fedora is documentation. There still is not even an installation manual for it. Unsupported, undocumented, untested...
Q: How do you determine what programs are 'essential' for end users and make their way into the product? Is the goal to offer compatible core programs only? Will you offer packages tailored to different environments -- ie. accounting, lite office, email only?
I added only Xdialog, apt-get and synaptic (and combined Cluster and Developer Suites to base distribution). Xdialog is used by LIFF installation program, apt-get solves dependency problems and is an easy substitute for Red Hat Network. Apt-get makes it also easy to install additional programs from LEL 3.0 DVD extras directories and from various Internet repositories. Again, the customer should have the choice. Red Hat has made some choices by providing a set of tested and supported packages in RHEL 3.0 and we are greatfull. If a customer wants more, it should be as easy as possible.
Q: What is your vision/expectation for the opportunity of Linux on the desktop?
My personal view is that a desktop distribution should also have a reliable base. For example, RHEL 3.0 is the first and only RH Linux with stable version of htdig. I indexed an earlier version of LIFF with beta version packaged by RH and it occasionally died without a clear reason. With stable htdig the indexing was 10 to 20 times faster and problems were gone. This is quite significant if indexing takes around 12 hours, dies near the end and the next time dies with different error. Stable version produces much bigger index files, but that is compensated by LIFF being on a compressed file system.
Many new Linux users like to experiment and have the latest and the prettiest distribution. I believe they get burned sooner or later by being on the bleeding edge.
(Click here for further information)
|
|
|
|
|
|
|
|
