Line 76: | Line 76: | ||
* http://www.devshed.com/c/a/BrainDump/Demystifying-SELinux-on-Kernel-26/ | * http://www.devshed.com/c/a/BrainDump/Demystifying-SELinux-on-Kernel-26/ | ||
* http://danwalsh.livejournal.com | * http://danwalsh.livejournal.com | ||
* [http://www.selinuxbyexample.com/ SELinux by Example; Prentice Hall] | |||
[[Category:SELinux]] | [[Category:SELinux]] |
Revision as of 15:27, 28 June 2010
Background
Karsten 'quaid' Wade gave a keynote[1] at the Southern California Linux Expo 8x in January 2010. Wade's presentation] discussed lessons learned from open source product launches. Two projects were cited as being less popular upon launch: 1. Fedora and 2. SELinux. While Fedora has impressive market share growth, SELinux continues to have a less than popular public opinion. Someone from the audience said, "you have a marketing problem," when it comes to SELinux. We agree.
Later that year, Dan Walsh gave several talks at Southeast LinuxFest on SELinux (when Klaatu uploads the videos to archive.org, we will link here.) There he theorized that because the SELinux "disable rate" was high on Red Hat Enterprise Linux products but not on the Fedora product, SELinux's unpopularity issues persist in the system administrator community.
Marketing Plan Execution
Step #1: Review stats. What do we like about our SELinux market share? What sort of change would we determine to be success in 6 months? 1 year? 5 years?
Step #2: Ask the target audiences why they do not use SELinux. Do customers feel betrayed due to issues during product launch? Are there outstanding myths? Or are there any current product concerns? Collect testimonials from fans of SELinux as we find them.
Step #3: Turn helpful critics into co-developers ie Linus' Law
Step #4: Produce material in many mediums that address the myths and concerns.
Step #5: Deploy materials to various audiences.
Step #6: Keep eyes on stats.
Step #7: Success! Celebrate it and give the credit to the community who made it all possible.
Myths
Let us seek to understand the reasons why system administrators choose not to use SELinux. Here are a few misconceptions we have heard. This is in no particular order. It would be interesting to do interviews of system administrators or online polling to see what is out there and what is prevalent.
- SELinux is a Red Hat/Fedora specific product.
- SELinux is not Free Software.
- No good documentation exists for SELinux.
- SELinux is at least as complicated, if not more complicated, as it was x years ago.
- SELinux is only for customers in the United States government.
- If I do not turn off SELinux then I will be blocked out of my own system.
- I run several virtual machines but I think SELinux is overkill for securing my hyper-visor.
- SELinux does not have a GUI... even if it does, it will not be powerful.
Concerns
Real and legit issues that prevent people from running SELinux shall be parked here.
- none yet.
Audiences
This is where we can reach people who may have a better opinion of SELinux if open communication.
- http://www.lospa.org - The professional society of System Administrators with a large chat community, training events, and a free blog.
- http://www.usenix.org - has many conferences and a professional journal
- http://www.linuxjournal.com
- http://www.linuxpromagazine.com/
- Conferences where Fedora exhibits
- People the Ambassadors encounter
- Podcasts: Mind of Root and FOSS/Linux-specific podcasts.
- http://www.opensourceway.com
Mediums
- Marketing interviews of Dan Walsh and Karsten Wade. Any other key folks?
- Video/Audio distributed under our values of freedom with ogg and archive.org.
- Educational handouts: A bookmark, desk side pamphlet "cheat sheet," or poster.
Documentation
Education is a critical key to the success of this campaign. Here we will list documents that we may end up quoting in an article, brochure, or poster.
Good Free Docs
Not Free
While these docs provide good information on SELinux, we can not quote them in marketing materials due to licensing issues. Maybe the copyright holders will work with us if we ask nice enough.