Challenges of a Product Manager in the Open Source World – Part Two

When we discussed, how the interaction of customers, partners, and communities influences the product manager’s decisions and his ownership of products, I promised to clarify this by sharing a recent example:

For more than a decade the standard graphical environment on Linux and other UNIX(tm)-like systems, the X Window System (X11, X.org), did run the graphics driver in userland, this means without any special dependency between the graphical environment and the Linux Kernel.

This separation had security implications (well controlled and proven safe though), and limited the flexibility of the graphical environment.

To follow the requirements of modern graphics hardware available on portable devices, desktops and servers, to make X.org more flexible with respect to multiple input and output devices and to improve speed, the Linux Kernel community decided to implement a new technology called Kernel Mode Setting (KMS) beginning with Linux Kernel 2.6.28; this technology is pretty well established since Kernel 2.6.30.

A majority of the X.org community was pushing for this change, supported by major graphics vendors.

While investigating the scope of SUSE Linux Enterprise 11 Service Pack 1, due mid 2010, those partners communicated to us that future graphics drivers would not work without KMS – well knowing that implementing KMS on SUSE Linux Enterprise 11’s Kernel 2.6.27 – while not impossible – would be highly unmaintainable, and thus unwise.

With SUSE Linux Enterprise we are used to follow the fundamental rule of not changing the version of core components over the lifetime of a product.

In this case now, our only choice seemed to be between

• following the rule and living isolated from future development with respect to graphics drivers
• looking for a radical change and upgrading the Kernel version.

Is that the only choice? Or are there other options?

Should we close SUSE Linux Enterprise 11 and cut it off from some open source streams?

Or should we remain on the communities’ trains and try to influence where they’re heading to?

How many trains are worth following?

Which train can be let go without negative impact?

What would you have done?