Storage, chip slingers pledge allegiance to Linux, open source

SanDisk, Seagate, and Western Digital...

Wait, does that mean decent linux drivers for their disks somewhere in foreseeable future???

oh, and http://www.theregister.co.uk/2010/08/05/wd_windows_only_diagnostics/

Any ecosystem...

gains resilience through diversity, i.e. monocultural tendencies are dangerous to long-term survivability.

While superficially attractive, the desire for the OS equivalent of a monoculture is wrong-headed. Windows OS is a real world demonstrator---albeit a pretty crap one---of the dangers of wanting everything to 'just work' by being 'the same'---it simply opens the door to large scale attacks and the rapid propagation of faults generally.

A highly diversified system where there is always an alternative way of getting something done is inconvenient and apparently 'inefficient', but only when taking a short term or localised view.

The old fashioned 'Unix' way of small open source tools doing one thing simply and well, that can be strung together in any number of arbitrary combinations is still one well worth hanging on to.

It allows anyone to sit down and craft the tools they need, if what is on offer commercially or in house isn't quite scratching the itch. And it keeps the ecosystem lively and diverse.

If that applies at the local level it applies just as much, and even more importantly on a global scale. We really don't want to make ourselves even more vulnerable to pwnage than we already are.

There are no panaceas, but I would put up with a fair degree of 'inconvenience' for the sake of a robust diverse application/OS ecosystem that actually inconveniences the state/criminal scum and assorted other ratbags by an order of magnitude.

For What it's worth--trying again.

Apologies--the editing period timed out before I finished the recently withdrawn post.

1. The design, size, complexity of systems, and the need for constant patching of security holes and programming defects have created an untenable situation. System configuration management is impossible and proof testing only good until the next patch. We see evidence of the problem in the increasing reports of major failures of security and reliability here in the Reg.

2. Linux is an obvious alternative path. But open source is not a panacea. Some of the security boffins I work with opine that because of multiple sources generating, distributing and patching code, open source software is more vulnerable to hackers. It just hasn't yet been a big enough target to draw their fire.

3. In any case, I believe that if current trends continue the barrier of sunk investment in legacy software is going to become less relevant. If system stability, security, and reliability continue to degrade, at some point the economics will change sufficiently to force a major change.

4. What that change will look like is impossible to predict. To large extent it will depend on whether any the device hardware alternatives for "After CMOS" emerge as practical. Otherwise, at some point, as practical limits of scale and reliability are reached, the cost of business as usual on the software side will become prohibitive. Someone will come up with better operating systems options than any we have today.