Posts by Geoff Stevens

These guys are clearly humourists

A quick scan of their "100% Service Level Agreement" reveals they've got the company's registered office address wrong, plus a few other mysteries like:

7.3 The Company shall not be obliged to pay compensation in accordance with this Paragraph 7 where the client has been attached by a third party or has not correctly configured the equipment to

And there it ends, after the "to"...

"We recommend that you download copies of our Policy Documents for your records." In case you need a good laugh sometime?

One thing I haven't seen...

...what's the plan if the motor doesn't fire? If it works, the thing blows itself apart, job done. If it doesn't, someone who's possibly feeling a little crestfallen has to take some delicate decisions.

Sorry to be a pessimist, but I don't want to hear of anyone getting their face melted off in this venture. Not even A Orlowski.

I'd like to see a risk assessment and hazard mitigation plan for disassembly on failure.

six sigma != negligible

"Now, it's easy *now* to smugly sit there and say, well, those silly buggers should have built 15m walls, had water tight generators, additional backups to the backup for the backups, put generators on the roof of some building somewhere and been prepared for a 9.0 quake. But before all of this, most people would have scoffed at the chances of a 9.0 earthquake and 10+ meter Tsunami since they go way beyond the 100 year event magnitudes."

Recently, some people may have noticed the world financial system blowing up because of inappropriate application of Gaussian models.

I suggest saying "model shows a six sigma probability of event in 100 years" is the wrong thinking, given the high cost consequences. More severe events are well documented in Japan. Whether it's likely in the 100 year span of your model seems academic the day after it's happened.

http://www.cognitive-edge.com/blogs/dave/2011/03/the_mantra_of_resilience.php

Paris, because she's got an eye for the main chance.