Shuttle costs

Bill wrote:

> According to the CMU software process tests, the linux kernel group
> got a four out of five. This means that technically as a group they
> can bid on NASA contracts.

Four is a damn fine score. Still, I think the space shuttle software
group is rather better. They've had 17 bugs in the last 11 versions of
420,000 lines. That's 17 total. Not per version. Total.

Anyway, it's the process, not the people. I'm sure the Linux group
could do this stuff too if they had too. But the current Linux process
isn't right for hard realtime, safety critical work.

See http://www.fastcompany.com/online/06/writestuff.html for more
info.

> By the way of all the software houses that has ever taken the test
> there are only 16 companies that have got a four or above. This test
> is a requirement for NASA contracts. You must get above a 4.0 to bid
> on contracts.
>
> Bill
>
> --- In spacesettlers@y..., Ian Woollard wrote:
>
>>Ryan Healey wrote:
>>
>>
>>>Would linux type reliablity be good enough?
>>>
>>According to what I read about this, Space Shuttle software has
>>SERIOUSLY low bug rates. Like an average of 10^-5 bugs per line of
>>code. The software industry 'standard' is 10^-2.
>>
>>Linux wouldn't be in the same ballpark. To get that low needs a
>>really seriously heavyweight process. Or atleast that is the
>>only currently practical way to get that.
>>
>>--
>>- Ian Woollard (ian.woollard@t...)
>>
>>"Is a planetary surface the right place for an expanding
>>
> technological
>
>>civilization?"
>>- Gerard O'Neill
>>
>

--
- Ian Woollard (ian.woollard@...)

"Is a planetary surface the right place for an expanding technological
civilization?"
- Gerard O'Neill