SpaceX COTS Demo 2/3 Updates (THREAD 2)

[kirghizstan]

I'm The furtherest thing from a spacex hater but didn't awhile back one of their people say something along the lines that their software has no issues and laughed at the even consideration of it?  (please note the question mark because I could be completely mistaken)

[mr. mark]

Probably an ex employee  Musk has a way of shuffling employees.

[Ben the Space Brit]

I'm The furtherest thing from a spacex hater but didn't awhile back one of their people say something along the lines that their software has no issues and laughed at the even consideration of it?  (please note the question mark because I could be completely mistaken)

Even if there isn't a problem with the software doesn't mean that it isn't a long pole.  Writing software takes a long time; debugging it so it works properly in its operational environment takes even longer.  Worse, the process isn't foolproof.  There's always at least one scenario that an end-user inflicts on your masterpiece that you never considered and crashes the whole shebang like it was a badly-constructed house of cards.


[edit]
Fixed a particularly funny auto-censor change

[kevin-rf]

[edit]
Fixed a particularly funny auto-censor change

Sounds like a software bug

[kirghizstan]

This is the quote I was referencing
http://www.spacenews.com/civil/111021-spacex-soft-review-softens.html

“The SpaceX software presentation was unsettling to the review team,” the summary continues. “There was no Capability Maturity Model Integration (CMMI) accredited capability or process, and the software chief [at SpaceX] said he didn’t worry about errors because ‘there were no mistakes in the software.’ In the Review Team’s experience, this is unlikely.”

[AnalogMan]

This is the quote I was referencing
http://www.spacenews.com/civil/111021-spacex-soft-review-softens.html

“The SpaceX software presentation was unsettling to the review team,” the summary continues. “There was no Capability Maturity Model Integration (CMMI) accredited capability or process, and the software chief [at SpaceX] said he didn’t worry about errors because ‘there were no mistakes in the software.’ In the Review Team’s experience, this is unlikely.”

For anyone interested, you can see the original International Space Station Advisory Committee & Aerospace Safety Advisory Panel open meeting report from September 9, 2011 which contains this quote.

http://oiir.hq.nasa.gov/asap/documents/ASAP_Public_Meeting_Minutes_9-9-2011.pdf

[cuddihy]

This is the quote I was referencing
http://www.spacenews.com/civil/111021-spacex-soft-review-softens.html

“The SpaceX software presentation was unsettling to the review team,” the summary continues. “There was no Capability Maturity Model Integration (CMMI) accredited capability or process, and the software chief [at SpaceX] said he didn’t worry about errors because ‘there were no mistakes in the software.’ In the Review Team’s experience, this is unlikely.”

Please let's not get into a CMMI debate again, there are reasonable arguments on both sides of the issue ( of whether or not CMMI actually adds anything of value) & it's been covered ad nauseum in a previous thread.

Also it's not clear from the quote whether the SpaceX software chief was speaking about *any* errors in the code (which would be laughable) or about larger errors in the software architecture or design (which would be a reasonable position).

Based on the way the ASAP information was originally released, the framing of the quote is suspect.

*edit* spelling corrected

[Lurker Steve]

This is the quote I was referencing
http://www.spacenews.com/civil/111021-spacex-soft-review-softens.html

“The SpaceX software presentation was unsettling to the review team,” the summary continues. “There was no Capability Maturity Model Integration (CMMI) accredited capability or process, and the software chief [at SpaceX] said he didn’t worry about errors because ‘there were no mistakes in the software.’ In the Review Team’s experience, this is unlikely.”

Please let's not get into a CMMI debate again, there are reasonable arguments on both sides of the issue ( of whether or not CMMI actually adds anything of value) & it's been covered ad naseum in a previous thread.

Also it's not clear from the quote whether the SpaceX software chief was speaking about *any* errors in the code (which would be laughable) or about larger errors in the software architecture or design (which would be a reasonable position).

Based on the way the ASAP information was originally released, the framing of the quote is suspect.

I won't debate the merits of CMMI, but I have been with several organizations where a CMMI Level 3 or above certification was a requirement stated in the RFP. If your organization doesn't have a certified quality process, then you can't compete for the contract. I wonder if SpaceX manufacturing has gone through ISO 9001 certification either ?

I'm not saying it leads to higher quality products being manufactured. It's just a requirement for being in this business.

[deltaV]

the software chief [at SpaceX] said he didn’t worry about errors because ‘there were no mistakes in the software.’

That quote almost makes Musk sound humble by comparison! I would only take the assurance of absence of bugs in the software seriously if they had a machine-verified proof that the software met the requirements. Even if they did that you'd have to check carefully for bugs in the requirements. From the ASAP meeting minutes shortly after the "no mistakes" quote:

Another comment was "we don't set requirements, we just do coding." The very essential part of software development is understanding the requirements so as to identify missed requirements, unexplained actions, and possible unsafe conditions.

[cuddihy]

I won't debate the merits of CMMI, but I have been with several organizations where a CMMI Level 3 or above certification was a requirement stated in the RFP. If your organization doesn't have a certified quality process, then you can't compete for the contract. I wonder if SpaceX manufacturing has gone through ISO 9001 certification either ?

I'm not saying it leads to higher quality products being manufactured. It's just a requirement for being in this business.

Except it wasn't a requirement for COTS anywhere in the RFPs or award criteria.

[oldAtlas_Eguy]

There is a world of difference between a failure to identify the requirements and a failure to meet the identified requirements. The requirements specified for the ground control software was found to be lacking and new requirements were identified as the result of an integrated test which in turn will cause the generation of new software and changes to existing (all of which was related to the ground software and not Dragon). So in actual fact the SpaceX software chief’s statement was truthful. The identification of the requirements for operations ground support software is the responsibility of operations personnel not the software coders and testers. It’s operation’s job to know what they need in order to do their job of Dragon operations.

As to CMMI it is an organizational/process model to manage the software life cycle. Almost everyone uses at least some or all of it informally even on very small software projects. For complete in-house projects including complex software projects a formal implementation of CMMI is not necessary as long as the effective management of the items that concerns the model and even others that the CMMI does not cover is done. A software project that has subcontractors that are developing pieces of the software that then has to be integrated together to make it work, a more formal CMMI would be necessary to reduce contractual problems, delays and errors of understanding of requirements. The second case is the normal one that NASA usually is overseeing (multiple software developers [contractors] to develop one software implementation) not the first case where SpaceX is the only software developer and formal management interfaces between the personnel in the software organization is overkill. Most of the CMMI effective software life cycle management is done by the software development tools if they are setup and used correctly without even having to think about performing them for things such as source and version control which is now completely automated by modern development tools (actually these tools have been around and in general use for several decades).

Edit: Spelling

[Antares]

I won't debate the merits of CMMI, but I have been with several organizations where a CMMI Level 3 or above certification was a requirement stated in the RFP. If your organization doesn't have a certified quality process, then you can't compete for the contract. I wonder if SpaceX manufacturing has gone through ISO 9001 certification either ?

Except it wasn't a requirement for COTS anywhere in the RFPs or award criteria.

CRS required 9001 and 9100 and some others.

http://www.nasa.gov/centers/johnson/pdf/418857main_sec_nnj09ga04b.pdf

See pages 29, 31, 76, 147

[Robotbeat]

I won't debate the merits of CMMI, but I have been with several organizations where a CMMI Level 3 or above certification was a requirement stated in the RFP. If your organization doesn't have a certified quality process, then you can't compete for the contract. I wonder if SpaceX manufacturing has gone through ISO 9001 certification either ?

Except it wasn't a requirement for COTS anywhere in the RFPs or award criteria.

CRS required 9001 and 9100 and some others.

http://www.nasa.gov/centers/johnson/pdf/418857main_sec_nnj09ga04b.pdf

See pages 29, 31, 76, 147

The weight of paper documentation and certifications, etc, must outweigh the spacecraft by several times before NASA will let them fly.

[anik]

http://www.nasa.gov/directorates/heo/reports/iss_reports/2012/02212012.html

"04/30/12 -- SpaceX Falcon 9/Dragon launch (target date)
05/03/12 -- SpaceX Falcon 9/Dragon berthing (target date)"

[Comga]

Adapting anik's Current schedule of ISS flight events that would be:

March 9 - ATV-3 "Edoardo Amaldi" launch
March 19 - ATV-3 "Edoardo Amaldi" docking (to Zvezda)
March 22 - ISS orbit's test raising by ATV-3 engines
March -     ISS orbit's raising by ATV-3 engines
March 30 - Dragon (SpX-D) launch
April 3 -    Dragon (SpX-D) capture and berthing (to Harmony nadir) by SSRMS
April -      Dragon (SpX-D) unberthing (from Harmony nadir) and release by SSRMS
April -     Dragon (SpX-D) deorbit and landing
April 19 - Progress M-14M undocking (from Pirs)
April 20 - Progress M-15M launch
April 22 - Progress M-15M docking (to Pirs)
April 25 - ISS orbit's raising by ATV-3 engines
April - Progress M-14M deorbit

[aga]

http://www.nasa.gov/directorates/heo/reports/iss_reports/2012/02212012.html

"04/30/12 -- SpaceX Falcon 9/Dragon launch (target date)
05/03/12 -- SpaceX Falcon 9/Dragon berthing (target date)"

does this schedule mean that kuipers can be the second operator for berthing? or are the soyuz flights moved?

[Antares]

The weight of paper documentation and certifications, etc, must outweigh the spacecraft by several times before NASA will let them fly.

There's a $100B space station up there.  Try explaining you've crashed a spacecraft into it to the people who authorized the activity and appropriated the money.

[Robotbeat]

The weight of paper documentation and certifications, etc, must outweigh the spacecraft by several times before NASA will let them fly.

There's a $100B space station up there.  Try explaining you've crashed a spacecraft into it to the people who authorized the activity and appropriated the money.

I understand why they want that much documentation.

[beancounter]

The weight of paper documentation and certifications, etc, must outweigh the spacecraft by several times before NASA will let them fly.

There's a $100B space station up there.  Try explaining you've crashed a spacecraft into it to the people who authorized the activity and appropriated the money.

I understand why they want that much documentation.

Documentation lots of it or less doesn't mean any system is safer, more efficient or effective.  In my experience, it's useless if it just adds more work and consequently cost and doesn't result in benefits.  At the end of the day, documentation isn't what counts, it's what's done with it that does.

[pippin]

Depends.
Documentation actually is what helps you when something goes wrong and you need some forensic analysis about what it was.