data comes from the devil?

57 views
Skip to first unread message

metasoft

unread,
Feb 24, 2012, 1:19:26 PM2/24/12
to Guerrilla Capacity Planning

DrQ

unread,
Feb 24, 2012, 1:45:46 PM2/24/12
to Guerrilla Capacity Planning
Screwed by a 50 cent connector. (or is it 'unscrewed' in this case?)

The OPERA group claims there was also a mistake in the analysis
involving how the GPS timing was inferred.

Although I was trained as a theoretician, I've also built some
electronics in my time. You very quickly learn that, no matter how
clever your design or how careful you are during its construction, you
are going to get screwed by 17th century mechanics, usually involving
metal. The two biggest gotchas are (1) cold solder joints and (2)
faulty connectors.

To be fair to the OPERA group, with the shear complexity and expense
of their equipment, I have to believe that each sub-team checked for
these 17th century problems during setup, baselining and, more
importantly, when the anomalous superluminal neutrinos were detected.
However, much worse than the gotchas I just mentioned is when those
faults are *intermittent*. Shit happens, I guess. Even when it's
expensive shit. :/

On Feb 24, 10:19 am, metasoft <metas...@yahoo.com> wrote:
> http://news.sciencemag.org/scienceinsider/2012/02/breaking-news-error...

James Newsom

unread,
Feb 25, 2012, 3:11:03 PM2/25/12
to guerrilla-cap...@googlegroups.com
Sometimes Murphy likes to be the middle man delivering the Data from the Devil. :-)

DrQ

unread,
Mar 30, 2012, 12:54:18 PM3/30/12
to guerrilla-cap...@googlegroups.com
The incriminations begin. First, shoot the messenger.

http://physicsworld.com/blog/2012/03/spokesperson_for_the_opera_col.html

DrQ

unread,
Apr 4, 2012, 11:59:03 AM4/4/12
to guerrilla-cap...@googlegroups.com
What that 50 cent optical-fiber connector looks like http://profmattstrassler.files.wordpress.com/2012/04/operafiberplug1.png

Left:   Not screwed in properly (Voila! Superluminal neutrinos)
Right: Contact! (And those superluminal neutrinos disappear)

And in further recriminations, the OPERA team leader has resigned "But the Italian scientist says he remains an active member of the 150-member team."

steve jenkin

unread,
Apr 4, 2012, 11:07:14 PM4/4/12
to guerrilla-cap...@googlegroups.com
This is the difference between Science and Engineering.

- Scientists just look to get a result, or just enough data, to
publish/prove and move on.

- Engineers need to design stuff that they can be sure "Does What it
Says on the Tin",
is able to be checked and maintained and is Safe and Reliable.

The world needs both, neither is more or less important.
And yes, I know that there would've been Engineers and technicians
working on this gear.

But the Scientists would've designed it and been in charge.
They've created this whole mess for themselves by not understanding the
necessity of Quality Control and Quality Assurance.

I've been exactly there: having stand-up fights with Scientists about
Quality and Design issues.
The ones I've had to deal with, were almost to a man arrogant and
insular. Very much "Doctor knows Best" because they had a PhD. I was
never able to get through to them, even after I had to spend months
redoing work because of exactly the errors I'd wanted to check for were
found... Not in any work I'd done, but by one of these Bozos who hadn't
bothered to read the (very clear & simple!) dataset description.

Testing? QA? "We gave the program to some friends and they were fine
with it."...
When I took 4 days to do some structured testing, it was swiss-cheese:
more holes than substance...

In 15 months I did 3 major things for them, completely out of their
range/abilities, that garnered them National recognition, provided a
good deal of that golden-coin of researchers: live media interviews, and
gave the commercial partner a real product to sell after years and years
of broken promises... All the scientists involved either got promotions
that year for 'top of the line' performance grades.

So in the tradition of these things, they learnt nothing from me and
actively worked to sack me.
Yep, I can believe that a connector wasn't screwed in and they missed it...
Wouldn't expect less of them.

DrQ wrote on 5/04/12 1:59 AM:


> What that 50 cent optical-fiber connector looks like
> http://profmattstrassler.files.wordpress.com/2012/04/operafiberplug1.png
>
> Left: Not screwed in properly (Voila! Superluminal neutrinos)
> Right: Contact! (And those superluminal neutrinos disappear)
>
> And in further recriminations, the OPERA team leader has resigned "But
> the Italian scientist says he remains an active member of the
> 150-member team."


--
Steve Jenkin, Info Tech, Systems and Design Specialist.
0412 786 915 (+61 412 786 915)
PO Box 48, Kippax ACT 2615, AUSTRALIA

stev...@gmail.com http://members.tip.net.au/~sjenkin

DrQ

unread,
Apr 8, 2012, 12:22:36 PM4/8/12
to guerrilla-cap...@googlegroups.com
Skipping the cost of doing performance analysis or capacity planning should always raise the question, "At what expense?"

Similarly, how does the cost of that 50 cent connector compare with the expense of not checking connections properly? (i.e., the measurable cost of the experiment). Here are some published numbers:
  • Cost of OPERA in SwFr 71 million = $77 million USD. Eh, let's call it $100 million USD in round numbers.
  • Cost of OPERA = 1/100th of LHC @ ~$10 billion USD (in round numbers).
  • Cost of OPERA = 1/4 of a single NASA Space Shuttle launch @ ~$400 million.
  • Greece 2nd bank bailout: €130 billion ($171 billion) or 2000x OPERAs.
Reply all
Reply to author
Forward
0 new messages