RE: [boost] Interval Wrinkles

From: Paul A. Bristow (boost_at_[hidden])
Date: 2003-04-16 16:23:11

As an aside, my attempts to persuade manufactuers and designers of balance etc
to transmit to computers (but not display which would be irritating) an extra
'noisy' decimal or binary digit(s) to avoid this effect was greeted with total
incomprehension  :-(

Paul

Paul A Bristow, Prizet Farmhouse, Kendal, Cumbria, LA8 8AB  UK
[phone]   Mobile [phone]
Mobile mailto:pabristow_at_[hidden]
mailto:pbristow_at_[hidden]

| -----Original Message-----
| From: boost-bounces_at_[hidden]
| [mailto:boost-bounces_at_[hidden]]On Behalf Of Guillaume Melquiond
| Sent: Wednesday, April 16, 2003 10:07 PM
| To: Boost mailing list
| Subject: Re: [boost] Interval Wrinkles
|
|
| On Thu, 17 Apr 2003, Ross Smith wrote:
|
| > The interval library isn't useful for computations involving uncertain
| > measurements, since it treats the interval as a square wave with rigid
| > limits, not a gaussian distribution.
|
| You shouldn't be so clear-cut on what interval arithmetic can or cannot
| handle. It all depends on what kind of uncertain measurements you deal
| with. For example, if you know seven (and only seven) digits of your data
| are correct because the sensor is accurate but only displays seven digits,
| then you don't get a gaussian distribution but a square wave. Each time
| precision is reduced due to data sampling (this is the important word),
| this phenomenon occurs. And consequently interval arithmetic is able to
| handle this king of uncertainty.
|
| Guillaume
|
|
| _______________________________________________
| Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
|
|