I think that is a good choice as except for very high precision
measurements the assumption is that the errors in the constants are
uncorrelated. In the case where the measurement/calculation is sensitive
to the correlations the user should be carefully checking those
correlations and would ha
Now that I look through
http://physics.nist.gov/cuu/Constants/RevModPhys_80_000633acc.pdf, I've
decided not to implement uncertainty (except as an attribute of the
constant object), since taking into account the correlation coefficient
between any two constants is more difficult than I thought
Yes, I know. But it's not implemented as Sage objects.
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
U
Scroll back in this discussion. The math/arithmetic problems, other than
getting good symbolic expression, have been solved in the "uncertainties"
package ( http://pypi.python.org/pypi/uncertainties/) you mentioned before.
Jonathan
On Friday, March 30, 2012 7:11:20 PM UTC-5, Eviatar wrote:
>
>
The auto-update should be easy to implement. However, now that I've looked
into error propagation, an interval seems like a very wrong way to
represent uncertainty in physical measurements (apparently Mathematica does
*not* do this; I don't know why I thought that). So the only alternative I
se
Sorry that I wasn't clear. I meant the actual object representing a number
with uncertainty; for example, 4.0 +- 0.25. It's not quite an interval,
since operations are handled very differently.
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this grou
David Roe writes:
> I'm confused by what you mean by "error propagation class." Should you just
> inherit from Exception, ValueError, etc?
I imagine he means in the sense of something like interval arithmetic,
except with fuzzy error bounds (i.e. probability distributions) rather
than strict int
Evitar,
As the code I use for gaussian error propagation is about 5 lines, this
seems like major overkill. I still think auto-update of a list of physical
constants is much higher on the list of things scientists would use. As
someone who is not at all expert on the complete OO structure of S
I'm confused by what you mean by "error propagation class." Should you
just inherit from Exception, ValueError, etc?
David
On Fri, Mar 30, 2012 at 03:17, Eviatar wrote:
> Hello,
>
> I'm working on implementing an error propagation class, which is quite
> easy. However, the type system/coercion
Hello,
I'm working on implementing an error propagation class, which is quite
easy. However, the type system/coercion is confusing me. Any ideas about
what the parent should be?
Thank you.
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, s
By the way, this ticket might be of interest:
http://trac.sagemath.org/sage_trac/ticket/9763
-Keshav
Join us in #sagemath on irc.freenode.net !
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@g
On Sat, Mar 17, 2012 at 7:23 AM, Volker Braun wrote:
> Just to rephrase things, there should be
> a) a new ring analogous to RIF that does error propagation (instead of
> interval arithmetic)
> b) the elements of this ring should have units attached and check that they
> match in ring operations
On Sat, Mar 17, 2012 at 10:08 AM, David Roe wrote:
> That sounds really useful actually. I still don't think it's enough for a
> GSOC project; is there something related we could add?
> David
This sounds like arithmetic with *random variables*, which is
something a probabilistic should
implement
That sounds really useful actually. I still don't think it's enough for a
GSOC project; is there something related we could add?
David
On Sat, Mar 17, 2012 at 10:23, Volker Braun wrote:
> Just to rephrase things, there should be
> a) a new ring analogous to RIF that does error propagation (inst
Just to rephrase things, there should be
a) a new ring analogous to RIF that does error propagation (instead of
interval arithmetic)
b) the elements of this ring should have units attached and check that they
match in ring operations
On Saturday, March 17, 2012 9:45:58 AM UTC-4, Keshav Kini wro
Keshav Kini writes:
> Jonathan writes:
> Is this sufficient for your purposes? You could store a Gaussian
> distribution with mean m and standard deviation s as RIF(m-s, m+s), say.
Wait, what am I saying... of course standard deviation doesn't behave in
the same way as Real Interval bounds when
Jonathan writes:
> As a scientist who would actually use this for research if it worked well, I
> have some thoughts. The thing that would make me use this instead of my
> personally maintained list of constants is if there were a command to update
> the constants automagically from the NIST datab
17 matches
Mail list logo
