Annular Solar Eclipse May 2012 accuracy

I moved my location in Stellarium that the moon would be centered. With that I'm around 30miles South of the center line and 2min later than the predictions from NASA.

Which means somewhere is a bug. it's whether on NASA/Google or Stellarium.

CalSky seems to be almost in agreement with the Espenak application from Nasa... so It appears Stellarium either is not rendering the moon centered properly or something else is amiss. Stellarium has proved extremely accurate for catching the ISS transiting the sun and moon, so I'm surprised it seems "off" for the annular eclipse.

"Which predictions are more accurate? The one from Fred Espenak on the google/nasa website or from Stellarium?"

Let me see, which one has access to better knowledge base, powerful computers and more detailed planetary motion models that are not constrained by time, and which one has to use models optimized for recalculating a planet's position tens of times per second?

The problem may be with Stellarium's orbit models becoming outdated or not simulating some factor, or with its 3D graphics models gaining a bug or two after the last alterations (refraction, extinction, etc.) It can also be the result of a rounding error somewhere.

Well I'm not expecting from a free/donation based software to be acurate but I couldn't find any information if this behaviour of Stellarium is normal or not.

People are probably not using Stellarium to find a location fro the eclipse but maybe to verify or visualize how it'll look. To be that far off in calculations is very confusing if you don't know why.

If it is something that can be fixed, great, if not please make a note somewhere in the documentation, wiki, readme etc. how acurate it is or what not to expect.

Thank You.

This is the closest we have to a precision guide. Incomplete, but something to go on. http://www.stellarium.org/wiki/index.php/Precision

Thank You Matthew. I found that wiki article already but that does not explain why Stellarium is so much off.
Because NASA is apparently using the same calculation methods.

VSOP87/ELP2000-82
http://eclipse.gsfc.nasa.gov/SEgoogle/SEgoogle2001/SE2012May20Agoogle.html

Thus, the only thing left that would explain the differences is a bug in Stellarium.

It appears some people ARE using Stellarium to get a centered eclipse... they're going to Reno. Nasa says that Reno is about 30km too far southwest of the centerline.

I've gotta say that Stellarium has really grown up nicely since I first used it about 5 years ago. Really an impressive application. My great thanks to those who've contributed to its development and support!

For solar eclipses, you need esp. to take DeltaT into account. Currently this is missing from the computations, so earth's rotational state is more than a minute off. This is best noted with solar eclipses...

It disturbs me as well, is solvable and on my long list, but not for this week, sorry. For now, you are better off trusting Espenak.

Kind regards,
Georg