Comment 2 for bug 827443

Thank you very much for the information on the Xephem database format. I have read the catalog entry for 176P from http://www.minorplanetcenter.net/iau/Ephemerides/Comets/Soft03Cmt.txt using the readdb function and found that it gives different RA and DEC compared with JPL Horizons on-line system. When I try to read directly the orbital parameters from JPL into pyephem using the command below I still get a different answer by about 1 degree:

# JPL orbital elements
# EPOCH= 2453784.5 ! 2006-Feb-18.00 (CT) Residual RMS= .5342
# EC= .1924588793996613 QR= 2.581086276205783 TP= 2453661.856807971
# OM= 346.5854859391472 W= 36.15336596682724 IN= .2378390038753938
# A= 3.196228910655303 MA= 21.15391261693265 ADIST= 3.811371545104822
# PER= 5.71433 N= .172483383 ANGMOM= .03017897
# DAN= 2.66388 DDN= 3.64413 L= 22.7386168
# B= .1403125 TP= 2005-Oct-18.3568080

linear = ephem.readdb('176P/LINEAR,e,0.2378390038753938,346.5854859391472,'+
    '36.15336596682724,3.196228910655303,0.172483383,0.1924588793996613,21.15391261693265,'+
    '2/18.00/2006,2000,g 15.0,2.0')

What is the expected relative error in the coordinates calculated in pyephem with respect to the Horizons ephemeris?