Observing and recording is great and all, but if you want to produce an accurate ephemeris or make an accurate orbital determination, time is important. For calculating my own ephemeris, I usually start by converting to the Julian Day Number and from there the number of centuries since the reference epoch, taking into account delta-T. For the current period from the year 1900 up to 2070 I use a polynomial function I derived to approximate delta-T:

1729.71-(2.29234*year)+(0.000728854*year^2)

I find that attains a suitable amount of accuracy to within a few seconds. NASA gives a series of polynomial expressions each valid for only a 20 year interval for the 20th century, I decided instead to create a single function as I find the above is accurate enough for most purposes, even telescopic observations of the moon. How about you?