A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
From: Frank Reed
Date: 2019 Sep 24, 09:54 -0700
Thanks, Sean, for all those details. Very helpful! By the way, all the things that you probably like about DOS programs still exist in most modern operating systems, including Windows and MacOS. They're called "console apps" now, and they produce simple line-oriented text output just like old DOS apps. We could definitely produce a modern version of that old product without having to revert to an obsolete operating system.
John, no, you wouldn't change the UT by one second for a one-second error in Delta-T.
An incorrect Delta-T value will not produce any navigationally significant problem until the error has reached a few dozen seconds. This will be apparent almost entirely in the position of the Moon. There's no problem with GHAs because UT is reset every year or two by tossing in a leap second. This keeps the Earth's orientation aligned properly with the stars when measured by UT. But the relative positions of astronomical bodies will slowly become un-synchronized if Delta-T is wrong. And since the Moon is by far the fastest in angular terms, it shows up in the Moon's position first. The Moon moves on the celestial sphere (relative to the background stars) at a rate of a tenth of a minute of arc in a dozen seconds. So if Delta-T is wrong by two dozen seconds, the Moon's position would be wrong by 0.2' (still insignificant, for everything except lunars).
How likely is it that software coded 25 years ago with an estimated prediction table of Delta-T extending decades into the future would be wrong by 24 seconds or more today? You can go back and look at some of the published predictions. Some predictions suggested Delta-T would be 100 seconds by the year 2020. Instead it's about 70 seconds. So yes, by now, you might expect some issues in some products. It's worth noting that some products were easily modified to include Delta-T in an external data file, probably just plain text. I would bet that the version of the software "behind the scenes" on the USNO web app does, in fact, use a Delta-T file. Unfortunately, that's not the only possible code problem that could kill off this web app after 2035. For example, for the sake of memory usage and computing speed efficiency way back then, the positions of Solar System objects might have been generated from polynomials that are only valid for a specified time frame (I'm not saying that this is the case -- just pointing out that there is more to it than getting Delta-T right).
And let's not forget that there are some serious problems with the computed altitude corrections in that web app. It's not perfect, just useful!