NavList:

Herbert Prinz wrote [7634]:

The sole subject of Bremiker's article that I cited is the numerical
accuracy of a given algorithm in a specific implementation (such as,
say, a work sheet for Dunthorne's formula, broken down into such and
such terms, evaluated by, say, 5 place logarithmic tables).

 

Thanks for providing the pages from “Astronomische Nactrichten”. There might be one kind of explaination there. When you take a look at the “Fehlergrenze” for Dunthorne’s formula you find that these “Fehlergrenze” seem to be lowest for distances between say 60 and 100 degrees independant of the parameter d (difference between altitudes). Also it seems that the “Fehelrgrenze” is rather constant between these two values. It might be a matter of choice to conclude that the “Fehelrgrenze” is lowest between 70 and 110 degrees and then state the limitation in the “Lehrbuch”. I certainly know that this explaination is far from scientific and ask you to comment whether you find my conclusion realistic?

 

Anyway I have found the pages of interest because my own lunar distance programming is done with Bremiker’s method including use of log’s (which is of course not necessary but was a way to compare results from old manuals with examples using log’s).

 

I have also taken a quick glance on the “Encyklopadie der Physik” but as Herbert has pointed out there are no discussions about accuracy to be found.

 

So for the time being this leaves me with a couple of possible explainations from Frank and George together with my own.

 

Kent N