UNIX for Dummies Questions & Answers

I have a doubt in date conversion. Can anyone explain the logic behind the date2jd and jd2date in datecalc function specified in this forum. I could not understand what the number are actually representing. Please explain...bcoz i am doing a program with this one..

Note: No need to specifically explain the function in datecalc...in general also...u can explain...i want to know the logic behind it in a logical manner...

I wrote those functions and I cannot explain them. I surfed the web and found date calculation routines in a variety of languages. I converted to ksh and benchmarked them. I took the fastest routines and used them. I gave datecalc a finite range, 1860 to 3999. 1860 is when the US bought Alaska and mandated that Alaska migrate from the julian to the gregorian calendar. Whether or not the year 4000 will be a leap year is controversial.

I manually verified that Jan 1, 1860 works right. Then I used a verification script to generate the dates between 1860 and 3999 by incrementing the day, month, and year. Each date was fed to datecalc for conversion to modified julian day. The script verified that the mjd incremented by 1 each day. The mjd was fed to datecalc for conversion back to year, month, and day and this final output was compared to the original date. So I know for sure that the routines work between the limits I set.

I don't understand the algorithms. But I know that they work. Calendar algorithms are not worth studying. Why not study an interesting algorithm instead? For example: how to accurately calculate x^y in extended floating point using only extended floating point for all intermediate calculations? I've been trying to figure that one out all week. It's much harder than it looks.