Customer Reviews

Standard C Date/Time Library: Programming the World's Calendars and Clocks

5 star:		(4)
4 star:		(1)
3 star:		(0)
2 star:		(2)
1 star:		(0)

 

 

I purchased this book out of interest, but then, as always, there comes a time when you reach for a reference book to help out when you need something coded. Well, that's what I did, and unfortunately the book has basic errors in its Julian to Gregorian conversion routine. Minor, I know, but this book is billed as "the Y2K buster" to get on the bandwagon, but the supplied (tested! hah!) routines don't actually work.

I went to the support web site to look up or report this errata. It's gone. Unforgivable. The correct algorithm is available from (the U.S. Naval Observatory's website). The transcription error from this algorithm in Fortran to C is minor, but enough to get me irate. A book that is sold on accuracy and Y2K busting should get dates right. You'd have thought that given a function and it's inverse, that the author would have checked that f(f^-1(x)) == x and f^-1(f(x)) == x. But no...

Apart from that, you'll never use many of the other calendric functions. My faith in this tome is severly shaken.

If you are a computer programmer, don't write any (more) date/time code until you see how Lance Latham does it! This book contains comprehensive, organized solutions to all the date/time problems you are likely to encounter, including tricky ones such as "What is the date N days after date X?," "How many days, exactly, are there between date X and date Y?," and even "What is my client's holiday schedule next year?"

Even if you think you already have an adequate library of date/time code in your shop, don't be too sure until you've tested it using the methods illustrated in Lance Latham's own test programs, included in the CD that accompanies the book. The Year 2000 bug is not the only one infesting date/time code!

For programmers with historical, international, or religious calendar problems to solve, this book is an invaluable reference for a wide range of past and present calendars and timekeeping systems. Lance Latham also includes a section on the forward-looking ISO 8601 standard, which is enjoying increasing usage in Europe.

I recommend this book even if you write computer programs in some language other than C. The routines in this book are easy to translate into other languages, even for programmers with only a "reading" knowledge of C. The correct usage of the routines is exhaustively documented, with all parameters, returns, and limitations spelled out for you. Whichever language you use, following Lance Latham's approach will lead to concise, fast, and reliable code.

My only caveat about this book is that the algorithms underlying the routines are not always adequately explained. Sometimes we are told no more than that an algorithm is a "standard Gregorian proleptic calendar conversion routine." Testing will verify that it does somehow give the correct results, but you may still be tempted to spend hours studying it to figure out how it works.

This isn't merely a list of algorithms (though the number of included date/time programs is extensive); it gives a background and an explanation for the major calendar systems and many of the obscure ones. So I'd recommend it not just for "C programmers" but to anyone who cares about calendars and their vagaries.

This thick book is heavy in research. Not only does it have much hard to find historical data about many of the world's calendars, but it contains exhaustive details on the calculations (with computer algorithms). Although there is some Y2K stuff - this is not the emphasis of the book. Much more importantly this book should be on the reference list of everyone interested in the history and exact details of almost all the world's calendars, ancient and modern. So much stuff that it would take years to research. Simply, a very excellent book for anyone the least bit interested in date and time calculations!!

The author, Lance Latham, is also easy to reach via e-mail and always responsive to questions and comments on calendar and time issues.

Another excellent book along the same lines is Calendrical Calculations by Nachum Dershowitz and Edward M. Reingold, but does not cover as much history and as many calendars as Latham's book.

Review by Ira J. Lund, author of the Universal Calendar Calculator.

I must admit that I'm not a programmer any more but I have read Lance's impressive book as I'm very interested in calendar problems. The book reveal Lance as an extremely competent programmer who writes his program codes in an easily understandable form ready for use for many date/time purposes.

Not often do one see date/time routines combined with the historical background of the calendars and their problems and again, Lance reveal himself as an expert in calendars from many, many countries in his discussion of the calendars. The description of the various calendars shows his deep understanding of the problems his codes solves. Included is a CD-ROM with the entire text of SCDTL, program codes, explanations and many calendar details that are not listed in the book.

It has been a pleasure to read the book and learn from it.

Toke Nørby

I'd like to clarify the bug that was reported in an earlier review. The bug is in JD0_to_gregorian. Don't worry, though. You can just use JD0_to_gregorian_proleptic which works just fine! The strange thing is even the comments say the 2 functions are identical. I suggest copying the code of the proleptic version to JD0_to_gregorian because a lot of other functions call into it, such as JD0_to_date. It's odd that just this one function has a bug (and a very pivotal function at that) because everything else seems to work great.

Otherwise the book is very complete and has a lot of interesting explanations of different calendars. It even has moon functions, although they are limited to about +/- 100 years from AD 2000 (it says the full versions are too processor intensive). Great book!

When I looked at this book, I checked the terms and conditions on using the code, and bascially, those Ts&Cs say "thou shalt not alter a darn thing in the code". I think those who attempt to fix bugs in it should be very wary -- the licence as I read it (but I'm not a lawyer) said that you could do nothing with it. And the licence makes it difficult to read it as well -- see for example a search at ... with key words Latham Standard Date, news group comp.std.c, and a date range from say 2000-01-01 to 2002-12-31 will pick up three different threads mentioning this issue.