From: Giorgio Pastore on 11 Aug 2010 18:12 Vincenzo Mercuri wrote: .... > any suggestion for a good book (and reference) on Fortran 2003? > > A hasty search on google lead to 3 main results: > > 1. The Fortran 2003 Handbook: The Complete Syntax, Features and > Procedures - by Richard E. Maine, Jeanne C. Adams et al. > > 2. Fortran 95/2003 Explained (Numerical Mathematics and Scientific > Computation) - by Michael Metcalf, John Reid and Malcolm Cohen. > > 3. Fortran 95/2003 for Scientists & Engineers - by Stephen Chapman. My 2 (euro)cents: I have all of three in my shelves. A linear combination of 1 and 3 or 2 and 3 (according to tastes) would be the best. Unfortunately it does not exist (yet). 3 is definitely the most pedagogical but (a part the C interoperability story, already discussed by others) it contains a few inaccurate statements. Probably one does not notice them immediately, but soon or later one realizes that things are a little different. For example the description of control mechanism of a do loop looks reasonable at a first glimpse but actually it is not very neat: it may generate misconceptios. This book contains many examples and problems. Probably the OO part is the less convincing (but in a way this is true for all). 2 and 1 are the most authoritative. 2 is concise but covers every corner of the language. For a very beginner it is probably a little too essential. For the experienced programmer (maybe assuming a previous knowledge of C, see for example the discussion about arrays of pointers) it is probably the best compact reference for the language. 1 is probably the best effort to provide kind of an understandable, annotated and complete version of the standard. Maybe it contains too much for the beginner. All of three are good but share the same weak point, in my opinion, i.e. they cover the syntax of the language but they do not provide enough information for the beginner in order to understand how to *use* the language in a consistent programming paradigm. For such a purpose, in the direction of object orientation, the (now a little obsolete) book by Ed Akin "Object-Oriented Programming via Fortran 90/95" went much better in the right direction. Also resources like http://www.cs.rpi.edu/~szymansk/OOF90/F90_Objects.html are valuable, although some updating to take into account the 2003/08 standard would be welcome. Summarizing my own viewpoint: for teaching Fortran and numerical programming to physics undergraduate students, I use Chapman's book. As reference text for research work, I use mainly 2 and 1 (in this order). Giorgio
From: glen herrmannsfeldt on 11 Aug 2010 19:28 Giorgio Pastore <pastgio(a)units.it> wrote: (snip) > For example the > description of control mechanism of a do loop looks reasonable at a > first glimpse but actually it is not very neat: it may generate > misconceptios. Could you be more specific on that one? -- glen
From: Uno on 11 Aug 2010 21:08 baf wrote: > #3 is a college textbook written by someone who uses Fortran in private > industry. It is really one of the few classic textbooks on "modern" > Fortran (95/2003) currently available. > Really? What makes it classic? -- Uno
From: Vincenzo Mercuri on 12 Aug 2010 01:57 Giorgio Pastore wrote: > Summarizing my own viewpoint: for teaching Fortran and numerical > programming to physics undergraduate students, I use Chapman's book. As > reference text for research work, I use mainly 2 and 1 (in this order). Great, thanks for your hints! -- Vincenzo Mercuri
From: Vincenzo Mercuri on 12 Aug 2010 02:02
Paul van Delst wrote: > With the caveat that C-interop is not covered at all in the Chapman book. Someone please correct me if I'm wrong, but I > just went through my copy again to doublecheck and could not find it covered. Actually I'll need to deepen this topic. Is it fully covered in the other 2 books? If not, could you suggest any other resource? -- Vincenzo Mercuri |