|
|
|
|
|
|
by Jach
410 days ago
|
|
|
Huh, didn't realize that beach didn't like it so strongly. Still, hardly "much maligned"... My impression was that most people enjoyed it, even though it definitely has some weaknesses and could (especially now 15 years later) use a second edition, if for no other reason than to move off CLISP to SBCL. This seems supported by its high reviews on Amazon and GoodReads. (Of course overall I liked it too, and we got the best music video about lisp out of it.) FWIW my own criticisms were largely style-based: too much of it felt like Scheme code rather than Lisp code, even outside of emphasizing/educating about the more primitive features, with lots of things like inner functions, recursion where looping would have IMO been clearer, and so much use of raw car/cdr/caadr/caddr/whatdr instead of more clearly structured structs or classes or just helper functions called intuitive things like get-foo. (The book Calendrical Calculations: The Ultimate Edition uses lists for all its data structures but helpfully creates many functions to both construct and access their parts. e.g. generic dates are a (year month day) list, but the definition and exclusive use of standard-year, standard-month, and standard-day for getting at them would let one refactor it into a class. There are also functions like gregorian-date, julian-date, and egyptian-date that have exactly the same implementation (making a list of the 3 passed params) but serve as 'typed' list constructors, and indeed could be refactored into something that carried along type information without changing other code.) |
|
|
The abstractions used in Calendrical Calculations sound good - and echo what I've seen elsewhere - so, based on your comment I'm now more likely to read it, so thank you for that.