Effective Java 157
| Effective Java Programming Language Guide | |
| author | Joshua Bloch |
| pages | 252 |
| publisher | Addison Wesley |
| rating | 8/10 |
| reviewer | Ben |
| ISBN | 0201310058 |
| summary | 57 pieces of Java wisdom. |
Introduction
Effective Java is a book very much in the style of Scott Myers' earlier C++ "Effective" series. The book contains 57 individual snippets of Java wisdom, broadly categorised into 10 sections including Classes and Interfaces, Exceptions, Threads and Serialisation. Scott Myers' books are classics; I was interested to see how this would compare.
The author, Joshua Bloch, has been involved in writing many industrial-strength Java libraries. His background is very much evident, in this, his first text. He consistently demonstrates the virtues of favouring libraries, clean APIs and advance design. I found the author very readable, and able to make a convincing argument, even in his more 'controversial' pieces. As with Scott Myers' books, there is a real-world, rather than purist approach taken to the language, with most of the code examples having a real-world feel to them. This is a breath of fresh air when lots of programming books tend to use more contrived examples.
The items
The author has endeavoured to keep the book accessible to less-experienced programmers throughout, while providing food for thought for the more advanced reader. For the most part this is succesful, but a small percentage of articles tend toward the simple side. Examples include 'Minimise the accessibility of classes and embers,' 'Write doc comments for all exposed API elements,' and 'Know and use the libraries.' We've all heard this advice many times and I don't feel that these add value. The vast majority however, are pitched at the right difficulty level. The selection of items is well balanced and broad, although unfortunately there are none pertinent to GUI programming.
Many of the articles are fundamentally based on known design patterns and idioms. Although a useful index to these patterns is included, I would have liked to see the virtues of design patterns summarised and demonstrated to a greater extent, perhaps in the introduction.
I was highly impressed with all code examples. Where used, they are consistently short, relevant and concise, with more verbose examples included on the website. The chosen code examples only ever assist in explaining complex concepts clearly.
The strongest area of the book for me was the section on threading. The author clearly demonstrates, for instance, how overuse of synchronised methods can lead to deadlock. He also provides food for thought on how the thread scheduler might trip us up. A section on moving from C constructs, which initially struck me as an odd category, proved very interesting and thorough. 'Replace enum constructs with classes' is a particularly interesting item, demonstrating the fragility of C enums, and indicating why the often-used replacement in Java (a bunch of public static constants) suffers from the same failings.
In conclusion
Ideally I would have liked to see some of the thinner items removed, and perhaps replaced with a section on the GUI libraries. I also liked the short prose sections, and thought the author could have spent more time setting out his stall before launching in to the items. Having said this, this is one of those rare books which could help a good programmer become an excellent one. Many of the books currently out there are aimed at either the beginner or the guru, and this book fills a gap.
I find this style of book very useful, in that I could foresee meeting the vast majority of the described situations at some point or another. So long as you aren't looking for tips to help you with your GUIs, this title is more than worth the investment.
For anyone interested, those sections in full:
- creating and destroying objects
- methods common to all objects
- classes and interfaces
- substitutes for C constructs
- methods
- general programming
- exceptions
- threads
- serialisation
You can purchase Effective Java from bn.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.
Re:Practical Applications (Score:3, Insightful)
The abortion that is
Really, I would say that it depends on the environment. If the back end is Oracle or DB2, I'd probably go with Java; if it's SQL Server I'd be more inclined to go with
Re:Practical Applications (Score:1)
Besides, if your skill set is based around MS products why not look into C#. I can tell you from experience that C# is a hell of a lot easier (or maybe more intuitive???) than Java. I like Java a lot, but damn is it painful to do GUI based apps. If you do decide to go the Java route take a look a the Eclipse IDE, it's not quite like using an MS IDE but it's clean and stays out of the way (cough**Borland**cough).
Re:Practical Applications (Score:2, Informative)
I'm in a similar situation, developing and maintaining a data access application in VB6 that has reached 'critical mass', has gotten too big for it's britches, and needs to be rewritten from the ground up.
I've been doing a lot of thinking about what it will be rewritten in, and so far C# and VB.net are leading the pack. All Java has to offer me is some platform independence, which is pretty much irrelevant in my particular case, and I doubt it would be worth headaches like throwing out all the code that's written.
Anyways, this is all contingent on the dorks in marketing not selling imaginary software that I have to write 2 weeks before shipping, that I may have time to make the actual software work.
Re:Practical Applications (Score:1)
Re:Practical Applications (Score:1)
I like Java a lot, though primarily for use in building web applications, not stand-alone client applications.
Re:Practical Applications (Score:2, Informative)
Visually, there are lots of similarities with VB, but with more power. You can go under the hood and get as low-level as you want, short of device drivers (where at least a small part must be written in C/C++).
As a bonus, data-aware controls actually work! (even MS recommends not using them in production VB code.)
There is a version (C++ Builder) that uses C++ instead of Pascal as the basic language, and another package, Kylix, that can compile Delphi/C++B source code to run on Linux.
Borland also hosts a fantastic developer community (especially newsgroups), and most components, even commercial ones, have source available. Look at Torry's [torry.net] for thousands of components/tools/apps.
Re:Practical Applications (Score:1, Informative)
Don't learn Java. Learn OOP, and implement what you learn, as you learn, in Java. That is the best advice no one ever gave me. When you learn OOP, it doesn't matter what Java-like language you learn, as the real skill will be portable to other languages, like C#, SmallTalk, Python and Ruby. Let me explain.
I recently decided to go into the field of programming, and probably don't have 1/10th your development experience. However, my research skills are equal to anyone that I know, and they've certainly helped me in getting a grasp on the fundamentals of software engineering (or problem solving as I've learned to call it).
I chose Java because of its touted strengths, and because I wasn't sure of my target platform yet. I soon understood that as an introductory language, it has some conceptual hurdles that may get in the way when first learning to program. After a while, I began to understand the basics of programming by supplementing my self-study with older languages like BASIC. Essentially, diving straight into Java without first learning the standard structures of programming (sequence, selection, and repetition) was a handicap. However, for someone who already knows Visual Basic, such as yourself, the biggest hurdle will be the concepts of Object Oriented Programming.
If you think you "know" Object Oriented Programming but have only implemented it in VB, then I would suggest that you may not know everything that you need to program in OOP properly. There are plenty of Java programmers out there, but looking at some of their code, I see a definite lack of understanding in how to implement OOP. I see procedural programming techniques being applied to a language that will allow for it, but the developers lose the real advantages of using OOP.
I would recommend that you get a "beginners" book on Java (John Smiley is an excellent resource with his Learn To Program Java) to become familiar with the mechanics. If you prefer, you can try a Java in 21 Days/24 Hours type book. It doesn't matter, because if you follow my advice, you will actually learn to program OOP which is by far the largest hurdle, not syntax/grammar/classes. The most beneficial book I've found on the subject is from a master on the subject:
Sams Teach Yourself Object Oriented Programming in 21 Days [amazon.com]
by Anthony Sintes
ISBN: 0672321092
This book will not only teach you how to program OOP, it will teach you to do it the right way. This is done by showing you how not to program OOP, and immediately showing you an alternative.
The author has some serious credentials: "Tony Sintes has worked with Object-Oriented Technologies since 1995. In that time, Tony has been part of many object-oriented development efforts. Currently, he works for BroadVision where his main responsibility as team mentor, building the skills of less-experienced developers. He brings his years of experience and ability to teach to different projects in order to guarantee their success.
Tony Sintes has written for JavaWorld, Dr. Dobb's Journal, LinuxWorld, JavaOne Today, and Silicon Prairie, where he produced a highly-regarded monthly column on Object-Oriented Programming. Tony currently writes JavaWorld's monthly Q&A column."
About the book: "Sams Teach Yourself Object Oriented Programming in 21 Days differs from other OOP books in two main ways. Many classic OOP books are designed for software engineers and teach at an academic level. Sams Teach Yourself Object Oriented Programming in 21 Days presents accessible, user-friendly lessons designed with the beginning programmer in mind. Other OOP books work to present both OOP and to teach a programming language (for example: Object-Oriented Programming in C++). Although Sams Teach Yourself Object Oriented Programming in 21 Days uses Java to present the examples, the book is designed to present concepts that apply to any OOP environment."
Days 1-7 cover the core OOP concepts with every other chapter covering the implementation of the concepts in the previous chapter and a boatload of do's and don'ts. For instance, in Chapter 2 you learn about Encapsulation theory (more importantly, what NOT to do). Chapter 3 you write code implementing Encapsulation...and so on.
Days 8-14 covers how to apply the OOP you've learned using design patterns like MVC.
Days 15-21 take the things you learned in the past 14 days, and make a GUI-based application. You apply the MVC pattern in making the app. Very cool.
Re:Practical Applications (Score:2)
Re:From the people who brought you: (Score:5, Insightful)
java is like any other language, its a tool to get your job done, and contrary to popular belief, it gets it done rather nicely.
of course i don't mean to wax religious here... there's enough of that here.
Re:From the people who brought you: (Score:2, Informative)
I've been required to use Java non-stop from late 1995 until quite recently for school, internships, and then for my career. I've used every version of Java, and every API and just about every class available in the SDK. I've written applets, mathematical programs, games, medical database apps for the department of defense, class projects, enterprise-level call center applications, and just about anything else you could possibly conceive.
I've found in my experience that Java is good when it's highly specialized and kept very small and simple (yes, Java is good for APPLETS!). For enterprise applications, it _sucks_.
It's not that the language itself is bad (although it certainly has some gaping holes). It has more to do with the poor (buggy) implementation of the APIs and the real-time interpretation. Java isn't nearly as portable as they would have you believe (writing a 100-line applet and then running it on two different web browsers will tell you that.) Additionally, in my opinion, Java would have been much more useful if had been designed as a compiled (not interpreted) language.
In my current job, I use a lot more C++. I like it. It does what I tell it to do, and I don't have to second-guess the tools every time there's a problem.
Re:From the people who brought you: (Score:3)
Interesting observation...
Why don't you give us some specific examples to back up that sweeping statement?
Re:From the people who brought you: (Score:2)
Which APIs? Which bugs? Nothing is bug free. Also, your comment about "real time interpretation" is off base, see below.
Java isn't nearly as portable as they would have you believe (writing a 100-line applet and then running it on two different web browsers will tell you that.)
Applets apply to Java portability how again? Are you under the impression that some significant amount of Java development today relates to applets? Answer: not really. Further, if you use a decent JRE in both browsers you'll find you have no applet portability problems either.
Additionally, in my opinion, Java would have been much more useful if had been designed as a compiled (not interpreted) language.
Now you are exhibiting true ignorance for the world to see. There are many traditional, ahead of time, Java compilers. gcj, TowerJ and Jet come to mind. Look back at the recent article on x86 Linux numerics. Check out the linked article [coyotegulch.com] from that post - note gcj is only 9% slower than g++. Supposedly gcj 3.3 and 3.4 both improve gcj speed substantially over the 3.2 compiler used in that article...one guesses the more mature g++ will see less speedup.
In short, your reasons for liking C++ over Java don't hold water except for verbosity. ;-)
(BTW, I'll give you a strong reason for disliking C++ - you can't turn on bounds checking in STL arrays using operator[] without hacking STL. Ugly.)
Re:From the people who brought you: (Score:1)
I remember the AWT and then, especially, Swing being extremely buggy. Trying to get a combobox to display correctly when you put it at the bottom of a window is one of the many problems I can remember right now. I understand that the APIs are still "new" (relative to the STL?) and constantly under development, but that makes using Java a poor management decision.
Applets apply to Java portability how again?
The Java platform is ported both to operating systems and to browsers. Using applets as an example of the poor portability of Java bytecode across different Java platform implementations is a perfectly good example of the over-hyped "portability" of Java. "A lot of people don't use Java applets" is not a good rebuttal.
Now you are exhibiting true ignorance for the world to see.
You should know that when you attack the person on the other side of the discussion, you're making it easier for the more intelligent readers to dismiss your arguments. As I said (and as you quoted), Java was designed to be an interpreted language. Yes, there are "real" Java compilers out there, and they're getting pretty good. I'm excited for what the future has in store for the maturity of free Java platform implementations. But the Java compilers available have never been good enough for management to say "let's use gcj instead of Sun's JVM".
Notice I never said that "Java is slow", although you spend an entire paragraph trying to convince me otherwise. The dynamic recompilation (JIT) that is available in most Java compilers is certainly a good thing, and compiled java is even better. But memory consumption, API compatibility (!), and yes, even speed, are things that need to be considered in a language that's going to be the foundation of a substantially large application.
I'll give you a strong reason for disliking C++ - you can't turn on bounds checking in STL arrays using operator[] without hacking STL.
Are you saying that C++ is a bad language because the STL doesn't do bounds checking on arrays?? I can personally think of several reasons for disliking C++, but that's never been one of them.
Re:From the people who brought you: (Score:1)
Which is a very sensible sentiment. The problem is that, because Sun has spent millions of dollars marketing it, management now thinks that it is the only tool to get the job done, when there are many other fine tools, such as C, Tcl, Python, etc... that each have good features.
That's what I find frustrating about Java. That and the fact that I like free software, and the Java comunity doesn't seem to have the strongest of relationships with those ideals.
Re:but (Score:2, Informative)
Except that verbosity and (contrived?) structures goes a long way towards keeping source maintainable. By requiring this verbosity and not to mention strongly recommending the Sun Java coding standards + Javadoc, code becomes magnitudes easier to read. Sure, it adds work and a bit of pain to the original coder/programmer, but adds substantially more ease of use for those down the line. Don't get me wrong it's still very easy to write hard-to-read-and-or-maintain code, but it's a bit easier to go the correct route.
But if that's how you get your kicks, gagged and bound with Sun sticking it to you, then so be it...
That depends on what you are referring to. You are right that Sun tends to stick it to the end user now and then and ultimately Sun can be said to be Java's worst enemy. For example, Mac OS X trying to do some amazing things with Java showing how it can be used viably, getting very little support from Sun.
After all that's said, I too still prefer C++ (even C#) over Java, but just not for the reasons mentioned.
Re:An intelligent reply? What gives?? (Score:2, Insightful)
whenYouCan(keepReferencingYourObjects)->LikeThi
Actually, it should be:
whenYouCan(keepReferencingYourObjects).LikeThis
remember, no pointers... *wink, wink*
Anyways, I have yet to see anything that bad (I know you were making a point), nor have I seen anything similar that can't be fitted within 80 columns on 2 lines. You can't seriously tell me you have never seen a C/C++ line that long. You can always write horrible code no matter the language, blaming it on the language itself is just an excuse.
What are your reasons for preferring C++ (or even C#) over Java? Speed, standards-compliance, readability, lower memory usage, saner language constructs, not controlled by Sun?
Speed - used to be, IBM is doing some amazing things with their JIT compilers. I think someone else mentioned they were getting 90% of native code. So this is no longer an issue for me. I've even writing some fairly advanced 3D apps using GL4Java.
Standards-compliance - well... not sure what to say here. There's not ASCI Java if that's what you mean. Never been much of an issue for me.
Readability - as I have been arguing above, I prefer Java's... just me I guess?
Lower Memory Usage - For sure! Java's memory footprint is horrible. Price you pay for a virtual machine. I do hate this.
Saner Language Constructs - see Readability.
Not Controlled By Sun - someone has to control it, I'd say they are doing a decent job of maintaining the API. They also seem to do a fine job of listening to the needs of the users. I think sun does a fine job with the API itself, they need to improve their application of it (Sun's JVM pretty bad, etc).
My main reasons for preferring C++...
- lower level, like the control I think.
- Sun's Swing and AWT API's suck IMHO.
- Java is very large API, I get a bit overwhealmed at times.
Re:An intelligent reply? What gives?? (Score:1)
Shouldn't that be :
?Smart guys, those parser builders. Not to mention that if *that* is your problem, you can pay a little performance penalty and give names to all those intermediate objects.
Re:An intelligent reply? What gives?? (Score:2)
whenYouCan(keepReferencingYourObjects)->LikeThi
This code is a horrible violation of the
Law of Demeter [neu.edu]. If you're doing this, formatting is not your biggest concern.
Re:An intelligent reply? What gives?? (Score:2)
Re:An intelligent reply? What gives?? (Score:1)
Secondly, I'd be interested in knowing exactly what 'powerful and useful' functionality Sun is holding back?
Or 'Maintainable Perl'. (Score:1)
Java is one of the most productive languages in the world. End of story.
I Stand Corrected! (Score:1, Offtopic)
Re:bravo! (Score:2)
Re:bravo! (Score:1)
thanks benjiboo (Score:1, Interesting)
Effective Java? What is this Java? (Score:3, Funny)
Do not talk of this Java Sanctuary. There is no such thing.
Flamebait? Oh, come on (Score:2)
Re:Flamebait? Oh, come on (Score:1)
Re:Effective Java? What is this Java? (Score:2)
[Sorry I couldn't keep your Logan's thing going...]
Where's the TOC??? (Score:1)
Re:Where's the TOC??? (Score:1)
I tried to post it. I really did. The slashdot filters just wouldn't let me. Seems like 37 characters per line should be plenty. Whatever.
Re:Where's the TOC??? (Score:1, Informative)
1 Introduction
2 Creating and Destroying Objects
1: Consider providing static factory methods instead of constructors
2: Enforce the singleton property with a private constructor
3: Enforce noninstantiability with a private constructor
4: Avoid creating duplicate objects
5: Eliminate obsolete object references
6: Avoid finalizers
3 Methods Common to All Objects
7: Obey the general contract when overriding equals
8: Always override hashCode when you override equals
9: Always override toString
10: Override clone judiciously
11: Consider implementing Comparable
4 Classes and Interfaces
12: Minimize the accessibility of classes and members
13: Favor immutability
14: Favor composition over inheritance
15: Design and document for inheritance or else prohibit it
16: Prefer interfaces to abstract classes
17: Use interfaces only to define types
18: Favor static member classes over nonstatic
5 Substitutest for C Constructs
19: Replace structures with classes
20: Replace unions with chass hierarchies
21: Replace enum constructs with classes
22: Replace function pointers with classes and interfaces
6 Methods
23: Check parameters for validity
24: Make defensive copies when needed
25: Design method signatures carefully
26: Use overloading judiciously
27: Return zero-length arrays, not nulls
28: Write doc comments for all exposed API elements
7 General Programming
29: Minimize the scope of local variables
30: Know and use libraries
31: Avoid float and double if exact answers are required
32: Avoid strings where other types are more appropriate
33: Beware the performance of string concatenation
34: Refer to objects by their interfaces
35: Prefer interfaces to reflection
36: Use native methods judiciously
37: Optimize judiciously
38: Adhere to generally accepted naming conventions
8 Exceptions
39: Use exceptions only for exceptional conditions
40: Use checked exceptions for recoverable conditions and run-time exceptions for programming errors
41: Avoid unnecessary use of checked exceptions
42: Favor the use of standard exceptions
43: Throw exceptions appropriate to the abstraction
44: Document all exceptions thrown by each method
45: Include failure-capture information in detail messages
46: Strive for failure atomicity
47: Don't ignore exceptions
9 Threads
48: Synchronize access to shared mutable data
49: Avoid excessive synchronization
50: Never invoke wait outside a loop
51: Don't depend on the thread scheduler
52: Document thread safety
53: Avoid thread groups
10 Serialization
54: Implement Serializable judiciously
55: Consider using a custom serialized form
56: Write readObject methods defensively
57: Provide a readResolve method when necessary
References - Indexes of Patterns and Idioms - Index
The author (Score:5, Informative)
Java Rules (Score:3, Funny)
Man, I'm barely done with Douglas Dunn's [amazon.com]
Java Rules and now I gotta read this?
Actually, I'm glad to see more book son coding effectively as opposed to the dummy approach which is a sure way to Shoot yourself in the foot [healyourch...ebsite.com] when the maintenance phase rolls around.
That said, one thing copiously missing from the review is whether or not the book covers J2EE at all -- which by and far requires some guidance in the developing the most effective Java applications.
Re:Java Rules (Score:2)
Re:Java Rules (Score:1)
To be fair, that would be outside the scope of the book. It covers Java itself, the core language as an O'Reilly book might call it. I find it to be quite handy and well-written and agree with the reviewer for the most part.
I may check out Java Rules now. To easy to make fun of that name: Hey dude, Java Rules!
Try this one for J2EE (Score:1)
Chapters
The authors include Hans Bergsten, William Crawford, Dave Czarnecki, Andy Deitsch, Robert Eckstein, William Grosso, Jason Hunter, Brett McLaughlin, Sasha, Nikolic, J. Steven Perry, George Reese, and Jack Shirazi.
This not a bad book. For $34(USD), there are some good learnings. This is one to have on the shelf.
Another good choice along these lines... (Score:2, Informative)
Absolutely the best Java book I've read (Score:5, Informative)
Written by the Joshua Bloch, the acknowledged expert on the subject, it is as authoritative as they come and extremely well-written. After six and a half years as a senior developer architecting and implementing algorithms and class libraries in Java, this book shocked me by summarizing much of what I knew about how to use the language effectively, while teaching me much that I did not know. It continues to top my recommended reading list for all new software developers at my company.
-dk
Re:Absolutely the best Java book I've read (Score:5, Interesting)
This book will show you any bad habits that you have picked up, and teach you why they are bad. I've not found a single developer of any level of experience who has not found this book useful.
Re:Absolutely the best Java book I've read (Score:2)
Easily one of the top five Java books ever, along with Thinking in Java (2nd. ed., Eckel) and Concurrent Programming in Java (2nd. ed., Lea)
Re:Absolutely the best Java book I've read (Score:2)
Agreed; I found it useful as a Java newbie, because it gave me a view of the language "from the other side", a sense of where its complexities lay. It is a fine example of how to think about design issues in the context of a set of linguistic constraints, and for this reason has something of value even for non-Java coders (as the Scott Meyers series has much of value for non-C++ coders).
This book is invaluable. (Score:5, Interesting)
When I was first learning Java, I often had that nagging feeling that I was making things harder on myself than need be. This book cleared up a lot of those feelings, and helped get me on the right track for some of the great tools hidden away in the Java API. The API documentation is great for showing you what members and functions are available to an engineer, but this book shows you how to use the API. The review cites the author's avocation of knowing the libraries, I would contend that this book would help you get the most out of those libraries, and increase your understanding of them.
This book has become a must-have around my office, and if you are looking to get over-the-hump and move from an intermediate to advanced java software engineer, pick up this book. It is dense with information, and will save you a ton of time and energy. What more could you ask for from a book?
Hahah (Score:2)
Scott Myers (Score:5, Informative)
This is about the highest praise you can give a language-specific programming guide. In his books Scott not only listed many of the ways you could get in trouble in C++, but also gave clear explanations of why they were trouble and why his recommendations were good practices. His books are top of the list I recommend to people who know the language and want to write solid fast code. If Effective Java is as good, I'll soon be adding a copy to my bookshelf.
--Jim
Re:Scott Myers (Score:5, Informative)
The topics it discusses are much more obvious than the traps that Meyers covered. Java's a simpler language, that's fair enough but there are areas of Java, especially around concurrent programming and network and io apis that may surprise even experienced developers. Those areas are precisely where 'Effective Java' is thin on content and I found it disappointing overall.
Just because it's sectioned the same way as Scott Meyers' books doesn't mean it's just as good!
Re:Scott Myers (Score:2)
there are areas of Java, especially around concurrent programming and network and io apis that may surprise even experienced developers.
Do you have any suggestion for a book or online resource where I could learn more about these things?
I mean, yes, I know a little about those areas, and I know where to find books that teach them, but since you seem to have a lot of experience (I mean it, no sarcasm), do you have any tips for a source that cover these important areas you say Boch neglects, just as well as he does the rest?
Re:Scott Myers (Score:2, Informative)
Regarding concurrent programming, I would go for "Taming Java Threads" by Allen Holub.
A lot people would also cite "Concurrent Programming in Java: Design Principles and Pattern" by Doug Lea.
Personally I prefer the first. The second one is far more complete in its exhaustive analysis of concurrent programming, but the first is more "hands-on".
Confused by expression (Score:3, Funny)
What does 'setting out his stall' mean? I picture somebody settling on to the toilet and opening the newspaper, but in the context I don't think that's what you meant.
Re:Confused by expression (Score:1)
Good review (Score:4, Informative)
The code and the text are both very clear and concise, and you can read the book cover to cover, coming back later to study each advice more in depth.
I think you can get it from reading between the lines in the review, but just to make it clear to any curious beginner, this is not the first book you should buy if you are just starting to learn java. Use [sun.com]
The Java Tutorial instead, and maybe [oreilly.com]
Learning Java.
A few more good words (Score:5, Informative)
The reviewer complains about some of the thinner or more "obvious" items, but I disagree. Heaven only knows how many times I've wished that really good and experienced programmers follow what seems the obvious maxim: "Minimise the accessibility of classes and members."
It's true that "Know and use the libraries" seems rather obvious and vague advice, but Bloch's exposition drives home the fact that you may not follow this advice as well as you think. As always, his examples are excellent: he shows how an innocuous-seeming abuse of java.util.Random creates serious problems, and how proper use of the libraries fixes the problem. How often do you write a loop to print the contents of an array? I never realized until Bloch pointed it out that System.out.println(Arrays.asList(array)) accomplishes the same thing much more simply.
If you're a Java programmer, get this book. If you're a technical author, aspire to it.
Re:A few more good words (Score:2)
Only six years, so no, I suppose not, in the grand scheme of things.
The point is that sometimes very obvious things escape even the most experienced programmers.
A "Must Have" (Score:3, Informative)
Many important Java techniques and idioms are described so well in this book that I have been known to insert comments to the effect of:
(or some such) into my code. Most of the author's items should be as engrained in the mind of a Java developer as terms like 'singleton' and 'event listener'.
I have read a little bit... (Score:5, Informative)
Best of all, it's not 3 feet think like Effective Java Unleashed or The Effective Java Bible would be. You get lots of info with minimal fluff.
Read this book a lot (Score:3, Interesting)
I rarely find a tech book that offers more than a few examples I ear-mark. This book however has me re-reading it from time to time. The lessons picked up take time to sink in, and IMO no one can truly pick up all of the lessons on one read through.
This book is great (Score:4, Informative)
Re:This book is great (Score:1)
Re:This book is great (Score:2)
While I appreciate your comment, I want to humbly submit that, this being slashdot, you might have used a subject line more in line with the way the editors pick headlines. For example, instead of the mundane-but-descriptive subject "This book is great", you should have used something like:
"Dangerous book leads to shortages and overcrowding"
The only book I recommend. (Score:2)
I read books like other people eat hot dinners, and when I recommend Java books in work I only recommend this one. (I tend to find that a surprisingly number of coders only read one or two tech books a year anyway). "Effective java" is very well written; it's nice and short without sacrificing any exposition of the hairer parts of the language.
And as another poster has said, at work it's often sufficient to say in a review, "HashCode as per Bloch please"
Yep, it's a good one (Score:2)
It's nice to see a book review that actually reviews the book, rather than just regurgitating the table of contents.
Sample Chapters (Score:5, Informative)
My favourite quote from the review... (Score:1)
So sorry. My weird sense of humour strikes again.
You can catch some of the problems.... (Score:4, Informative)
http://pmd.sf.net/
PMD will find places where you've used concrete collections rather than interfaces, left unused code lying around, etc., etc.
Tom
Buy this book (Score:1)
Why Java can be faster than C++ (Score:1)
(F.e. you can eliminate a virtual method call just because you dynamically know what method you are actually calling; you can inline methods from system and third party classes, etc.). Just don't forget to add -server to your java run.
Worth the money (Score:2, Informative)
- Effective Java
and
- Java Platform Performance by Steve Wilson and Jeff Kesselman
Of the second book there is a HTML version available at this link [sun.com]
It won the 2002 Jolt Award (Score:2, Informative)
If you're interested, you can get a free [sdmagazine.com] subscription to the print edition of Software Development Magazine. It's one of the few high quality freebees (and no, I am not affiliated with them). Also, as Scott Myers got mentioned here a few times: He writes for SD Magazine once in a while.
Great for quoting (Score:2)
But the most useful thing about the book is that you can quote it to people. It has been a useful tool for applying to the Open Source mailing list I have to participate in.
And if you are writing an API in Java you need this book shipped to you overnight delivery. You can thank me later.
question for all you Java experts (Score:3, Insightful)
I dug up some old benchmarks (the BYTEmarks, which includes a jBYTEmark, normalized to that same P90) that I can use to fairly compare Java and C performance.
I've got a 1.2Ghz Duron, and the C version of the BYTEmarks seems to reflect this accurately--with scores of 26.6 Integer and 21.4 FP, with a P90 being 1.00.
I'm comparing this against two separate Java implementations--Sun's JVM for Linux, and gcj (I also tried out a program that translates Java bytecodes to C, but the resultant executable didn't perform correctly). Here's a table of results:
C Bytemark: INT: 26.6 FP: 21.4
SUN (--server) INT: 4.21 FP: 1.80
GCJ INT: 3.63 FP: 1.71
SUN (Default) INT: 2.18 FP: 2.14
Now, I'm not surprised that the C version ran so fast, but I am surprised the the Java benchmarks did so badly, considering that both benchmarks were normalized to an index of 1.0 for the same Pentium 90! Maybe I got a bad copy of the jBytemarks--it was hard to find a copy at all--but it looks pretty legit.
Do any of you Java people have some amazing tips for me? I did use all the optimization flags I could find. Or is this typical performance for Java, making a 1.2Ghz Duron run as slow as a P200 or a K6/300?
update (Score:2)
IBM INT: 5.8 FP: 6.4
These results still aren't as good as I'd hoped, but it does show that all of the JVMs aren't up to par yet, and therefore should be careful in choosing one!
Re:question for all you Java experts (Score:2)
At best you might be able to compare jvms with it, but since the benchmark is designed to measure cpu performance (isn't it?) it might not give meaningful results for that either.
--
Benjamin Coates
Re:question for all you Java experts (Score:1)
You're right about the JVM performance issue--I think it was originally normalized with Symantec's Visual Cafe? I can only assume they tried to use the best Java implementation they could at the time, but judging from straight CPU performance, I can only infer that JVMs have gotten worse, or that there are other factors at work. Maybe I should try those JVMs under Windows to see how drastic the performance difference is.
Re:question for all you Java experts (Score:1)
Re:question for all you Java experts (Score:2)
Sure. Wipe the words milliseconds, GHz, performance, benchmark, and optimization from you memory. They don't matter in about 99.99% of all programming, and caring about them where you shouldn't is likely to lower the quality of your program code. In particular they don't matter when comparing languages. Thank you for your consideration.
Re:question for all you Java experts (Score:1)
Obviously the times to care about benchmarking are 1) before you start coding -- if performance is important, find a good set of tools for your project and 2) when you're done with a cycle of development, to test the performance of your application, and see how it can be improved.
So far, I've found that C implementations haven't (or can't?) improve too much more (i.e., an order of magnitude) while Java implementations easily can (and, in fact, can easily show orders of magnitude performance differences).
But most of the time, your tips are applicable, much like the first rule of optimization.
Re:question for all you Java experts (Score:2)
I don't know how you did your tests, but Java benchmarking is a funny thing. Java should be run in a HotSpot or equivalent runtime optimizer, and when it is you get a slow startup time, slow runs for the first few runs of an app, and then quite fast performance thereafter after the code has been compiled, optimized, and cached by the runtime system.
On a server, that's how a real Java app operates. It is started up, then it is run again and again within an ongoing process. It doesn't have to restart and recompile, so it spends most of its time running quite fast after a slow start. Benchmarks that look at the first run instead of the 1000th are okay for C but completely misrepresent Java. Measurements should reflect the 99.9% case, not the exceptional case (startup).
In my own informal experiments, I've found that Java gets pretty close to C++ after it gets rolling. (I haven't benchmarked against C.) The differences are interesting to me, too, but are usually swamped by other issues such as database connections that dwarf the language delta.
Re:question for all you Java experts (Score:1)
I initially liked Java... (Score:1)
Useful not just for Java mavens.... (Score:2)
Great book! (Score:1)
Also, if you ever get a chance to see Joshua speak, go for it! He's a really engaging speaker, much better than the rest of us nerds
Re:Short Book (Score:4, Informative)
Re:Short Book (Score:1)
Re:Short Book (Score:1)
Re:Short Book (Score:1)
For reasonable GUI performance SWT [eclipse.org] is the way to go. The object model approximates AWT's so there's not a huge learning curve if you want to experiment.
Re:Short Book (Score:1)
Re:Used it for what? (Score:1)
"Clicky-clicky-drag" does not mean that you're not developing apps the proper way. A key idea (ideal?) in software engineering is code reuse and rapid development. If you have to write 34 methods to accomplish a text box for every app you write, then you're just wasting time. Sure C is great for mathematically intense stuff like DSP, and it's great fun for memory mis-management, but I wouldn't use it for a frontend. Use a good platform independent front-ending language, like java, to provide an interface to your powerful computational engine written in C or Lisp or Perl.
Re:Used it for what? (Score:2)
Mind if I borrow that?
I didn't realize that there were so many Java fan(atics) out there. I do realize it's importance to the computer industry, and especially the acedemic industry, but there is a lot of bloat to the language. Java has become a wishlist of functions. Great for learning and teaching OOP, but as it's been abstracted too far (in my opinion) it has lost a little bit of flexibility (and a lot of its speed). Personally, I don't care about overloading methods, total abstratcion, or write once debug everywhere capabilities. Garbage collection is nice though, as well as the additional security Java has to offer, but as far as it being the last word everywhere, just ask Corel.
Your suggestion to Use a good platform independent front-ending language, like java, to provide an interface to your powerful computational engine written in C or Lisp or Perl does make sense though, and you didn't have to come across as a zealot.
I'm burning Karma to reduce my heating bills...
Actually... (Score:1)
Re:Short Book (Score:1)
Re:Advice: Stick with a real OO language (Score:1, Insightful)
Java and C++ are popular because they're popular. Managers are often idiots, but most are smart enough to know that when you leave they'll have to find a replacement with your skills. There are lots of Java/C++ people out there, lots of books, lots of classes. Smalltalk, not so much...
And yes, I agree, BeOS and Betamax were better, but don't forget...there's a REAL WORLD out there!
Re:Advice: One new thing per project (Score:2)
Re:Advice: Stick with a real OO language (Score:5, Insightful)
Java is slow? Maybe that's just your coding. These guys (among others) don't agree with you...
http://www.sosnoski.com/Java/Compare.htmlhttp://www.javagrande.org/jgsc98/index.html
IBM has achieved 90% the speed of C in Java. I personally use the CERN numerical libraries ... very cool, very fast. Java certainly can be fast. I've seen distributed Java beat C++ implementations hands down (that was hard to everyone to believe ... problem was C++ CORBA marshalling and the RTTI overhead in C++).
I've used Squeak, and have a background in coding Smalltalk (Digitalk, ParcPlace, IBM VA). Squeak is a great up-and-coming environment, but FAR from commercial. Can you name one commercial implementation in Squeak?
I think that lack of commercial viability makes Squeak impractical. Java, on the other hand, has scads of drop-in commercial libraries and components. This makes economic sense in business. Java also has a proven track record on the server side, and there are many, many successful commercial implementations in mission critical environments.
And NO, I don't think Java is the perfect enviroment; far from it. I've yet to see my ideal language, and I keep looking on the horizon for a dynamic functional/object/aspect language that performs and holds up in a commercial setting. In the meantime, I've got real work to do.
Re:Advice: Stick with a real OO language (Score:1)
Commercial Implementation (Score:2)
Good point. However how many commericial applications are written in Java? I am not talking about custom client/server apps or installer programs but popular shrink wrap applications that are written in Java? Just curious as I never seem to encounter any and often wonder why not - given the Java advantage of cross-plateform capability.
Re:hrm? (Score:2, Informative)
Ok, maybe you should view what I posted, and re-read the links.
IBM makes claims to Java at 90% speed of C++ in their numerical benchmarks. JavaGrande includes a number of well-respected organizations doing intensive computation, including GRID work. The tests quoted show a wide variety of performance, and you picked on file I/O; yes, one of the worst aspects in those tests. However, note that these are OLD JVM tests, not the significantly improved I/O found in JDK 1.4. (I don't have stats handy but have seen and experienced the claims of improved performance).
Java as an interpreted, cross-platform language will always be at a disadvantage to C++ in terms of raw speed. The again, many equate C++ as "Object Assembly" in the OOP world. That has it's own disadvantages.
Everything depends on how the language gets used. I have written many distributed C++ infrastructures and applications, and find Java a much welcome alternative. In many distributed scenarios it has proven faster, to the surprise of many seasoned collegues. I couldn't even imagine the jirations required to do an equivalent of JINI in C++.
So, why is it that most financial institutions have dropped C++ in favor of Java for server applications? It must suck something fierce, eh?
Re:Mod that son of a bitch down. (Score:1)
Thinking in Java (Score:1)
You forgot to mention that most of Bruce Eckel's books are available online [64.78.49.204] in a variety of formats, for free.