Search in Book Reviews

This book covers plenty of material related to distributed computing. Its main chapters include the following:

• Networking in Java (basically java.net)
• Distributing Objects (why do it?, RMI, CORBA)
• Threads (multithreading applications)
• Security
• Message Passing Systems (fixed + adaptable protocols)
• Databases (JDBC)

Overall I was disappointed with the book. There were many reasons for the disappointment. I would have liked to have seen more treatment of RMI. I sometimes felt that a flow diagram could have been used to elucidate rather lengthy explanations (there are few diagrams in the book). The explanations some-times seem to ramble on and become a bit wishy-washy.

The Threads chapter seems slightly inappropriate as the book assumes Java programming experience. I think it would be better referring to some other fuller treatment of multithreading.

The Message Passing chapter seems unnecessarily long. It also seems slightly dubious in advocating designing ones own protocol. The world seems full enough of protocols!

The Database chapter covers JDBC. The chapter explains how the programmer could simulate remote JDBC access using RMI if the JDBC driver does not support remote access. The JavaBeans specification seems to require that a JDBC driver should naturally provide access to remote databases. The chapter also mentions the programmer implementing his own caching mechanism. I find that slightly bizarre as the database and/or the JDBC driver should provide caching!

The examples are fair. However, there seem to be a lot of examples that fill the book up.

Niggles that I have with the coding style adopted in the book include:

there are no javadoc comments - I can't even remember an explanation explaining that all javadoc comments had been ripped out of the printed example for space reasons (unfortunately they are also missing from the online examples!) exceptions are sometimes caught and junked by use of:

• catch (Exception e) {}

  
  
  I don't like this because I think exceptions should generally be reported in some way or other. Either they should be thrown or at least have a System.err.println for debug purposes.
• the following line of code is used frequently:
  
  

  return success;

  
  
  so the operation is successful? No, because 'success' is a Boolean that may be true or false! At the least I'd rather see 'succeeded'.

• System.out

  is sometimes used for error reporting rather than

  System.err

  .
• some methods are marked

  throws IllegalArgumentException

  . As this exception is subclassed from

  RuntimeException

  these throws are both unnecessary and not idiomatic
• some

  StreamTokenizer

  processing is not robust, as it doesn't handle incorrectly formatted messages
• the use of

  System.out.println

  calls to explain what is going on at run time rather than suitable '//' comments

It is not really a bad book. It just doesn't identify its readership very well. It would probably be reasonable reading for a beginner or a low intermediate programmer.

I can't recommend the book in its current form.