Publisher: McGraw-Hill
Date: Jan 1999
ISBN: 0071348131
Purchase: Amazon
|
|
|
|
Author: Justin Couch Publisher: McGraw-Hill Date: Jan 1999 ISBN: 0071348131 Purchase: Amazon |
News: 18/8/99 JSDT 2.0 EA1 is now available. I'm struggling to keep up with the changes to the code and the book at the moment due to some impending work deadlines. I'm hoping to get the 1.5 updates complete shortly, but I can't promise anything unfortunately. I may skip these and head directly to the 2.0 stuff.
4/7/99 - Updates for JSDT 1.5 and other chapters are online. Although not yet complete, the pages are getting there. These should be complete within the next week or two.
20/5/99 - JSDT 1.5 is now available for free! The old 1.4 version was a for-sale product, but Sun have changed their minds again and made it free for all use - even commerical. I'm now going to pass over all the JSDT code and update it to 1.5. Stay tuned.
Following a couple of requests, I've put the code online from the CD. At this stage, I've just put a dump of the entire CD that I submitted for the book (I still haven't received my own copies yet). Its pretty hefty at around 7MB, so only if you dare :) I'll get around to splitting it all up as soon as possible.
| Return to top |
It is aimed squarely at Java 2. You won't find any Java 1.1 and "and now here's how it's done in Java 2" sections. I cover all aspects of Java networking and related topics such as threading and serialisation from the aspect of usefulness to networking.
Far from being an API bible, this book aims to introduce you to the essential parts that you need to know in order to do your job properly. Part of this is knowing the strengths and weaknesses of every option available to you. Also, it seeks to avoid as much of your pain as possible by showing all of the common pitfalls that besiege a networking novice. That is, we take you through the development of the code just as if we were writing it for the first time. The code makes mistakes, shows you where they are and then how to correct them. Thus when (not if) you make those mistakes again at a later time you'll be at least prepared to know how to deal with them.
| Return to top |
Work came back for review and final edit just after RC2 was released in late
November (1998). Assuming that not much had changed we went with this -
particularly for the first few chapters. Luckily, the final release came out
just before I handed in the last few chapters (not including the JSDT stuff).
Having handed everything in, then came the official announcement and the name
change - DOH! By this stage, nothing had hit the presses and the editor asked
if I wanted to change the name to indicate Java 2. Quickly tossing up the
options I decided against it because the cover would say Java 2 but we couldn't
guarantee that we'd catch every mention in the body of the book. So the book,
AFAIK, was to be titled Real World Java 1.2 Networking. About a month later,
I receive an email asking about the relation between my Java 2 networking book
and the Real World book. After some investigation, it seems as though the
publishers changed their mind and changed the title and made the reference
changes in the book. The only minor hiccup in this is that the code still has
the rwjn prefix to all the packages. Oh well...
So to answer any early queries - Yes, the book is fully Java 2 compliant, it is not some beta java version based. If there are any slip ups, I'll let you know through this page. Check back every now and again for updates, typos etc.
| Return to top |
I just read a review over on Amazon. Its funny, but that is the first time I have ever been accused of lacking character! Normally its the other way around - I've got too much of it. Well, at least my books aren't all like the Java Bible that I've just helped to update. There I was weeding out odd tales about a guy drinking Coffee and other really strange things. I suppose if you want a beginners book that appeals to the idiot user then fine, my books are not for you. Almost 3 years after I wrote it, Late Night VRML is still regarded as a bible for VRML, even without having done an update. In private email, I have yet to receive one email that has any negative comments so I can't be doing too bad. Can't please everyone I s'pose....
Kestrel, which is Sun's code name for the next release of Java (aka JDK 1.3) has already started work. There are a number of really interesting things coming along to allow better networking. One of these is the possible inclusion of URNs into the core libs, and also the asynchronous I/O classes becoming available.
| Original | FTP | HTTP. |
| JSDT Chapters |
| Return to top |
After introducing the concepts of where different standards fit in, we move slowly up the chain to deal purely with IP based networking. At this level we discuss the roles of the various standards and standards bodies that exist and how they work in the big picture. Also, it introduces the basic standards that effect a large majority of your networking applications such as MIME types, Various TCP protocols and URLs
Having established all the basic theoretical concepts, the Java interpretation of these is introduced. All of the important classes are kept in the java.net package so we do a wide but shallow look at every class and their capabilities in the third chapter.
Multi-threading is absolutely essential to any non-trivial networking app. However, in this treatment, we look at it from the perspective of how it helps to write networking applications. That is, we don't go into heavy theory OS implementations and all that sort of stuff. After all, you're going to be using threading pretty much through the rest of the book.
Serialisation is also important to many aspects of networked, as well as non-networked apps. Serialisation is at the core of RMI and many other future technologies coming out in the Java world - such as JINI.
In this part of the book, we go it a little slow. A simple introduction to all of the requirements of network programming and surrounding code. Unlike the later parts, we don't dive heavily into all of the little options that we could use to improve the code. We stick very much to just the core socket programming information - creating connections, sending and receiving information (both binary and text), and tearing them down again. First establish good coding habits, then learn to make it dance.
RMI, as a basic system, is still relatively complex to set up well and make it worth the trouble. Simple things like dealing with Java's new security model up to complex issues like how to optomise RMI to make it fly are all covered. While doing this, you are given a pretty thorough under the hood look at how RMI is implemented. Why is this necessary? Well, in the real world, making RMI acceptable to end users (that is, making it quick), depends a lot on you knowing everything about how something is implemented. If you don't want to know how things work under the hood, then Java isn't the language for you because this knowledge is essential to making anything in Java run with decent performance. (perhaps try Visual Basic otherwise...)
Once RMI fits like a glove, we come along and disturb you again by throwing a whole heap more at you. Activation, new to RMI for Java 2, is a very worthwhile tool, but it also costs you a lot in other things, unless of course you know your way around its internals. Again, we build the background information needed to make your job easier when the PHB types tell you to change the design a week before something is due. Oops, need encrypted links, no problem we can do that for you too in RMI. It's all there.
One of the least understood, but one of the most useful part of Java has to be the URL classes. I have never seen a decent treatment of them in any Java text. They seem to be some black magic thing that programmers don't approach for fear of upsetting the mystical beast. We remove the mysticism and show you how really simple it is and what this can really do. We build customised protocol handlers for a Telnet application and an image loader for PNG images (all code included!).
JSDT is one of those great time saving APIs that really finds itself used in all sorts of weird places. Its great for building a quick chat application, but building a lock-em-out, no-holds-barred secure conferencing facility takes only a few minutes longer to code. If you're in any doubt, we build a full application in the finale to the book. There's almost 100 classes just for this one app.
| Return to top |
| Return to top |