Re: a crazy idea -- the Book

[Meskauskas Audrius]

Typically, the best books are written by the authors of the methods or 
other things that are present in the book. For instance, the "Molecular 
Cloning: A Laboratory Manual" is present in near every serious 
biochemical laboratory, despite this book costs several hundreds of 
dollars. The book is popular because it consists of multiple chapters, 
written separately mainly by the authors of the described methods. It is 
more popular than many other books, describing the same methods and 
costing one order of magnitude less.

If this is true for java also, the really good books can be written 
either by Sun,  by active members of our project and probably by the 
authors of other Sun API implementations. For parts where the 
development is no longer active the sections could also be written by 
somebody else of (if there are enough material) maybe even skipped.

Between the proposed topics of the book it could be (the list is 
discussible):
1. Advanced use of the standard java API. For simplicity, we can just 
assume that the reader knows the java language and various simple things 
like how to display a JFrame or iterate over collection. I suggest to 
assume that the reader is somebody like Sun Certified Java Programmer 
(then the program of that exam can be used to check what the reader 
already knows).  We can focus mainly on packages and classes where the 
topics are difficult and  documentation is very incomplete or even 
missing: advanced networking, swing models, non trivial class loaders 
and so on.

There are already many - too many - books for "beginners", frequently 
describing details that can be learned from the web without any book at all.
2. Each free java virtual machine project can contribute a chapter about 
that virtual machine. These must be the informative, useful chapters, 
giving the potentially useful technical details - not just the lists of 
supported platforms.
3. The structure of the GNU Classpath project can be described - that 
autogen.sh does, that configure.sh does, how this library is connected 
to the java virtual machine and so on.

Other ideas can also be suggested.

Hence we can write such book if there are enough persons willing to 
contribute chapters about the Free Java parts that they have 
implemented. In the past, I wrote multiple scientific articles for 
various journals (in English), once needed to write, technically prepare 
and arrange printing (but not distribution) of the small book and twice 
was supervising the shared development of the multi-author book where 
each author was contributing a separate chapter. If required, I can take 
the co-ordination of this Classpath subproject.

In the first step, we need the chapter proposals (topic and possible 
date of submission). Using  the CVS history information it is easy to 
set the priorities if several potential contributors want to write 
chapter about the same (or they can be just told to co-operate).

In the second stage, the book is composed from the received chapters. 
Normally the authors then receive they sections as they would appear in 
the book for the additional editing and proofs.

Finally the printing and distribution must be arranged. People probably 
may be more motivated and write more material if at least part of the 
profit from that book would be returned back to the authors. If I am 
wrong with this way thinking here, the better. Surely there are other 
reasons to contribute.

Audrius.