Committing in .

[wxWidgets.git] / docs / latex / book / chap_intro.tex

\chapter{Introducing wxWindows}\label{chapintro}
\pagenumbering{arabic}%
\setheader{{\it CHAPTER \thechapter: INTRODUCING wxWINDOWS}}{}{}{}{}{{\it CHAPTER \thechapter: INTRODUCING wxWINDOWS}}%
\setfooter{\thepage}{}{}{}{}{\thepage}%

\section{The need for portability}

Not so long ago, it seemed as though Microsoft was on course to
take over the desktop for good with the Windows range of
operating systems. Unix was retreating to server territory and
had all but given up the fight against NT. Apple was sickening
and it seemed almost common sense that the quirky Mac had had
its day.

At the start of the new century, the picture has changed quite
dramatically. Linux is advancing from its position of strength
in the server market and is poised to make inroads on the
desktop, where the high price of Windows gives it an advantage.
Apple has made an impressive comeback and is revamping its own
operating system. The embedded market continues to explode and
has given rise to new niches for specialist operating systems.
The keyword bow is diversity, when the best guess of a few years
was that Microsoft would conquer all. More than ever,
application developers need to keep their options open and
targeting one platform only can be a recipe for commercial (or
open source) disaster.

In the light of this change in the computing landscape,
cross-platform development (or multiplatform development as we
call it in this book) has become quite a hot and sometimes
bitterly controversial topic. There are a number of ways to
achieve the goal of portability, and many tools and languages to choose from.
However, it is still generally agreed wisdom that C++ is the
language of choice for desktop applications that depend on
speed, ease of delivery and compatibility with billions of lines
of existing code. Plus, there are many hundreds of thousands of
developers now experienced in C++ programming.

The purpose of this book is to give you (the developer or
interested manager) a solid grounding in one increasingly
popular multiplatform solution: wxWindows. wxWindows is an open
source GUI toolkit that has been on the scene since 1993 and has
reached a high degree of stability and functionality. It is
available for Linux (and most other Unix variants), all desktop
versions of Microsoft Windows, and Mac. Other ports are in
progress.

\section{Why wxWindows?}

While wxWindows is labelled a GUI development toolkit, it is in
fact much more than that and has many features that are useful
for many aspects of application development. This has to
be the case, since the whole of a wxWindows application needs to
be portable to different platforms, and not just the GUI
component. So wxWindows provides classes for working with files
and streams, multithreading, management of application settings,
interprocess communication, and much else.

wxWindows is open source, with a vibrant developer and user
community, and may be used in any proprietary or non-proprietary
project. wxWindows encompasses the whole spectrum of users from
one-man software outfits to big-name companies such as Xerox,
Lockheed-Martin, Pratt and Whitney, and Motorola; from computer
science departments to medical research groups; from ecological
research, to the telecommunications industry, and in a myriad of
open source projects.

When you use wxWindows you tap into an astonishing talent pool,
with contributors from a wide range of backgrounds. Many aspects
of application development that you might otherwise have to
laboriously code yourself have been encapsulated by these
developers in easy-to-use classes that you can plug into your
code. Since there are a lot of people subscribed to the mailing
lists, you'll enjoy discussions not only about wxWindows but
often other matters close to the hearts of both experienced and
inexperienced developers. And hopefully, one day you'll join in the
success of wxWindows and become a contributor yourself!

\section{The history of wxWindows}

wxWindows was started in 1992 at the Artificial Intelligence
Applications Institute, University of Edinburgh, by Julian
Smart. Julian was designing a kind of meta-CASE tool called
Hardy which needed to run on Windows as well as X-based Unix
workstations. The existing commercial cross-platform tools were
deemed too expensive for an in-house experimental project, so
the only alternative was to build one. wxWindows (w for Microsoft
Windows, x for the X Windowing System) started off with support for XView and MFC 1.0, and
AIAI allowed it to be released to the Internet. As it became
clear that XView was doomed, a Motif port was written. Borland
C++ users began to request a version that was not dependent on
MFC, so the Windows port was rewritten to use the native Windows
API. Over time, a small but enthusiastic community of wxWindows
users was established and a mailing list set up. Contributions
and fixes were sent in, the largest contributions being the Mac
and Xt ports. wxWindows gradually picked up more and more users
from all over the world: commercial, government, and - most
gratifying of all - company users who found that wxWindows
offered a better product, and better support, than the
commercial products they had looked at or used.

During 1995, Markus Holzem released his port of wxWindows to Xt,
the X toolkit. This meant that software could be written that
would work on X systems without the need for installing Motif,
which is a commercial product.

In 1996, Julian left AIAI to pursue freelance consultancy, and
there were few new releases. The seemingly unstoppable advance
of Java was throwing some doubt on the future of C++ libraries.
After many contributions, the wxWindows code was looking rather
tired and in need of a serious rethink. At the start of 1997 it
was make or break time - call it a day and move onto other
things, or commit to re-engineering wxWindows with an API that
was flexible enough for current and future user interface
trends, with more 'widgets', and better use of C++. After an
exchange between Julian Smart and Markus Holzem (the author of
the Xt port), it was decided that it was worthwhile. Julian
would do the Windows port, Markus the combined Motif/Xt port,
with both contributing to the generic parts. wxWindows 2 API
ideas and code started to take shape.

During 1997 an effort to produce a standard Linux desktop
environment was underway - GNOME. Its widget set was GTK+, built
on top of X11 and it looked as though GTK+-based apps were to
become the standard in the Linux universe. Its one major problem
- GTK+ was C-based, and only a thin (and unportable) C++ wrapper
existed for it. In August 1997, Wolfram Gloger made a suggestion
that wxWindows 2 should be ported to GTK+ - Robert Roebling had
initial reservations, but in general supported the idea. He
became the prime mover for wxGTK and alphas were made available
at the beginning of 1998. In May 1998 the Windows and GTK+ ports
were merged and put into a CVS repository made available to all
contributors to wxWindows.

Markus Holzem had to drop out of the wxWindows 2 effort in early
1998, but Julian Smart started a new wxMotif port. The idea of
an Xt port was dropped, since there remain two possibilities for
free Unix programming with wxWindows: wxGTK, and wxMotif with
the Motif clone Lesstif.

In September 1998, Stefan Csomor started a new version of the
wxMac 2 port based in part on Greg Whitehead's initial work.
Several releases have been made and the code will be integrated
into the main CVS stream in Summer 2000.

David Webster and Stefan Neis are working on an OS/2 port, and
interest has been expressed in ports to other systems, such as
BeOS.

wxWindows 2.1.11 saw the bundling of Vaclav Slavik's impressive
wxHTML classes, with subsequent enhancements including
print/preview ability, and a helpview application and help
controller class for cross-platform HTML help. wxHTML gives
applications easy access to enviable formatted text viewing and
reporting capabilities.

During 2000 wxWindows will adopt Vaclav's XML resource system
as a replacement for the WXR system; this will allow much
greater flexibility and portability with support for
sizers and many more controls than WXR.

\section{How to use this book}

The chapters in this book are meant to be fairly independent, but
as usual, the ordering reflects a progression from basic concepts
to more advanced ones, so the first few chapters at least
should be read consecutively for best effect.