'(blog:HappyPeng)

In these posts I will try to explain the various reasons which lead me to develop a new component system like midtalk, from my personal experience, and also the problems the GNOME platform faces today.

About seven years ago, I started to learn programming. Like some other kids of my age, I'm not perfect so I used the brand new Windows 95 and I started to search the web to find something free (as in beer) I could develop with. I started with QBasic which was included on the operating system's install disk, but I wasn't satisfied because I wanted something which could make applications like those which were on my desktop.

The first thing I found when searching on the web was a free (as in beer) version of the famous Visual Basic from Microsoft - Visual Basic 5.0 Control Creation Edition. And it very was fun, I liked it. I could make pieces of software that ran on web pages Internet Explorer, little bits of widgets which could appear on Excel sheets.

Later, I found Basic too limitating for my ambitious needs. So, I managed to get the whole Visual Studio, and to learn Visual C++. It was great too,
because I like to learn new languages, and I was even happier than before :
I could myself write stand alone applications which could integrates the parts I had previously written in VB ! And when I had written such programs, I could launch them from VB and pilot them directly with scripts. It seemed very powerful, and very easy with all those wizards. I was still very young at that time, and impressed.

Time passed. I got tired from Windows, I discovered GNU/Linux and free (not as in beer) software. I decided I could no longer develop on Windows, because it was unfair, and after all not so easy - on GNU/Linux, you already have the tools, you have plenty of libraries, you can watch every piece of code, even at the first sight for me it wass a lot cooler than developing on Windows.

Great. I liked it, and I still do. I wouldn't be there if I didn't. I first used KDE, and I started learning C++ again. I wrote small applications, nothing very interesting. Things you do to learn. And then one day I read articles from a clever guy named Miguel de Icaza, about some wonderful thing called Bonobo. That was it. The things I did when I used Windows : Bonobo allowed to do them in a free environment.

I found that very cool, so I dropped KDE and started to use GNOME. I learned to write good C code, GTK+ interfaces, and so on. Bonobo seemed to work, and it was used in applications like Nautilus to load various components, you were able to pilot some applications from scripts, etc. I didn't understand very well how to write Bonobo code and people started criticizing it a lot, but I didn't care : as long as it was just like the things I loved before, I defended it and I was looking forward to learn more things to be able to do those things later.

In order to get experience I decided to start contributing to a project. We were writing lots of clean GObject code, the design seemed very good and clear to me, I liked it a lot. But a time happend when we wanted more features, more power, more coders. We wanted to have it extensible.

Fine, dlopen () exists. But... What the hell I'm I doing writing thousands of repetitive lines of code. When I used Windows, I could do it simply with only a few lines ! Maybe Bonobo can help me. Finally, the whole thing, Bonobo, CORBA in C, GObject finished seeming just like heavily time-eating over simple solutions we could imagine or remember. We still didn't have the simplicity and power I knew.

I looked around, found nothing. The next article will explain in details what the situation seems to me in GNOME today.

PS: that's true, Windows 95 evokes special feelings for me, that are not quite like other people's :-) But things in this article are facts, not nostalgy. I don't regret Windows and I don't like the OS nor the fact it's not free. I'm trying to improve free softwares.