The question is: how many Desktop (Homedesktop) OSes available?
Easy to answer: 3: Microsoft Windows, several Linux distros which seek to become a Desktop OS and BeOS which is a true Homedesktop os.
Neither SuSE nor Mandrake were able to achieve a home desktop. That is because of the packaging system: depencies over depencies… That makes everything not really easy. Before a proper packaging system has been created where the user doesnt need to install 20 libraries because he wants to use an application, Linux will have big problems arriving at home computers. Personally, I love Linux because I can use it, but others can’t.
It would be really nice if haiku would become a true choice for home computers without doing same mistakes which microsoft did: single user as an administrator and autoupdate not known to a newbie.
As soon as the autoupdate is by default on and can be turned off, the security is a simple problem since a newbie won’t turn it off. But such an autoupdate should be used only in case of remotely exploitable holes, other errors should be relatively optional.
Why force a user to run a limited username and use an administrator to solve installations etc? Well, this way simple virii would have a problem getting into the system and doing with it what they want.
To get a part of the market Haiku will have to create a standard distribution which can be used to create software. That means: if I create a media player, I should be sure that everyone will be able to install it without downloading additional stuff.
Are all of these 3 requirements fulfilled, Haiku can begin to get its users, users willing to work with a comfortable office suite (OpenOffice.org), wanting to listen to music easily, wanting to watch their favorite videos without having a lack of codecs (every codec developer including divx is interested in extending his influence onto the market and if he sees a good option, he will recompile his codec for this plattform, too)
As long as there is no proper serious alternative plattform with a standard distribution on the market, the industry will be not really generous at providing support for this plattform (drivers, hardware, software). Most of the hardware is sold to home users and small companies who are looking for the cheapest opportunity to get their computer working without going into a compromise of getting into a software lack. In the first stages of development there will be a lack of gamers on an alternative system: they can’t run their games on the system. Porting directx games to an alternative is not really easy, you have to either rewrite directx classes to create a c++ gateway to open libraries as OpenGL or OpenAL, or you have to create a directx clone with which the game is also compilable what makes even more work without bringing a better sollution. The same problem is with MFC, but on software level people tend to alternatives and don’t require a 1:1 copy - impossible with games
The current main aims of Haiku should be: providing a development IDE supporting GCC and Pascal (freepascal), MeTOS is a good application for such purposes. Without software no users, as many developers will be needed as possible. And of course, to create a standartisized, working (and stable) system which has the same advantages as BeOS had. I’m sure that the development team is very competent to create such a system … actually I have any doubts
The first release should contain also somethign for developers, the best would be a possibility to port gnu software easily. It isn’t a really nice idea to port GTK or Qt GUIs to BeOS/Haiku: it will eat a lot of RAM. Porting software to BeOS/Haiku should mean porting a console applications and writing a native GUI for it. Same has been done for example on MacOSX and on other systems and worked very good.
I just had to get rid of that, that were some advises, I just want you to read it through and discuss - I’m open for a discussion and I’m waiting for constructive opinions.