I'm used to working with several platforms. The most interesting part of it is the fact that I use the same platform with different operating systems, such as Win95, Warp and Linux on the same machine. Because of this, I can see some differences in the performance of some devices from one system to another. Commonly, this difference is caused not by the operating system itself, but by the drivers written for them. As an example, look at the ATI Mach 64 drivers. The ATI Mach 64 family of video boards is certainly very fast and full of capabilities, in all versions (from the old XPression to the newest 3D Series). However, not all the options are available in all the drivers, for a variety of o.s. Under Warp, few of the functions are used, and the driver-user interaction is limited to changing the resolution. Even for that, a boot is needed. Now look at the same board, under Win95. The driver, also made by ATI, comes with a full set of capabilities. Besides changing resolution during usual operation, it is possible to set the frequencies of operation, width/height of the screen, its position relative to the monitor, even the intensity ratio between the three basic colors.
So what is the purpose of this discussion? I think that on the long run, the success of an o.s. is determined by the support it provides for the several devices available on the market. It is essential that every major product be supported. Many drivers for OS/2 are only crippled versions, if compared with the Win95 ones. Consequently, people are more comfortable with Win95 since they can more readily customize their machines.
The hardware firms think it is not worth the time to produce a good OS/2 driver for their product. However, since under Win95 the drivers are very sophisticated, a product will only win in this platform if the support provided by the driver matches the one provided for other concurrent products. From this context, I think that a great effort is needed to make the hardware vendors support OS/2. And it can be as simple as asking the manufacturer for a good OS/2 driver for every product you buy. Since they usually seek numbers, if we can make a substantial number of requests, they will think it is worth the work. It worked with the Netscape (for developing a native Warp version), so why shouldn't it work with others?
An operating system can only survive if it has somewhere to be run. That's the key, and it will decide the future of OS/2.
Please do not hesitate to send me any Brazilian specific news and information you would like mentioned in this column.
Rafael Donnici de Azevedo is graduating in Systems and Electronics Engineering at PUC-RJ and does research in the field of Parallel Proccessing. Rafael has worked with OS/2 since 1993, and has made many presentations of OS/2 versions 2.1 and 2.11 in a project between PUC and IBM. Currently, he is the Web Page Editor for the Rio de Janeiro OS/2 Users' Group.
