I've seen this used to advantage under AUX (several years ago): one
monitor could be dedicated to X11 and the other to MacOS, for instance.
Many graphic artists who use the Macintosh use more than one monitor. Two
monitors are cheaper than a single monitor with the same real estate.
Suppose that Linux supported USB keyboards and multiple displays. Then
two, three, maybe four people could work around one workstation, taking
advantage of Linux' stability and multitasking abilities, with a quite
small incremental cost per person. How many offices have a printer on
everyone's desk? Printer switches and networked printers are ubiquitous.
Multiheaded PCs will become common once businesses figure out that similar
savings are possible with them. Upcoming versions of Microsoft Windows
will likely support this. If Linux does not, it will slip farther from
acceptance as a desktop OS in the cubicle zone.
> Second: PCI is much faster than VLB.
I'd say it depends, but this has been debated elsewhere.
> Fourth: My suggestion is to use the PCI card and not do anything, you
> can't merge the VRAM to 2 megs....
>From the GGI FAQ ( http://synergy.caltech.edu/~ggi/GGIFAQ.html ):
+ Because SVGAlib and X are user-space programs, there are many
things that are not currently possible to do with the video
card. For example, it would be very useful to be able to use
the video card's unused video RAM as a fast buffer cache, but
this is not currently possible since the kernel has no
control over the video card under SVGAlib or X.
___
Trevor Johnson <trevor@jpj.net>