[rescue] "bloated pig X11"

[Mouse]

> [...historical vs modern graphics...]  No need for graphics server
> because of IPC latency and memory latency of passing data and
> control.

You still need something a lot like it, though, to arbitrate screen
region access (in X terms, window size, location and stacking order,
and their implications for allocating screen real estate to clients)
and to demultiplex input events.

Also, the architecture you sketch completely ignores portability.  Of
course, this may be acceptable; I've seen more X code than I'd like
that blindly assumes the default visual is 24bpp TrueColor - and some
of it even blindly assumes it knows what the RGB masks are.

> So DPS would need to be interpreted in the application itself and
> would be a userspace API.

This will have trouble with the inter-client aspects of DPS, as I
sketched above.

> Also, current drawing models use vector and transparency (the "PDF"
> model), Postscript can use vector well but lacks transparency and has
> opaque colors only

That's not _that_ difficult to fix.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse at rodents-montreal.org
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B