Philosophical musings: interpreting models

Alaric Williams
Fri, 10 Sep 1999 11:14:43 +0100 (BST)

On Fri, 10 Sep 1999, Jim Little wrote:

> . Dial a phone number
> . Print a document
> . Display a picture
> . Calculate income tax
> etc
> I.e., real-world programming goals.

Reminds me of some musings I had about expressing the capabilities of a
computer to programmers. I mean, every machine can compute and remember
things, but some can also phone people, print things, display images, etc.

These extra facilities have to be:

1) Described/interfaced, even though there may be variations in the
quality of service a peripheral can provide. This is the job of low level
"drivers" that do nothing but convert the peripheral's internal interface
into a standardised interface for that operation as efficiently as

2) Enumerated by some device-independent portion of the system that knows
what the system can do in terms of a list of active drivers and their
capabilities, and then handles appropriate "multiplexing" and access
control to convert the list of things the physical hardware can do and the
list of what rights things have into a per-"thing" list of what tha
"thing" can do, including sharing resources in device-dependent ways
(printers are spooled, sound output devices are shared with some form of
mixing, consoles are shared with windowing environments, etc). This
includes such thorny topics as the physical devices "keyboard", "mouse",
"sound hardware" and "display" need to be merged into "console"; yet a
machine with dumb terminals hanging off of it already has these things
merged into "consoles". The definition of a "console" needs to be able to
expand with the invention of things like advanced sound output; UNIX
systems currently suffer from their "console" concept being limited to
beeping, with sound output being completely unrelated to a "console" per
se (you can't play MP3s back over a telnet connection).

> Jim

---==========================[ ]============---

	Almost three thousand years ago, I was on the comittee
	who designed Christianity... but it took us nine hundred
	years to get it past the risk assessment subcomittee :-(

---=========[ ]==================================---