Re: Heated agreement? (was) Re: Canvas shortcomings

[Mark <jamess1 wwnet net>]

On Thu, 21 Jun 2001, Nathan Hurst wrote:

> On Wed, 20 Jun 2001, Mark wrote:
> 
> > Another possiblity is no interpreter at all. The interpreter allows a
> > clean seperation between the display server, and client application. Mac
> > OS X seems oriented to the desktop user, so they probably don't care about
> > this. They might use a hybrid interpreter/API, where the API has fine
> > control over the interpreters actions. This would probably be closer to
> > most graphics APIs available.
> 
> Here are some more questions to think about(nobody has really answered my
> previous questions, besides Mark, so I can only assume that the other
> participants haven't actually thought them through).
> 
> Is there a difference between an interpreter and a serialised protocol?

I guess you mean a difference between Postscript and X/PDF. Yes, but very
slight. X and PDF have a fixed set of operators and are non programable,
where Poscript is a general language which is exceuted on the server.

> Why do we want an interpreter?
>  (What will be the benefits and costs associated?)

The benifits would be generalality, and reduced network traffic. It's hard
for me to predict the cost of interpetation.

> Why do we want a serialied protocol?
>  (What will be the benefits and costs associated?)
> 

Possible speed increase at the cost increased network traffic and loss of
generalality.

I'm not sure how heavely each of these weighs though. That this the speed
loss from interpretation, the average amount of increased network traffic,
and wheather the full power of postcript is wanted or a subset would be
sufficient.

Mark