Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!helios!bcm!dimacs.rutgers.edu!seismo!uunet!shelby!agate!miro.Berkeley.EDU!seth
From: seth@miro.Berkeley.EDU (Seth Teller)
Newsgroups: comp.graphics
Subject: Re: help with NURBS on SGI h/w
Keywords: NURBS, surface evaluation
Message-ID: <1991Feb21.230008.28233@agate.berkeley.edu>
Date: 21 Feb 91 23:00:08 GMT
Sender: usenet@agate.berkeley.edu (USENET Administrator)
Reply-To: seth@miro.Berkeley.EDU (Seth Teller)
Organization: University of California at Berkeley
Lines: 66

Blake Freeburg wrote
>> I need some help using the SG lib on an RS/6000 and drawing a nurbs surface.

David F. Rogers wrote
> You don't need to use the SGI hardware to do this efficiently. Look in
> the December issue of Computer Aided Design for the paper Dynamic Rational
> B-spline Surfaces by Rogers and Adlum. It's all there for the taking.

Seth Teller wrote
> this seems like poor (or at least not necessarily good) advice.  hundreds
> of person-hours have gone into the sgi library so that people like blake
> can draw surfaces with simple procedure calls.  why refer them to something
> unknown (with all due respect)?
>
> i didn't write the sgi nurbs implementation, but i have several years'
> experience working with it and contributing to the development of its
> interface and functionality.  please don't sell it so short.

David F. Rogers wrote
> Because I have tried it and don't like it. It is too restrictive, the
> explanation in the manual is terrible, etc.

Seth Teller responds:

i'll say this (again) as politely as i can manage.  you're recommending
that an inexperienced person, _in isolation_, transcribe an unknown, 
unproven technical paper into _isolated_ software, rather than use 
working, released, supported software.

your justifications are: 

1) an opinion;
	> Because I have tried it and don't like it.

2) an essentially meaningless (and in my opinion unfounded) statement; 
	> It is too restrictive,

3) a (possibly legitimate) complaint _about the documentation_!
	> the explanation in the manual is terrible, etc.

now, lots of people read this group and form opinions from it.
in my opinion, your posting ill-serves them.  most people who 
wish to render surfaces do not write papers about surface 
rendering.  the gl nurbs implementation is fine for, say,
95%+ of its users.  technical people may perhaps desire
something different.

my point:

it is simply a _disservice_ to the original poster to send 
him off on several weeks' worth of hell wading through 
Yet Another Spline Paper, instead of sending him what he
originally _asked_ for: some code to get surfaces drawn.

take care.

seth

ps.  some technical points:

	the sgi implementation does view-dependent subdivision.

	sgi hardware is used for low-level evaluation, but
	obviously such evaluation works by another mechanism
	on non-sgi boxes.  therefore not even the gl "needs" 
	sgi hardware to draw nurbs.


Brought to you by Super Global Mega Corp .com