Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!cs.utexas.edu!sun-barr!rutgers!psuvax1!news
From: melling@cs.psu.edu (Michael D Mellinger)
Newsgroups: comp.sys.amiga.advocacy
Subject: Re: NeXT Press Release
Message-ID: <1o4G7xmu1@cs.psu.edu>
Date: 21 Apr 91 01:02:19 GMT
References: <.-2G2vju1@cs.psu.edu> <11226@uwm.edu>
Sender: news@cs.psu.edu (Usenet)
Distribution: comp
Organization: Penn State Computer Science
Lines: 51
In-Reply-To: gblock@csd4.csd.uwm.edu's message of 21 Apr 91 00:03:36 GMT
Nntp-Posting-Host: sunws5.sys.cs.psu.edu


In article <11226@uwm.edu> gblock@csd4.csd.uwm.edu (Gregory R Block) writes:

   > 
   > Why exactly is Display Postscript slow?  I've heard quite a few Amiga
   > droid say that it is, but do any of them have an answer to my
   > question, or did another Amiga droid tell you that it is slow?
   > 

   1)  Walk through the code, and you'll understand.

But I don't understand, explain it to me.

   2)  How large is it?  Something that large takes up processor time,
   like it or not.  It simply does.  PostScript itself is rather clunky
   and Display Postscript is better--but not by much.

It must be better.  I've seen a LaserWriter spend a couple of minutes
imaging a page to print.  Display Postscript on the NeXT doesn't have
this problem.

   3)  If it truly does take up 10% of a 68040, as you said, it _IS_.
   May not be that much to someone who has never had to squeeze every
   inch of a productive session out of a computer.  But for those who
   have used Suns, Apollos, or ANYTHING except a NeXT know that a slow
   GUI is BAD for productivity.  And the solution, as you would like to
   think, is NOT to throw in a bigger cpu.

I didn't claim that it took 10% of the CPU time.  I imagine that the
actual amount of time varies.  If you are watching your program
compile, I seriously doubt that 10% of the CPU time is being used by
Display Postscript.  

   I believe someone said that you should code for the smallest, fastest,
   most optimized programs, no matter what you plan on running it on.

What a nice saying.  Should I write in 68000 assembler?  Wait!  NeXT
is going to move to a RISC chip in a year or so.  Oh well, then I'm
SOL.

   The Unix kernal in general is big and obtrusive.  NeXT's kernal is
   larger than most, if not all, I believe.  That fact in itself makes me
   a tad wary.

Buy more RAM, it's cheap(< $200 for 4MB).  Unix is like having a big
powerful engine in your car, except that gas is getting cheaper
everyday.  Don't worry, it's environmentally safe to use a RAM
guzzling NeXT.


-Mike