Path: utzoo!attcan!uunet!lll-winken!lll-lcc!ames!oliveb!pyramid!prls!philabs!sbcs!root
From: root@sbcs.sunysb.edu (root)
Newsgroups: comp.sys.amiga
Subject: Re: Studebakers
Summary: Hardware really isn't that capital intensive
Message-ID: <1326@sbcs.sunysb.edu>
Date: 10 Jun 88 21:33:03 GMT
References: <4400@gryphon.CTS.COM> <56089@sun.uucp>
Organization: State University of New York at Stony Brook
Lines: 73

In article <56089@sun.uucp>, cmcmanis%pepper@Sun.COM (Chuck McManis) writes:
> In article <4400@gryphon.CTS.COM> (Richard Sexton) writes:
> > As a related question, given you had ONE improvement you would make
> > to the amiga, what would it be ? 
> 
> Well I contend that it is a *lot* more capital intensive to develop 
> hardware than it is to develop software. So I the improvement I would

	The first time around hardware is indeed very expensive
	to develop.  To do a reasonable job of things, one needs
	a prom/PAL programmer (~$0.5K-$10K), 'scope (~$2K-$6K), Logic
	analyzer ($10K-$20K), decent software system to do driver development
	on (~$4K -> $20K), misc lab stuff (~$10K or so).  After that
	it costs just around $5K per board layout + parts.  So it is not
	a lot.  *Any* software development will burn up far more
	people dollars than hardware capital equipment investment.
   	Looking at the whole picture, the only difference between
	hardware or software development costs is just how the money is 
	paid out; for hardware you drop ~$50K at first, for software
	you pay it out in $50K/year/programmer chunks.

> Some examples of things that would be much more common today if we had
> this thing :
> 
> 	o Networks - Ethernet, StarLan, Arcnet, LocalTalk, and MAP.
> 	    Sure Ameristar has done ethernet and done it well, but they 
> 	    could have done it faster and cheaper if they had an off the

	Cheaper perhaps if we could leverage off of the volume present
	in the 'PC market & use a 3Com or WD PC bus Ethernet card.
	Of course, it is arguable who has a better Ethernet board 
	design, eg 3Com/WD or ours :-).  Faster is a function of the 	
	number of programmers you can afford.  Afford is driven by whether 
	you're starting with venture capital or whether you can bootstrap 	
	yourself off sales of simpler devices that require a lot less software,
	eg memory boards, disk controllers, etc.  Looking at the history
	of Sun, you folks had mongo venture capital $$ in the early
	days plus the original OS was purchased canned from Unisoft
	I believe.  I suspect that if dollars were a bit smaller or
	Sun had to hire enough programmers to field a Unix product
	themselves the growth curve would have been a lot less spectacular
	that it was.

	By the way, we've done both Arcnet & Ethernet.  Localtalk
	needs to be built into the machine, ala Mac to be cost
	effective.  That said, who needs Starlan :-)

	And, while we're comparing costs of PC Ethernet vs Amiga Ethernet,
	remember the salient point is *system* cost.  Sun PC-NFS + a 3Com card
	from Sun lists at $999.  Our package (functional equivalent of
	Suns) costs $899.  So even though PC-NFS ships probably 4-5X
	our volume, we cost $100 less.

> 	    shelf ethernet card they could OEM and just concentrate on 
> 	    the software.
> 
> 	o Frame Buffers - Targa or Truevision cards could be running *today*
> 	    on the Amiga, generating *fantastic* ray traced images after doing
> 	    the 'check' on the Amiga side.

	The tough part in building Targalikes for the Amiga is that
	the framebuffer standard on the machine pretty much dictates
	custom VLSI/gate arrays to make it really compatible, or you 
	declare yourself wholly incompatible.  This contrasts with 
	the IBM/PC for example, where it is fairly easy to build in 
	at least CGA compatibility.  

> --Chuck McManis
> uucp: {anywhere}!sun!cmcmanis   BIX: cmcmanis  ARPAnet: cmcmanis@sun.com
> These opinions are my own and no one elses, but you knew that didn't you.

					Rick Spanbauer
					Ameristar