Xref: utzoo comp.sys.mac:33336 comp.graphics:6130
Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!apple!well!nagle
From: nagle@well.UUCP (John Nagle)
Newsgroups: comp.sys.mac,comp.graphics
Subject: Re: Macintosh ROM Sources
Message-ID: <12153@well.UUCP>
Date: 12 Jun 89 17:12:51 GMT
References: <1210@tnoibbc.UUCP> <2801@wheaties.ai.mit.edu> <89160.093417DN5@PSUVM> <109301@sun.Eng.Sun.COM>
Reply-To: nagle@well.UUCP (John Nagle)
Lines: 43

In article <109301@sun.Eng.Sun.COM> mae@sun.UUCP (Mike Ekberg, Sun {GPD-LEGO}) writes:
>>I think that this may actually make cloning the Mac Roms and system harder.
>
>I think a well accepted technique of legally reverse engineering technology
>is to have two groups, working in isolation from each other. First group uses 
>whatever(legally, of course {:-), it takes to figure out what the X does, 
>laser slice the die, disassmble the ROM, etc. This first group then generates 
>a *complete* spec, warts and all.
>
>The second group then takes the spec. and generates the code, using only the
>spec.
>
     That's how Chips and Technologies, the primary maker of parts for PC
clones, does it.
>
>Note under this technique, a complete, legally obtained  specification is more 
>dangerous to the owner of the technology.
>
      That's quite true.  Given a complete specification of the interface
offered to a Mac application, it would be straightforward to develop 
compatibility packages allowing one to run Mac programs on other 68000-based
machines, such as Suns, etc.  There are still "look and feel" issues, but
it's not legally settled yet whether copyright protection really can be used
to protect interfaces.  It may well turn out that doing so is an antitrust
violation.  IBM tried to use patents and trade secrets to prevent others
from building IBM-compatible peripherals for IBM mainframes in the 1960s
and 1970s, and they lost in court consistently.  Today, many vendors build
equipment compatible with the machines of others, and it is settled law that
one can do so.  The same reasoning may turn out to apply to software 
interfaces.  The effort by Atari to make Nitendo cartridges may well
produce the litigation that settles this issue.

      As Mac applications get cleaned up, and more run under both UNIX and
the Mac OS, the application interface will inherently become better defined.
It should be possible today to develop an interface kit which allows the
execution of "well-behaved" applications on Suns, Amigas, and such.  A
working definition of "well-behaved" is that the application must run
on the full range of Mac hardware and operating systems.  If it does so,
it probably conforms to the interface spec as defined in "Inside Macintosh,
vols I-N".  And, of course, nothing in "Inside Macintosh" can be a trade
secret, since those are published books.  

					John Nagle