🚀 go-pugleaf

In article <381@tut.cis.ohio-state.edu>, elwell@tut.cis.ohio-state.edu (Clayton Elwell) writes:
> 
> You're a funny guy.  You would have to rewrite the memory management
> stuff completely (not a fun time, since Sun has a custom MMU that
> doesn't look much like a 68851), not to mention all of the device
> drivers.  In other words, it would be as easy as doing your own port
> of, say, System V.  You'd also loose (a) streams, (b) the toolbox, (c)
> more solid networking code (although Sun 3.4 looks better than their
> previous efforts), ...

	A sense of humour sometimes helps when dealing with the ``cards''
	you meet in our industry.  Anyways, I fail to see why doing a
	full port of SysV is the same amount of effort as recutting 
	the *machine dependent* part of the BSD MMU code, scsi driver, 
	NuBus management, etc.  The idea is that since Sun is already 
	on a 68020 platform, one doesn't have to spend time recompiling + 
	porting all the applications, which is where the bulk of Unix source
	code is (albeit, the most portable part).

	As for the issue of networking, you gotta be really strange
	if you think SysV networking is in anything resembling a
	useable state.  I've a guy in our EE department with ~100
	3B series machines that would gladly trade his machines 
	for Sun iron if he had the chance - let's just say that
	the sole network application supplied with the machines
	("sorta" ftp) just doesn't meet their needs.

> 
> It's a nice idea, but I'm much more interested in porting NeWS first.
> 

	It's been done for both the MAC/II under A/UX, and, I think
	for the MAC/II under the "old" OS. 

> -- 
> 							      Clayton M. Elwell
>        The Ohio State University Department of Computer and Information Science

					Rick Spanbauer
					SUNY/Stony Brook