[GRLUG] multiple login on Mac

[Bob Kline]

Re "way off topic," the Wednesday
meets are billed as forums for Linux and
other geek talk, so it might be
fair to say this venue is as well.  Sometimes
a thread will be billed as not strictly Linux, and
sometimes topics just drift, but a few will find
almost anything computer related of interest.
I do.  Some will fuss about all this.  But as the
venerable philosopher Bill Cosby once said,
"I'd don't know what a success is, but I do
know what a failure is.  That's someone who
tries to please everyone."  So do your thing
and take your chances.

On Sun, Jan 8, 2012 at 4:04 PM, Robar Philip <philip.robar at gmail.com> wrote:

> Getting way off topic, but being at Sun and working in the two major
> system software groups (Windows and OS/Net) was a big part of my life so I
> like to talk about it. :-)
>
> On Jan 8, 2012, at 10:53 AM, Bob Kline wrote:
>
> > Clearly "working in" trumps working with. i.e., Sun's Graphics Group. A
> fellow I
> > knew years ago became Sun employee number 36. Jerry Evans? Worked in
> > the same group, if there were groups yet in a company with 36 employees.
> The
> > same field anyway. He used to got to SIGGRAPH every year. I remember
> > visiting the plant in CA, and seeing a hush-hush project: a machine with
> a
> > SPARC processor.
>
> Jerry was well known throughout Sun (he’s still with Oracle as far as I
> can tell), but he was in the graphics/imaging part of Sun, not the Windows
> group itself if I recall correctly.
>
> > The first Sun workstation I had used a Motorola 68020. Seemed spunky at
> the
> > time, but wasn't by today's standards. The display was the big new
> feature
> > for people not so long before using ASCII terminals.... Later I got a
> SPARC I
> > workstation, and that was peppy indeed. A SPARC 5 for home use was less
> > people, but still compatible with SunOS, which was really the point.
> Plus it was
> > built like a brick privy.
>
> My first Sun was a Sun 3/50 (with a smoking 16 MHz 68020 CPU) I used at a
> startup I worked at for a year before I went to Sun. I still remember the
> day we upgraded the 3/50’s and 3/60’s from 1 to 4 megabytes. (Yes, MHz not
> GHz and MB, not GB.) It made an amazing difference in performance, but it
> cost a small fortune.
>

I had a Sun 3/60 about that time.
Some president had accused the
Japanese of dumping memory, so
the Japanese turned around and
either raised prices or cut supply.
In my case the company shelled out
$200 a MB for memory for my workstation,
in the dollars of 25 years ago.  I have
8GB in my home machine today, and
for $50 to $100 could double that if there
was any reason to...

>
> > Anyway, I recollect that NeWS was short lived, despite that fact that it
> > was intended to become Sun's standard windows system. Didn't happen as
> > I remember, and there were one or two others to choose from.
>
> Actually NeWS was around for quite a while. Its development started in the
> mid-80’s and it lived on in the merged X11/NeWS server until 2002. SGI
> actually shipped it as their main desktop, and it was ported to the Mac and
> OS/2. The reasons for NeWS’s failure were many and varied, both from a
> technical and business standpoint. The Wikipedia article does a pretty good
> job of summarizing them. That being said, NeWS was very cool and the AJAX
> style of development clearly traces its history directly back to it.
>
> > Jumping around a bit, my take is that while Sun dabbled with the idea
> for a
> > while, Solaris never became open source. And with Oracle now owning
> > what's left of Sun, probably never will.
>
> Solaris was open sourced, though Sun did a poor job of managing and
> working with the external community. It was called Open Solaris. The llumos
> (core OS) and OpenIndiana projects (illumos and desktop) forked from it
> when Oracle killed it.
>

I wasn't sure of the detailed reasons,
but Open Solaris never got much traction.
Linux came on fast and furious about that
time. Solaris was a wonderful OS, if less
so than the original BSD, but it never
developed much of an outside community
of developers and users, and too it was
never clear where Sun wanted to go with it,
or what it would ever put in to it.  There
seems to be no game at all now that
Oracle owns Solaris.

>
> > Just how FreeBSD came to be I only know vaguely, but it was the result of
> > Bill Jolitz and his wife - nominally a rewrite of BSD Unix, and somehow
> legally
> > open.
>
> Dr. Dobbs magazine documented the development of 386BSD in an 18 part
> series. (http://en.wikipedia.org/wiki/386BSD) BSD/386, FreeBSD, NetBSD
> and OpenBSD are all direct descendants.
>

Thanks for the URL.


>
> > … And I'll repeat that the big advances over the years are in silicon,
> not software….
>
> I’ll politely disagree here by example: threads and SMP, micro kernels and
> message passing, shared memory/RPC/Doors, NFS, Yellow Pages/NIS and LDAP,
> ZFS, virtualization, Sun’s Crossbow, jails/zones/containers, DTrace, ACLs
> and RBAC, Sun’s Service Management Facility, DBUS, Pulse Audio, HTML and on
> and on and on. System software developers continue to innovate at a very
> healthy rate. (As do application developers.)
>

Indeed.  Another set of discussion for
another day. Micro kernels go back a
ways - Mach?  Something using micro
kernels - herd? - was supposed to become
GNU's answer to Linux, but only Thomas
Busnell ever worked on it much, and
Richard Stallman got rid of him after some
tiff.  Busnell is now a monk, part time anyway.

Some of the things on the list might or
might not be strictly computer related -
e.g., HTML, which arguably made the
Internet a hit.  But TCP/IP dates to the
late 1970s or so.  As you point out,
Wikipedia is a never ending source of
history on computer history and technology,
and one can bone up on TCP/IP there, as
well as Linux, IPv4,6, etc.

My sense is that file systems are a
dime a dozen today, and every so often
one has some advantages over what
exists.  The same with languages and
compilers.  Any advanced CS student
worth their salt once ginned up a language
and created a compiler, few of which were
ever used for anything other than to get a
thesis.  Rob Pike, now at Google, still
mostly creates new languages, and a
few are bound to handy, while most likely
come to nothing.  yacc and lex are a
couple of age old Unix tools that are of
course used to create new languages
and compiles.  yacc = yet another
compiler compiler - I doubt many today,
used to using some of the higher level
languages you mention, have ever used
these.  How about using assembly
language?

So yes, lots of changes.  NFS, Yellow
Pages, and a bunch of other stuff came
out of Sun, and reflected its "the network
is the computer" motto.  There again,
hardware was developed to support that
philosophy.  Ethernet did not, but supports
local networks.  Message passing has
been a Unix kernel feature for a long time.
I think Minix leans on it heavily.

Anyway, yes, there's a lot of interesting
history.  And it is in part a chicken and
egg game.  Bigger drives, more memory,
faster CPUs and more capable GPUs,
faster ethernet, and on and on, all with
greater reliability, have been the enablers.
An amusing question is whether either
drives the game, or do both simply react
to new opportunities?  I think it's fair to
say that ever bigger, cheaper, more
reliable memory, storage, and, well, most
basic hardware are essentially are given
good things.

Standards, and the industries acceptance
of them, is also a major new thing that
somewhat stands on its own, but has turned
many aspects of computers into practical
realities.  Standards are probably
under appreciated.

    -- Bob


>
> Phil
>
>
> --
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.
>
> _______________________________________________
> grlug mailing list
> grlug at grlug.org
> http://shinobu.grlug.org/cgi-bin/mailman/listinfo/grlug
>

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://shinobu.grlug.org/pipermail/grlug/attachments/20120108/ad447e48/attachment-0001.html>