I seem to recall reading that the power of Unix stems from the wise
choice of a few design principles rather than the endless accumulation
of special cases. However, I cannot find where this is stated. I tried
a Google web and a Google Scholar search using the terms "unix endless
accumulation special cases", and I also asked ChatGPT for a publication
associated with this phrase. I also searched for "special" in
D.M.Ritchie's "The Evolution of the Unix Time-sharing System" and "The
UNIX Time-sharing System A Retrospective". (Amazingly, both have
several parts that still highly relevant.)
Can anyone help? Am I misremembering something?
Howdy folks, I just wanted to pass on word that there is currently a stack of 1983 System V documents on eBay, here's a link to the User's Manual: https://www.ebay.com/itm/225659365754
The same seller has these documents available in their shop (as well as a bunch of other old computing documents and magazines):
- User's Manual
- Administrator's Manual
- Error Message Manual
- Operator's Guide
- Graphics Guide
- Transition Aids
- Release Description
So not the whole lot, but a nice spread nonetheless. I've already got all of these and they're already all scanned in some fashion or another, or I'd have scooped em up already.
Anywho, figured that might pique folks' curiosity. The pricing is very agreeable might I add, sometimes I see single volumes from this set going for over $100.
- Matt G.
I have a copy of Lions's "A Commentary on the UNIX Operating System" It is
an n-th generation photocopy that I've had for some number of decades. It
also has the source code, as I guess all of those copies did back in the
day. I could just recycle the paper -- it's yellowed with age -- but I'd
rather pay it forward and ship it to someone who promises to make copies
for others who want it. This is also on bitsavers, so you're only doing
this to preserve the sense of history in passing photocopies around.
I'll take bids on how many copies you'll provide. I'll post the winning bid
(without their address), so that the community can pressureXXXXXX ensure
they meet their commitment. I'll send it out by the end of the week.
I will give preference to US/Canada domestic addresses since that is easier
for me. How much preference is left to my discretion. :)
I hope this doesn't make me sound like a jerk; it's intended to be
lighthearted community service.
> What is the name of the mathematic symbols
Here are some readings, not exactly names
incl (used with partial orderings) is included in, or is less than
|> is not greater than
|< is not less than
<wig
>wig
wig is approximately, or asymptotically approaches
~wig is approximately
Doug
I had an experience similar to Tom London's:
To: alice!rob
Subject: you've spoiled me
I can't believe it. I'm sitting here at home in front of my
2621, and I can't work.
Damn it. I've got to get a blit at home.
When I left Bell Labs, I had an X11 workstation at work, but
only a simple terminal at home. Having used a Jerqblit5620 for
years at both work and home, I found it incredibly limiting.
After a year or so I came across a reseller who had a lot of
off-lease 5620s for sale cheap (like USD 150 or so each). I
asked around the university I now worked at, found a handful
of other people who wanted in, and then a local small company
who did System V system-administration consulting who wanted
some for themselves, and were willing to handle all the paperwork.
All that allowed us to negotiate the price down even more.
In the end I bought six, of which I think four are still working,
though I haven't turned any of them on for years.
None of the Unixes I used at the time came with 5620 support,
but the protocol for the basic window system built into the ROM
was well-documented and I managed to roll my own host support.
I also managed to cobble up my own binary-loading tools sufficient
to get sam working (I forget how I compiled the binary for the
terminal); that was rather more work, but it was worth it to
be able to have sam and multiple windows from home, even though
it was the ROM OS and therefore mpx rather than mux.
I remember porting my version of the host code to Ultrix,
SunOS 4, and IRIX.
My workplace at the time had a little bit of VAX/VMS around.
I didn't use that much but wanted to try porting my host code
to VMS as well. VMS had had a C compiler for some years and
some sort of pseudo-terminal for a shorter time, so it ought
to have been possible. I didn't get around to it before we
finally left VMS behind in the dustbin of history. I wish I'd
found time to do it, just to show that there really was nothing
Unix-specific about the idea or the implementation. It's just
a multiplexing protocol; it needs no kernel support except that
needed to run a command-line session not attached to a physical
terminal, and networking has long since made that available on
any competent OS.
Norman Wilson
Toronto ON
The standard routine for drawing menus on the jerq was
'menuhit'. Items in the menu were centered, and the menu was
scrollable when a certain threshold number of items was
reached, and in addition, when the mouse pointer was in the
top (bottom) item of the menu and it was possible to scroll
in the appropriate direction, the menu was scrolled up or
down 1 line. The structure associated with these menus is
'Menu'.
There was however another menu-drawing routine, 'mhit', the
menus drawn by this being hierarchical, the structure NMenu,
which no longer contained text strings but an array of NItems.
NMenus also had provision for 'help' text to be displayed, a
simple string displayed on the screen, when button 1 was
pressed while an entry in the menu was selected.
The earliest version in the Eight Edition jerq code, also has
one function in the NMenu structure which is called when the
mouse pointer invokes a hierarchical menu. By Ninth Edition
this has been expanded, with 3 functions, one as above, one
invoked when an item is selected ('hit') and one when a
hierarchical menu is exited.
In the jerq code directories, under 'lib/jj', is a small 'ms'
document, 'A Library of Goo for the 5620', which lists
routines available in the library, and their authors. Andrew
Hume is listed as the author of 'mhit'.
Are there examples of code using these three menu functions
('dfn', 'hfn', 'bfn')?
There seems to have been little interest in hierarchical menus
at the labs, their use was quite limited. I found a program
in the Tenth Edition archive, 'bubble' (which seems to be a
program for displaying the three-dimensional structure of
molecules) which uses them. 'samuel' made heavy use of them,
including use of the 'hit' function, and Tom Cargill used
basically the same code in 'pads' wherein the routine was
called 'scripthit'.
The plain 'menuhit' survived into Plan9, but as far as I
know, it is only used by 'sam'.
Available at https://www.skeeve.com/bell-labs-cstrs.tar.gz
Warren and Brantley and anyone else, feel free to retrieve.
I have two sets - both are in the tarball so there are undoubtedly
duplications. If someone else can curate them into single canonical
set that'd be helpful, I just don't have the time right now.
Enjoy,
Arnold
> once you've got M3 and M4, you've got a naming convention; I'd
> think it a safe bet that there was an M5 that was an internal
> experiment, and that M6 was simply the next in line
M6 came first, created by Andy Hall as a portability tool for Altran.
I always assumed the name echoed Ken Knowlton's L6 (BelL Labs low
level list language, with a superscript 6). I seem to recall that M6
was endowed with a very labored acronym.
Doug
