Path: sparky!uunet!sunquest!venus.sunquest.com!terry
From: te...@venus.sunquest.com (Terry R. Friedrichsen)
Newsgroups: comp.unix.bsd
Subject: Sigh - now it'll *never* be free
Summary: CSRG disbanding as of 4.4BSD
Keywords: 4.4BSD   CSRG
Message-ID: <41950@sunquest.UUCP>
Date: 12 Jun 92 18:33:51 GMT
Sender: n...@sunquest.UUCP
Followup-To: comp.unix.bsd
Distribution: usa
Organization: Sunquest Information Systems, Tucson
Lines: 27

I just read a report that CSRG is *disbanding* with the 4.4BSD release,
which the report says will be released fairly soon.  According to this
report, the release will still be burdened with USL kernel sources.

Can anyone comment on the accuracy of this report?

If true, this greatly saddens me.  It looks like I have just lost any
chance of running BSD free with sources on my DECstation.  I was really
looking forward to running an operating system with full source code
again ...

Does anyone see any chance that either of the 386-style BSDs discussed
here will ever be upgraded to run on real computers?  ;-)

(Please, PLEASE note that the above remark was made in jest.  I greatly
applaud the efforts of both groups to bring full Unix OS sources to the
masses.  I do not mean to denigrate their massive efforts AT ALL (I only
mean to denigrate the 386 :-).)

Terry R. Friedrichsen

te...@venus.sunquest.com  (Internet)
uunet!sunquest!terry      (Usenet)
te...@sds.sdsc.edu        (alternate address; I live in Tucson)

Path: sparky!uunet!spool.mu.edu!agate!boulder!ophelia.cs.colorado.edu!drew
From: d...@ophelia.cs.colorado.edu (Drew Eckhardt)
Newsgroups: comp.unix.bsd
Subject: Re: Sigh - now it'll *never* be free
Keywords: 4.4BSD   CSRG
Message-ID: <1992Jun15.202726.21177@colorado.edu>
Date: 15 Jun 92 20:27:26 GMT
Article-I.D.: colorado.1992Jun15.202726.21177
References: <41950@sunquest.UUCP>
Sender: n...@colorado.edu (The Daily Planet)
Distribution: usa
Organization: University of Colorado at Boulder
Lines: 44
Nntp-Posting-Host: ophelia.cs.colorado.edu

In article <41...@sunquest.UUCP> te...@venus.sunquest.com (Terry R. Friedrichsen) 
writes:
>I just read a report that CSRG is *disbanding* with the 4.4BSD release,

This is correct.  Funding problems, the loss of both Karels and Bostic
(Keith just got married, and is moving out of state.  Berkeley
will not allow him to telecommute), and the political situation all 
have taken their toll.  

>which the report says will be released fairly soon.  According to this

Kirk said ~1 month for an alpha, at the BSD BOF at Summer USENIX.

>report, the release will still be burdened with USL kernel sources.

Technically, this is correct too.  There will be a 4.4bsd, requiring 
at least a 32V license, as well as a 4.4bsd lite,
or a net3 if you will - in the same state as net2 as 
far as detoxification.  The CSRG folk's opinion they gave
at the BSD BOF was that allthough the Jolitz, etc work
isn't there, and can't be because of political considerations, 
anyone can get it off the net and can fill in missing pieces.



>Can anyone comment on the accuracy of this report?

It's very accurate, only missing the details.

>If true, this greatly saddens me.  It looks like I have just lost any
>chance of running BSD free with sources on my DECstation.  I was really
>looking forward to running an operating system with full source code
>again ...
>
>Does anyone see any chance that either of the 386-style BSDs discussed
>here will ever be upgraded to run on real computers?  ;-)

There are seven files missing, for which should be able to get the Jolitz 
stuff to work.  This is not on the standard distribution because
Berkeley requires all contributors to sign a release, and given
the situation between Jolitz and Berkeley, this will not
happen.

The decsstation 3000 and 5000 serties will be supported platforms - 
I don't remember if they will make the alpha tape or not.

Path: sparky!uunet!spool.mu.edu!agate!pasteur!hermes.Berkeley.EDU!bostic
From: bos...@hermes.Berkeley.EDU (Keith Bostic)
Newsgroups: comp.unix.bsd
Subject: Re: Sigh - now it'll *never* be free
Keywords: 4.4BSD   CSRG
Message-ID: <1992Jun17.001544.4017@pasteur.Berkeley.EDU>
Date: 17 Jun 92 00:15:44 GMT
Article-I.D.: pasteur.1992Jun17.001544.4017
References: <41950@sunquest.UUCP> <1992Jun15.202726.21177@colorado.edu>
Sender: n...@pasteur.Berkeley.EDU (NNTP Poster)
Distribution: usa
Organization: University of California at Berkeley
Lines: 53
Nntp-Posting-Host: hermes.berkeley.edu

A couple of comments:

In article <1992Jun15.202726.21...@colorado.edu> d...@ophelia.cs.colorado.edu (D
rew Eckhardt) writes:
> Technically, this is correct too.  There will be a 4.4bsd, requiring
> at least a 32V license, as well as a 4.4bsd lite,
> or a net3 if you will - in the same state as net2 as
> far as detoxification.  The CSRG folk's opinion they gave
> at the BSD BOF was that allthough the Jolitz, etc work
> isn't there, and can't be because of political considerations,
> anyone can get it off the net and can fill in missing pieces.

4.4BSD-Lite does have some new, freely redistributable software, including
C library modules, include files and utilities, among them dd, init and
join.  It may have an implementation of ex/vi, but the odds are currently
against that happening.

> There are seven files missing, for which should be able to get the Jolitz
> stuff to work.  This is not on the standard distribution because
> Berkeley requires all contributors to sign a release, and given
> the situation between Jolitz and Berkeley, this will not
> happen.

This isn't right, and needs to be corrected.  Bill Jolitz, Pace Willison
and others have already offered to contribute their versions of the
missing modules to the CSRG to make the distribution complete.  We have
chosen not to include them for the following reasons.

The first reason is that it would take a significant amount of time to
make them work in our current system.  Time that we just don't have.  The
NET/2 release, from which the free 386 releases are derived, was a year
ago.  Our system has changed in many significant ways.  As an example,
we've added a version of Sprite's log-structured file system, LFS.  LFS
uses the buffer cache code in very "interesting" ways.  Ways that it was
never intended to be used.  It took weeks to get the bugs out of the
current LFS/buffer cache interface, and I'm not eager to do it twice,
especially not a few weeks before a release.

The second reason is that we don't want reimplementations of the current
missing functionality.  Buffer caches were great for 1972, but for 1992
I want a "buffer cache" that's fully integrated with the VM.  The same goes
for exec, ptrace and the other missing functionality.

Finally, we expect that the groups that replaced the missing functionality
in NET/2 will immediately reintegrate their source into the system, and
they can probably do it as least as quickly as we can!

So, given that the CSRG's resources are extremely limited, the choice was
between reimplementing already existing, working functionality and making
the rest of the system correct and stable.  Since the CSRG will be going
away we felt that the latter was more important.

--keith

Newsgroups: comp.unix.bsd
Path: sparky!uunet!cs.utexas.edu!utgpu!torn!watserv1!wes.on.ca!tomh
From: t...@wes.on.ca (Tom Haapanen)
Subject: Re: Sigh - now it'll *never* be free -- and ex/vi
Organization: Waterloo Engineering Software
Distribution: na
Date: Wed, 17 Jun 1992 12:10:39 GMT
Message-ID: <1992Jun17.121039.19472@wes.on.ca>
Keywords: 4.4BSD   CSRG
References: <41950@sunquest.UUCP> <1992Jun15.202726.21177@colorado.edu> 
<1992Jun17.001544.4017@pasteur.Berkeley.EDU>
Lines: 19

bos...@hermes.Berkeley.EDU (Keith Bostic) writes:
> 4.4BSD-Lite does have some new, freely redistributable software, including
> C library modules, include files and utilities, among them dd, init and
> join.  It may have an implementation of ex/vi, but the odds are currently
> against that happening.

Never having looked at the source for ex/vi (while I had BSD source access),
I'm not really on firm ground here, but would it be foolish to assume that
some (most?) of the source modules for ex/vi would by now be free of AT&T
copyrights?  And, if so, couldn't the free source modules be distributed
in the manner of NET/2, so that net.volunteers could complete a "real" free
vi (just like the 386BSD and Linux projects are being done)?  

It's nice to have stevie, vile and elvis, but they're not quite complete,
and definitely not as robust as the real thing...

[ \tom haapanen    "i don't even know what street canada is on" -- al capone ]
[ t...@wes.on.ca                   "trust the programmer" -- ansi c standard ]
[ waterloo engineering software      "to thine own self be true" -- polonius ]

Newsgroups: comp.unix.bsd
Path: sparky!uunet!wupost!think.com!ames!agate!pasteur!hermes.Berkeley.EDU!bostic
From: bos...@hermes.Berkeley.EDU (Keith Bostic)
Subject: Re: Sigh - now it'll *never* be free -- and ex/vi
Message-ID: <1992Jun18.164531.10222@pasteur.Berkeley.EDU>
Keywords: 4.4BSD   CSRG
Sender: n...@pasteur.Berkeley.EDU (NNTP Poster)
Nntp-Posting-Host: hermes.berkeley.edu
Organization: University of California at Berkeley
References: <1992Jun15.202726.21177@colorado.edu> 
<1992Jun17.001544.4017@pasteur.Berkeley.EDU> <1992Jun17.121039.19472@wes.on.ca>
Distribution: na
Date: Thu, 18 Jun 1992 16:45:31 GMT
Lines: 50

In article <1992Jun17.121039.19...@wes.on.ca> t...@wes.on.ca (Tom Haapanen) writes:
>bos...@hermes.Berkeley.EDU (Keith Bostic) writes:
>> 4.4BSD-Lite does have some new, freely redistributable software, including
>> C library modules, include files and utilities, among them dd, init and
>> join.  It may have an implementation of ex/vi, but the odds are currently
>> against that happening.
>
>Never having looked at the source for ex/vi (while I had BSD source access),
>I'm not really on firm ground here, but would it be foolish to assume that
>some (most?) of the source modules for ex/vi would by now be free of AT&T
>copyrights?  And, if so, couldn't the free source modules be distributed
>in the manner of NET/2, so that net.volunteers could complete a "real" free
>vi (just like the 386BSD and Linux projects are being done)?  

Yes, most of the sources for ex/vi are freely redistributable, but the
original file buffering code is clearly derived from ed(1).  A couple
of other things are tainted as well.  The difficulty in doing what you
suggest is that ex/vi is an incredibly convoluted beast.  Trying to
isolate the specific portions that are tainted would be difficult, to
say the least.  And, even if the tainted portions were identified,
removing them without reimplementing the entire program would be even
more difficult as they include the fundamental ex/vi data structures.

--keith


Vi trivia time:

Q:	When is the command "2x" not the same as "xx"?

Q:	What does the command "3y4y" do?

Q:	If the cursor is on the last character in a line, is the "l"
	command ever legal?

Q:	When is the command "4e" not the same as "eeee"?

Q:	In the line "abc def ghi jkl", with the cursor on the 'd', the
	command "dfg" will delete up to, and including, the 'g'.  The
	command "dtg" will delete up to, but not including, the 'g'.
	What does the command "d/g<CR>" do?

Q:	What is the movement command that, when there are no more of it's
	searched for conditions in the file, goes to the beginning of the
	last line, not the end?

Q:	What are the three (yes, three) conditions where a join does not
	insert white-space between the two lines being joined?