This is G o o g l e's cache of http://www.lp.se/ftp/mailinglists/FREE-VMS.1997-01.
G o o g l e's cache is the snapshot that we took of the page as we crawled the web.
The page may have changed since that time. Click here for the current page without highlighting.
Google is not affiliated with the authors of this page nor responsible for its content. |
| These search terms have been highlighted: | levitte | programming |
|
|
Archive-Date: Fri, 24 Jan 1997 20:30:29 +0100
Sender: owner-free-vms@lp.se
Date: Fri, 24 Jan 1997 14:29:30 -0500
Message-ID: <199701241929.OAA09888@harvey.cyclic.com>
From: Jim Kingdon <kingdonc@harvey.cyclic.com>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
BCC:
Subject: VMSBACKUP 4.1 uploaded
A new version of VMSBACKUP is available at
ftp://alpha.wku.edu/free-vms/vmsbackup4-1.tar
It will shortly be moved to the usual place at
ftp://ftp.wku.edu/vms/free-vms/testing
The version number has jumped from 3.0.1 to 4.1 not because there are
a huge quantity of changes, but because I discovered that there were
versions numbered 3.1 and 4.0 floating around (the functionality from
both of which has been merged in).
As usual, please test it and send me your bugfixes, bug reports, etc.
I thank the people who have done so for the previous versions.
Changes since version 3.0.1:
* Merged in changes from version 3.1 from S.J. Tappin of the
University of Birmingham. .HLB is now on the list of file extensions
which are ignored by default (along with .TLB, .EXE, etc.). Filenames
are matched in a case-insensitive fashion. Added -B option to
extract files in binary mode.
* Merged in the -b change from Timothy Stark's version 4.0. This lets
one set the blocksize.
================================================================================
Archive-Date: Sat, 25 Jan 1997 13:53:08 +0100
Sender: owner-free-vms@lp.se
Subject: Re: VMSBACKUP 4.1 uploaded
To: Free-VMS@lp.se
Date: Sat, 25 Jan 1997 07:53:00 -0500 (EST)
From: "christopher f. chiesa" <lvt-cfc@servtech.com>
Reply-To: Free-VMS@lp.se
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Message-ID: <32ea022c0.6e6a@cyber1.servtech.com>
Content-Transfer-Encoding: 7bit
>
> A new version of VMSBACKUP is available at
>
> ftp://alpha.wku.edu/free-vms/vmsbackup4-1.tar
>
> It will shortly be moved to the usual place at
>
> ftp://ftp.wku.edu/vms/free-vms/testing
Hmm.. This is just about the FIRST message I've EVER received from the
"free-VMS" mailing list. So I don't have any idea what you're talking
about. Is this a VMS Backup clone that runs UNDER the Free-VMS OS? I had
no idea FreeVMS was that far along, that it could actually be run and
used. If so, where do I get it? What hardware does it run on? If it's
NOT available, where/how am I supposed to run/test the announced VMSBACKUP
program? Is this a VMS-backup-saveset-reader for OTHER OS's? Or...?
Thanks in advance for clarifying this for me. I've been waiting to hear
more about this project ever since subscribing to the mailinglist a month
or two, or more, ago, but as I said, this is the FIRST message I've seen!
Chris Chiesa
lvt-cfc@servtech.com
================================================================================
Archive-Date: Tue, 28 Jan 1997 04:08:21 +0100
Sender: owner-free-vms@lp.se
Date: Mon, 27 Jan 1997 21:09:08 EST
From: Hunter Goatley <goathunter@MadGoat.COM>
Reply-To: Free-VMS@lp.se
To: FREE-VMS@LP.SE
Message-ID: <009AF01E.CC94B0B7.1@ALPHA.WKU.EDU>
Subject: Updated STR and BTRAN
Kevin Handy <kth@srv.net> writes:
>
>>>I've placed new versions of 'str' and 'btran' on ALPHA.WKU.EDU.
>>>
>>>'str' has only a couple of bug fixes.
>>>
>>>'btran' has been modified to use the 'str' routines, as well as
>>>many bug fixes and enhancements.
>>>
These .TAR files have now been placed on FTP.WKU.EDU in
[.VMS.FREE-VMS.TESTING].
Hunter
------
Hunter Goatley, Process Software Corporation (TCPware)
<goathunter@MadGoat.com> http://www.wku.edu/www/madgoat/hunter.html
================================================================================
Archive-Date: Thu, 30 Jan 1997 02:58:07 +0100
Sender: owner-free-vms@lp.se
Date: Wed, 29 Jan 1997 20:58:34 -0500
Message-ID: <199701300158.UAA06131@harvey.cyclic.com>
From: Jim Kingdon <kingdonc@harvey.cyclic.com>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Subject: Re: VMSBACKUP 4.1 uploaded
> This is just about the FIRST message I've EVER received from the
> "free-VMS" mailing list.
Welcome to the list. Things have been slow, but if we want something
to happen we all need to chip in and do the work to make Free-VMS an
active project rather than a web site and a dormant mailing list.
There is a nice task list and other information on the Free-VMS web
site, so it should be quite feasible for anyone who wants to
contribute to find something which needs doing. Please consider
picking a programming project, testing the programs currently out
there, or whatever else seems to you to need doing.
> Is this a VMS Backup clone that runs UNDER the Free-VMS OS? I had no
> idea FreeVMS was that far along, that it could actually be run and
> used.
Your suspicion is correct; Free-VMS is a very long way from running.
If you go to the FTP site you'll find VMSBACKUP, a package of STR$*
functions and a BASIC->C translator. There are some DCL clones
floating around and perhaps another piece or two elsewhere, but just
about everything still needs to be written.
> Is this a VMS-backup-saveset-reader for OTHER OS's?
VMSBACKUP runs under VMS (Free-VMS will be compatible enough that the
differences between VMS and Free-VMS probably won't matter from
VMSBACKUP's point of view) as well as a variety of other OS's. The
most practical use for it at the moment is to read BACKUP savesets on
non-VMS systems. But the plan is to enhance it (and it needs many,
many enhancements for this) to be the BACKUP program for Free-VMS.
Remember that Free-VMS needs not only an operating system kernel, but
also all the user programs like COPY, DIFFERENCES, DIRECTORY, etc.
These are very good projects for volunteers without kernel hacking
experience, as each program is self-contained and need use only
documented interfaces such as the VMS System Services.
================================================================================
Archive-Date: Thu, 30 Jan 1997 05:16:52 +0100
Sender: owner-free-vms@lp.se
Message-ID: <32F02F18.3BDF@net-link.net>
Date: Wed, 29 Jan 1997 23:18:16 -0600
From: "Todd C. Campbell" <toddc@net-link.net>
Reply-To: Free-VMS@lp.se
MIME-Version: 1.0
To: Free-VMS@lp.se
Subject: Just a thought
References: <199701300158.UAA06131@harvey.cyclic.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Has anybody looked at Linux and it's binary compatiblity with VMS???
It does exist, and the source is very available. I understand that these
OSes are very different...I guess I just want to see a bootable
kernel...I feel that I cannot contribute anything untill such a kernel
has been constructed.
================================================================================
Archive-Date: Thu, 30 Jan 1997 14:24:56 +0100
Sender: owner-free-vms@lp.se
Message-ID: <1.5.4.32.19970130132447.0069a188@mail-1.ns.net>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Thu, 30 Jan 1997 05:24:47 -0800
To: Free-VMS@lp.se
From: John Mee <jmee@ns.net>
Reply-To: Free-VMS@lp.se
Subject: Re: Just a thought
At 23:18 1/29/97 -0600, you wrote:
>Has anybody looked at Linux and it's binary compatiblity with VMS???
>It does exist, and the source is very available. I understand that these
>OSes are very different...I guess I just want to see a bootable
>kernel...I feel that I cannot contribute anything untill such a kernel
>has been constructed.
I had this discussion with a friend at work yesterday. My thought is that
FreeVMS ought to be constructed to run natively, rather than on top of
another OS. He pointed out that there is a version of Linux that runs on top
of MACH.
My concern is that as MACH changes, then it could/would affect FreeVMS.
Further, it would seem that putting it in as a native OS would make porting
to different platforms easier (e.g. Alpha????). I realize that it would
substantially increase initial development time since there are all those
nifty calls that wouldn't have to be coded.
================================================================================
Archive-Date: Thu, 30 Jan 1997 16:58:11 +0100
Sender: owner-free-vms@lp.se
Date: Thu, 30 Jan 1997 10:51:55 -0500
Message-ID: <199701301551.KAA23018@harvey.cyclic.com>
From: Jim Kingdon <kingdonc@harvey.cyclic.com>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Subject: Re: Just a thought
> Has anybody looked at Linux and it's binary compatiblity with VMS???
Wait. What binary compatibility? I'm reasonably tuned into linux and
I haven't heard anything about any VMS binary (or source)
compatibility package.
If you are saying that such a package could be written, sure a lot of
things could be written and one of the favorite pasttimes of this list
seems to be to discuss the "best" architecture for a VMS-compatible
operating system. But speaking for myself, I'll pay more attention to
such discussions if they are backed up by reasonably firm offers to
write code, made by people with time and skills to do so.
> I guess I just want to see a bootable kernel...I feel that I cannot
> contribute anything untill such a kernel has been constructed.
Oh, there is a lot you can contribute. Take a closer look at the task
list. Just to pick one random example, enhance GNU diff (which
already runs on a variety of operating systems) to have the output
format of VMS DIFFERENCES. And/or look into the literature on
algorithms for diff programs and figure out what is involved in adding
the traditional VMS algorithm (I don't have a strong opinion one way
or the other about the necessity of this--GNU and unix use different
algorithms but people seem to think they are compatible enough--but at
least one poster to this list thinks the GNU algorithm is too
different from the VMS one for use in Free-VMS).
================================================================================
Archive-Date: Thu, 30 Jan 1997 18:06:49 +0100
Sender: owner-free-vms@lp.se
Message-ID: <v03007803af168305d51d@[140.186.88.14]>
References: <32ea022c0.6e6a@cyber1.servtech.com> (lvt-cfc@servtech.com)
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Thu, 30 Jan 1997 11:59:19 -0500
To: Free-VMS@lp.se
From: Dick Munroe <munroe@acornsw.com>
Reply-To: Free-VMS@lp.se
Subject: Re: VMSBACKUP 4.1 uploaded
>> Is this a VMS-backup-saveset-reader for OTHER OS's?
>
>VMSBACKUP runs under VMS (Free-VMS will be compatible enough that the
>differences between VMS and Free-VMS probably won't matter from
>VMSBACKUP's point of view) as well as a variety of other OS's. The
>most practical use for it at the moment is to read BACKUP savesets on
>non-VMS systems. But the plan is to enhance it (and it needs many,
>many enhancements for this) to be the BACKUP program for Free-VMS.
Which brings up an interesting question wrt Backup in particular. Acorn
manufactures robotics software for tape libraries, some of our partners
manufacture software for backup management, we all are thinking about the
NT/U*x/Mac space and I'm just wondering if a port of the Free VMS backup to
any of these OSs (in the context of the appropriate copy left stuff) would
be (1) possible in the legal sense and (2) of interest to members of the
Free VMS group. I, for one, am underwhelmed with the quality of the tape
archiving available with pretty much everything save OpenVMS and IBM's
larger boxes.
Dick Munroe
--
Dick Munroe Internet: munroe@acornsw.com
Acorn Software, Inc.
267 Cox St. Office: (508) 568-1618 x1
Hudson, Ma. 01749 USA FAX: (508) 562-1133
================================================================================
Archive-Date: Thu, 30 Jan 1997 18:53:35 +0100
Sender: owner-free-vms@lp.se
Date: Thu, 30 Jan 1997 10:53:33 -0700
Message-ID: <97013010533285@spike.cs.mci.com>
From: tucker@spike.cs.mci.com (TUCKER)
Reply-To: Free-VMS@lp.se
To: FREE-VMS@lp.se
I just found out about freevms a few weeks ago, and I've been
reading about Mach, including the prototype port of VMS to
Mach by Digital. This has been very interesting!
I really want to help, but I haven't seen any activity lately.
Is there a news group for freevms? If not, how can we start one?
I sure you have already talked about this, but why not write
FreeVMS along the lines of the current OpenVMS but rewrite it in C,
removing hardware dependencies using a Hardware Abstraction Layer
(HAL) type concept. You could do development on a current alpha
running OpenVMS replacing parts of the kernel until it was portable,
using the standard VMS debugger, linker, etc. until it was finished.
It would have to be suitably different to not get in trouble with
DEC. But then again, DEC may not care if we are advancing
VMS, and helping them to sell larger clustered VMS systems.
If something doesn't happen, VMS may slowly become like many
of those other forgotten operating systems (RSTS, TOPS, etc).
I guess I'm just not sold on using Mach yet, where each VMS
process must use a pair of Mach tasks. It's also so different,
from the VMS kernel that I'm not sure how or where to begin.
To work on freeVMS, I really need an alpha at home and I would
love it to boot VMS. When I was at the last DECUS there was a presentation
about how to build your own low cost alpha system that would boot VMS.
This presentation was to include a parts list, using off the
shelf IDE hard disks, memory, etc. The spokes-person to give
this presentation never showed up and left about 20 to 30
people wondering what he was going to say. DEC during the
closing session also indicated that they would make a hobbiest
license of OpenVMS for home use available for cheep (around $100).
Does anyone have any information on building or buying your own
alpha PC that would boot openVMS for a reasonable cost?
Has anyone heard of any information on a hobbiest license for openVMS?
Thanks,
Roger Tucker (tucker@spike.cs.mci.com)
Thanks,
Roger Tucker (tucker@spike.cs.mci.com)
================================================================================
Archive-Date: Thu, 30 Jan 1997 19:19:48 +0100
Sender: owner-free-vms@lp.se
Message-ID: <c=US%a=_%p=msft%l=RED-70-MSG-970130181738Z-1384@INET-03-IMC.itg.microsoft.com>
From: Chris Kaler <ckaler@microsoft.com>
Reply-To: Free-VMS@lp.se
To: "'Free-VMS@lp.se'" <Free-VMS@lp.se>
Subject: RE: VMSBACKUP 4.1 uploaded
Date: Thu, 30 Jan 1997 10:17:38 -0800
I believe that everything in FreeVMS to date is copyright free,
so I wouldn't imagine that is a problem. If VMSBACKUP is
based on the source floating around the Web, then it already
runs on several platforms.
We are still trying to get the VMS-On-Mach and Intrepid
sources released to the public domain for non-commercial
use, although this is more complicated now that Bill Davenport
has left Digital as well!
Included in these sources is a multi-threaded portable backup
utility that I wrote, which might help... if we ever get them freed.
Christopher
-----Original Message-----
From: Dick Munroe [SMTP:munroe@acornsw.com]
Sent: Thursday, January 30, 1997 8:59 AM
To: Free-VMS@lp.se
Subject: Re: VMSBACKUP 4.1 uploaded
>> Is this a VMS-backup-saveset-reader for OTHER OS's?
>
>VMSBACKUP runs under VMS (Free-VMS will be compatible enough that the
>differences between VMS and Free-VMS probably won't matter from
>VMSBACKUP's point of view) as well as a variety of other OS's. The
>most practical use for it at the moment is to read BACKUP savesets on
>non-VMS systems. But the plan is to enhance it (and it needs many,
>many enhancements for this) to be the BACKUP program for Free-VMS.
Which brings up an interesting question wrt Backup in particular. Acorn
manufactures robotics software for tape libraries, some of our partners
manufacture software for backup management, we all are thinking about the
NT/U*x/Mac space and I'm just wondering if a port of the Free VMS backup to
any of these OSs (in the context of the appropriate copy left stuff) would
be (1) possible in the legal sense and (2) of interest to members of the
Free VMS group. I, for one, am underwhelmed with the quality of the tape
archiving available with pretty much everything save OpenVMS and IBM's
larger boxes.
Dick Munroe
--
Dick Munroe Internet: munroe@acornsw.com
Acorn Software, Inc.
267 Cox St. Office: (508) 568-1618 x1
Hudson, Ma. 01749 USA FAX: (508) 562-1133
================================================================================
Archive-Date: Thu, 30 Jan 1997 19:22:38 +0100
Sender: owner-free-vms@lp.se
Message-ID: <c=US%a=_%p=msft%l=RED-70-MSG-970130182130Z-1427@INET-04-IMC.microsoft.com>
From: Chris Kaler <ckaler@microsoft.com>
Reply-To: Free-VMS@lp.se
To: "'Free-VMS@lp.se'" <Free-VMS@lp.se>
Subject: RE: Just a thought
Date: Thu, 30 Jan 1997 10:21:30 -0800
> Has anybody looked at Linux and it's binary compatiblity with VMS???
Wait. What binary compatibility? I'm reasonably tuned into linux and
I haven't heard anything about any VMS binary (or source)
compatibility package.
CK> Neither have I
If you are saying that such a package could be written, sure a lot of
things could be written and one of the favorite pasttimes of this list
seems to be to discuss the "best" architecture for a VMS-compatible
operating system. But speaking for myself, I'll pay more attention to
such discussions if they are backed up by reasonably firm offers to
write code, made by people with time and skills to do so.
CK> As I have ranted and raved before, I do think
CK> it is important to build a strong architecture, and
CK> I would try NOT to write a kernel.
CK> P.S. I'd love to contribute, but I am about as
CK> tainted as they come. We are still trying to get
CK> our previous sources released from DEC. Everybody
CK> think happy thoughts ;-)
> I guess I just want to see a bootable kernel...I feel that I cannot
> contribute anything untill such a kernel has been constructed.
Oh, there is a lot you can contribute. Take a closer look at the task
list. Just to pick one random example, enhance GNU diff (which
already runs on a variety of operating systems) to have the output
format of VMS DIFFERENCES. And/or look into the literature on
algorithms for diff programs and figure out what is involved in adding
the traditional VMS algorithm (I don't have a strong opinion one way
or the other about the necessity of this--GNU and unix use different
algorithms but people seem to think they are compatible enough--but at
least one poster to this list thinks the GNU algorithm is too
different from the VMS one for use in Free-VMS).
CK> I couldn't agree more! If you look at the
CK> evolution of FreeBSD you will see that the
CK> first step is to replace all of the utilities with
CK> free versions... the infrastructure can come later.
CK> Of course, it is often nice to have some of the
CK> RTLs available early on.
================================================================================
Archive-Date: Thu, 30 Jan 1997 19:24:13 +0100
Sender: owner-free-vms@lp.se
Message-ID: <c=US%a=_%p=msft%l=RED-70-MSG-970130182204Z-1433@INET-03-IMC.itg.microsoft.com>
From: Chris Kaler <ckaler@microsoft.com>
Reply-To: Free-VMS@lp.se
To: "'Free-VMS@lp.se'" <Free-VMS@lp.se>
Subject: FW: Just a thought
Date: Thu, 30 Jan 1997 10:22:04 -0800
> Has anybody looked at Linux and it's binary compatiblity with VMS???
Wait. What binary compatibility? I'm reasonably tuned into linux and
I haven't heard anything about any VMS binary (or source)
compatibility package.
CK> Neither have I
If you are saying that such a package could be written, sure a lot of
things could be written and one of the favorite pasttimes of this list
seems to be to discuss the "best" architecture for a VMS-compatible
operating system. But speaking for myself, I'll pay more attention to
such discussions if they are backed up by reasonably firm offers to
write code, made by people with time and skills to do so.
CK> As I have ranted and raved before, I do think
CK> it is important to build a strong architecture, and
CK> I would try NOT to write a kernel.
CK> P.S. I'd love to contribute, but I am about as
CK> tainted as they come. We are still trying to get
CK> our previous sources released from DEC. Everybody
CK> think happy thoughts ;-)
> I guess I just want to see a bootable kernel...I feel that I cannot
> contribute anything untill such a kernel has been constructed.
Oh, there is a lot you can contribute. Take a closer look at the task
list. Just to pick one random example, enhance GNU diff (which
already runs on a variety of operating systems) to have the output
format of VMS DIFFERENCES. And/or look into the literature on
algorithms for diff programs and figure out what is involved in adding
the traditional VMS algorithm (I don't have a strong opinion one way
or the other about the necessity of this--GNU and unix use different
algorithms but people seem to think they are compatible enough--but at
least one poster to this list thinks the GNU algorithm is too
different from the VMS one for use in Free-VMS).
CK> I couldn't agree more! If you look at the
CK> evolution of FreeBSD you will see that the
CK> first step is to replace all of the utilities with
CK> free versions... the infrastructure can come later.
CK> Of course, it is often nice to have some of the
CK> RTLs available early on.
================================================================================
Archive-Date: Thu, 30 Jan 1997 19:26:12 +0100
Sender: owner-free-vms@lp.se
Message-ID: <c=US%a=_%p=msft%l=RED-70-MSG-970130182539Z-1459@INET-05-IMC.microsoft.com>
From: Chris Kaler <ckaler@microsoft.com>
Reply-To: Free-VMS@lp.se
To: "'Free-VMS@lp.se'" <Free-VMS@lp.se>
Subject: RE:
Date: Thu, 30 Jan 1997 10:25:39 -0800
I sure you have already talked about this, but why not write
FreeVMS along the lines of the current OpenVMS but rewrite it in C,
removing hardware dependencies using a Hardware Abstraction Layer
(HAL) type concept. You could do development on a current alpha
running OpenVMS replacing parts of the kernel until it was portable,
using the standard VMS debugger, linker, etc. until it was finished.
CK> Writing a kernel is not a good use of your time...
CK> trust me ;-)
It would have to be suitably different to not get in trouble with
DEC. But then again, DEC may not care if we are advancing
VMS, and helping them to sell larger clustered VMS systems.
If something doesn't happen, VMS may slowly become like many
of those other forgotten operating systems (RSTS, TOPS, etc).
CK> DEC is concerned about highly secure, robust, available
CK> clusters. I don't you are going to challenge them in
CK> the near future. As long as you don't violate copyrights,
CK> I don't think they will care.
I guess I'm just not sold on using Mach yet, where each VMS
process must use a pair of Mach tasks. It's also so different,
from the VMS kernel that I'm not sure how or where to begin.
CK> Me either. That's why we did Intrepid. I believe the
CK> white paper is on the FreeVMS site.
================================================================================
Archive-Date: Thu, 30 Jan 1997 20:00:05 +0100
Sender: owner-free-vms@lp.se
Message-ID: <v03007811af169db6852a@[140.186.88.14]>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Thu, 30 Jan 1997 13:53:02 -0500
To: Free-VMS@lp.se
From: Dick Munroe <munroe@acornsw.com>
Reply-To: Free-VMS@lp.se
Subject: Re:
>To work on freeVMS, I really need an alpha at home and I would
>love it to boot VMS. When I was at the last DECUS there was a presentation
>about how to build your own low cost alpha system that would boot VMS.
>This presentation was to include a parts list, using off the
>shelf IDE hard disks, memory, etc. The spokes-person to give
>this presentation never showed up and left about 20 to 30
>people wondering what he was going to say. DEC during the
>closing session also indicated that they would make a hobbiest
>license of OpenVMS for home use available for cheep (around $100).
>Does anyone have any information on building or buying your own
>alpha PC that would boot openVMS for a reasonable cost?
>Has anyone heard of any information on a hobbiest license for openVMS?
Depends on what you mean by reasonable. My last universal box (nt/u*x/vms)
axp was bought for about 4K, it was loaded but that includes a pretty steep
developer's discount.
I'm seeing multia's going up for auction at www.onsale.com for between 600
and 1300 bucks. But they don't run VMS. I haven't done any reasearch as
to WHY. I'll leave that to the rest of the folks on the list.
I suspect that if you do some poking around on the net you should be able
to come up with a 3000-300 cheaply (we're still using one we bought 5 years
ago as a development engine) but it doesn't run nt/u*x...
Everything's a tradeoff
Dick
--
Dick Munroe Internet: munroe@acornsw.com
Acorn Software, Inc.
267 Cox St. Office: (508) 568-1618 x1
Hudson, Ma. 01749 USA FAX: (508) 562-1133
================================================================================
Archive-Date: Thu, 30 Jan 1997 20:20:14 +0100
Sender: owner-free-vms@lp.se
MIME-Version: 1.0
Content-Type: Text/Plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Date: Thu, 30 Jan 1997 20:20:07 MET
Message-ID: <13041.631.271759.feedmail.pl5.LEVITTE@devil.bofh.se>
From: Richard Levitte - VMS Whacker <levitte@lp.se>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Subject: Re: VMSBACKUP 4.1 uploaded
From: Dick Munroe <munroe@acornsw.com>
Which brings up an interesting question wrt Backup in particular. Acorn
manufactures robotics software for tape libraries, some of our partners
manufacture software for backup management, we all are thinking about the
NT/U*x/Mac space and I'm just wondering if a port of the Free VMS backup to
any of these OSs (in the context of the appropriate copy left stuff) would
be (1) possible in the legal sense and (2) of interest to members of the
Free VMS group. I, for one, am underwhelmed with the quality of the tape
archiving available with pretty much everything save OpenVMS and IBM's
larger boxes.
1) I don't see the problem, since it's supposed to be free.
2) Well, I'll say the same as RMS says about GNU software: the programs
for Free-VMS are (or should be) built to run on Free-VMS before anything
else. For programs like VMSBACKUP, it means the should run on VMS,
since Free-VMS is supposed to be compatible at that source level.
If Free-VMS programs are enhanced to work on other OS's as well, that's
fine with me, but is really up to the author of that particular program.
This is something I neither encourage or discourage.
--
R Levitte, Levitte Programming; Spannv. 38, I; S-161 43 Bromma; SWEDEN
Tel: +46-8-26 52 47; Cel: +46-10-222 64 05; No fax right now
PGP key fingerprint = A6 96 C0 34 3A 96 AA 6C B0 D5 9A DF D2 E9 9C 65
Check http://www.lp.se/~levitte for my public key. bastard@bofh.se
================================================================================
Archive-Date: Thu, 30 Jan 1997 20:30:51 +0100
Sender: owner-free-vms@lp.se
MIME-Version: 1.0
Content-Type: Text/Plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Date: Thu, 30 Jan 1997 20:30:44 MET
Message-ID: <13041.1268.271759.feedmail.pl5.LEVITTE@devil.bofh.se>
From: Richard Levitte - VMS Whacker <levitte@lp.se>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Subject: RE:
From: Chris Kaler <ckaler@microsoft.com>
CK> Me either. That's why we did Intrepid. I believe the
CK> white paper is on the FreeVMS site.
Look in http://www.lp.se/free-vms/docs/
--
R Levitte, Levitte Programming; Spannv. 38, I; S-161 43 Bromma; SWEDEN
Tel: +46-8-26 52 47; Cel: +46-10-222 64 05; No fax right now
PGP key fingerprint = A6 96 C0 34 3A 96 AA 6C B0 D5 9A DF D2 E9 9C 65
Check http://www.lp.se/~levitte for my public key. bastard@bofh.se
================================================================================
Archive-Date: Thu, 30 Jan 1997 21:10:07 +0100
Sender: owner-free-vms@lp.se
Date: Thu, 30 Jan 1997 14:45:24 -0500 (EST)
From: Elliot Lee <sopwith@cuc.edu>
Reply-To: Free-VMS@lp.se
To: "'Free-VMS@lp.se'" <Free-VMS@lp.se>
Subject: Re: FW: Just a thought
Message-ID: <Pine.LNX.3.95.970130143537.1843E-100000@helix.cs.cuc.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
On Thu, 30 Jan 1997, Chris Kaler wrote:
> > Has anybody looked at Linux and it's binary compatiblity with VMS???
>
> Wait. What binary compatibility? I'm reasonably tuned into linux and
> I haven't heard anything about any VMS binary (or source)
> compatibility package.
It doesn't exist, but it is theoretically possible. Linux (and *BSD, to be
perfectly honest) can handle
> CK> As I have ranted and raved before, I do think
> CK> it is important to build a strong architecture, and
> CK> I would try NOT to write a kernel.
>
> CK> P.S. I'd love to contribute, but I am about as
> CK> tainted as they come. We are still trying to get
> CK> our previous sources released from DEC. Everybody
> CK> think happy thoughts ;-)
>
> > I guess I just want to see a bootable kernel...I feel that I cannot
> > contribute anything untill such a kernel has been constructed.
>
> Oh, there is a lot you can contribute. Take a closer look at the task
> list. Just to pick one random example, enhance GNU diff (which
> already runs on a variety of operating systems) to have the output
> format of VMS DIFFERENCES. And/or look into the literature on
And what are you going to run this vmsdiff on? UNIX?
> CK> I couldn't agree more! If you look at the
> CK> evolution of FreeBSD you will see that the
> CK> first step is to replace all of the utilities with
> CK> free versions... the infrastructure can come later.
So let's just make a big house of cards and build it on air.
If someone comes along with (a) a basically working kernel and (b) a
basically working DCL(sp?) shell, I'll jump on their bandwagon, because
from experience I know that they are much closer to having a working
VMS-like system.
Top down implementation won't work here, folks. You have to have something
to work on. The main advantages that I see to VMS (clustering,
scalability, reliability) take a kernel to implement, not a diff that
works VMS-style ;-)
My $0.02, coming from a Linux user's point of view.
Hope this helps,
-- Elliot http://www.redhat.com/
"I'm a member of the Association of Federations of Linux Project
Initiators That Never Really Get Much Done (AFLPITNRGMD, for short)."
Just do it!
================================================================================
Archive-Date: Thu, 30 Jan 1997 22:01:06 +0100
Sender: owner-free-vms@lp.se
Message-ID: <c=US%a=_%p=msft%l=RED-70-MSG-970130210022Z-2497@INET-02-IMC.microsoft.com>
From: Chris Kaler <ckaler@microsoft.com>
Reply-To: Free-VMS@lp.se
To: "'Free-VMS@lp.se'" <Free-VMS@lp.se>
Subject: RE: FW: Just a thought
Date: Thu, 30 Jan 1997 13:00:22 -0800
I hear you, but . . . .
If we had taken that approach, we never would have gotten
anything done with either VMS-on-Mach or Intrepid. Instead,
we chose to exploit parallelism. Now our scale was somewhat
limited because we only had five people. The Web offers the
potential for some great scaling.
I would propose doing the following in parallel:
1) build user-level applications
2) work on APIs
3) work on DCL
4) work on environment
Now you need to have some clue as to what you are going to build
before starting 3 and 4. And some apps in 1 won't work until you have
2. But you can cheat. Build the apps in 1 on VMS leveraging the API.
This (a) gets them built, and (b) provides a great test bed for the people
working on 2. How you choose to implement 3 will depend on what you
are doing in 4. But the point is that 1 does not depend on 4.
This was our approach and I would highly recommend it. You make
forward progress very quickly. Which is important for keeping people
interested.
Christopher
-----Original Message-----
From: Elliot Lee [SMTP:sopwith@cuc.edu]
Sent: Thursday, January 30, 1997 11:45 AM
To: 'Free-VMS@lp.se'
Subject: Re: FW: Just a thought
On Thu, 30 Jan 1997, Chris Kaler wrote:
> > Has anybody looked at Linux and it's binary compatiblity with VMS???
>
> Wait. What binary compatibility? I'm reasonably tuned into linux and
> I haven't heard anything about any VMS binary (or source)
> compatibility package.
It doesn't exist, but it is theoretically possible. Linux (and *BSD, to be
perfectly honest) can handle
> CK> As I have ranted and raved before, I do think
> CK> it is important to build a strong architecture, and
> CK> I would try NOT to write a kernel.
>
> CK> P.S. I'd love to contribute, but I am about as
> CK> tainted as they come. We are still trying to get
> CK> our previous sources released from DEC. Everybody
> CK> think happy thoughts ;-)
>
> > I guess I just want to see a bootable kernel...I feel that I cannot
> > contribute anything untill such a kernel has been constructed.
>
> Oh, there is a lot you can contribute. Take a closer look at the task
> list. Just to pick one random example, enhance GNU diff (which
> already runs on a variety of operating systems) to have the output
> format of VMS DIFFERENCES. And/or look into the literature on
And what are you going to run this vmsdiff on? UNIX?
> CK> I couldn't agree more! If you look at the
> CK> evolution of FreeBSD you will see that the
> CK> first step is to replace all of the utilities with
> CK> free versions... the infrastructure can come later.
So let's just make a big house of cards and build it on air.
If someone comes along with (a) a basically working kernel and (b) a
basically working DCL(sp?) shell, I'll jump on their bandwagon, because
from experience I know that they are much closer to having a working
VMS-like system.
Top down implementation won't work here, folks. You have to have something
to work on. The main advantages that I see to VMS (clustering,
scalability, reliability) take a kernel to implement, not a diff that
works VMS-style ;-)
My $0.02, coming from a Linux user's point of view.
Hope this helps,
-- Elliot http://www.redhat.com/
"I'm a member of the Association of Federations of Linux Project
Initiators That Never Really Get Much Done (AFLPITNRGMD, for short)."
Just do it!
================================================================================
Archive-Date: Thu, 30 Jan 1997 23:49:41 +0100
Sender: owner-free-vms@lp.se
MIME-Version: 1.0
Content-Type: Text/Plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Date: Thu, 30 Jan 1997 23:49:35 MET
Message-ID: <13041.13199.271759.feedmail.pl5.LEVITTE@devil.bofh.se>
From: Richard Levitte - VMS Whacker <levitte@lp.se>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Subject: Re: FW: Just a thought
From: Elliot Lee <sopwith@cuc.edu>
On Thu, 30 Jan 1997, Chris Kaler wrote:
> > Has anybody looked at Linux and it's binary compatiblity with VMS???
>
> Wait. What binary compatibility? I'm reasonably tuned into linux and
> I haven't heard anything about any VMS binary (or source)
> compatibility package.
It doesn't exist, but it is theoretically possible. Linux (and *BSD, to be
perfectly honest) can handle
Care to rephrase that? Or rather, to expand?
> Oh, there is a lot you can contribute. Take a closer look at the task
> list. Just to pick one random example, enhance GNU diff (which
> already runs on a variety of operating systems) to have the output
> format of VMS DIFFERENCES. And/or look into the literature on
And what are you going to run this vmsdiff on? UNIX?
No, on VMS. Well, if someone wants to port it to Unix, that's to him/her.
> CK> I couldn't agree more! If you look at the
> CK> evolution of FreeBSD you will see that the
> CK> first step is to replace all of the utilities with
> CK> free versions... the infrastructure can come later.
So let's just make a big house of cards and build it on air.
If someone comes along with (a) a basically working kernel and (b) a
basically working DCL(sp?) shell, I'll jump on their bandwagon, because
from experience I know that they are much closer to having a working
VMS-like system.
Oh, come off it, will you? This project isn't only about writing a kernel.
It isn't only about writing a CLI. It isn't only about writing a lot of
utilities. It IS about all of these things.
I'll repeat it again and again, Free VMS is supposed to be at least
compatible in source with VMS down to the calls of system routines.
That's at least one of the goals I set up from the very beginning.
This means that utilities that you can compile on VMS today are expected
to compile on Free VMS with no or very little change (the changes may
be due to using another compiler).
I do not think you are going to tell us that you have use for a kernel
and a DCL if there aren't any utilities around. Say COPY, or SETP0, or
SETP1 (yes, that one'll be a little tougher...). To take Linux as an
example, would it have been any fun without all the GNU utilities that
come with the regular distributions today?
If you give this some thought, you'll find that this project has some
similarities with the GNU project. The GNU project was about making
a Unix-like system that would be freely available to everyone. What
really happened is that a LOT of utilities were built, from big programs
like emacs downto the tiniest shell utility like cp or cat, and the
kernel is still in beta (or is it alpha?), but is beginning to get a
lot of attention latelly. The HURD took about ten years of development,
if not more, or at least that's what I heard. Do you really expect this
project to be very different (except I hope we'll not take some ten years,
at least if we get the head start we hope for)? I don't.
As a matter of fact, I'd be rather amused to see a VMS box where everything
but the kernel is free software :-).
--
R Levitte, Levitte Programming; Spannv. 38, I; S-161 43 Bromma; SWEDEN
Tel: +46-8-26 52 47; Cel: +46-10-222 64 05; No fax right now
PGP key fingerprint = A6 96 C0 34 3A 96 AA 6C B0 D5 9A DF D2 E9 9C 65
Check http://www.lp.se/~levitte for my public key. bastard@bofh.se
================================================================================
Archive-Date: Fri, 31 Jan 1997 00:55:16 +0100
Sender: owner-free-vms@lp.se
Date: Thu, 30 Jan 1997 18:57:06 -0500 (EST)
From: EVERHART@Arisia.GCE.Com
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Message-ID: <970130185706.62@Arisia.GCE.Com>
Subject: vms on multia...
Agreed multia's cheap. But the bus stuff isn't there. Supposedly
the bus routines from the laptop are ALMOST what you need, but not
quite. At that time, the hardware guys weren't asking for VMS support.
That has supposedly changed and they ALL want it. Some mighty accurate
rumors have circulated...
================================================================================
Archive-Date: Fri, 31 Jan 1997 04:02:22 +0100
Sender: owner-free-vms@lp.se
Date: Thu, 30 Jan 1997 22:02:58 -0500
Message-ID: <199701310302.WAA05750@harvey.cyclic.com>
From: Jim Kingdon <kingdonc@harvey.cyclic.com>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Subject: Re: VMSBACKUP 4.1 uploaded
> We are still trying to get the VMS-On-Mach and Intrepid
> sources released to the public domain for non-commercial use
Well, I think eventually we'll want something which can be used
commercially.
Of course, we could always start with that and replace it one piece at
a time, sort of the way that unix turned into BSD turned into
FreeBSD. So it would not be a bad thing if the sources are released
for non-commercial use; it just wouldn't be as good a thing as if
something shows up which is OK for commercial use too.
================================================================================
Archive-Date: Fri, 31 Jan 1997 04:16:38 +0100
Sender: owner-free-vms@lp.se
Date: Thu, 30 Jan 1997 22:17:13 -0500
Message-ID: <199701310317.WAA05820@harvey.cyclic.com>
From: Jim Kingdon <kingdonc@harvey.cyclic.com>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Subject: How to help on Free-VMS
> I really want to help, but I haven't seen any activity lately.
Well, have you read the task list?
I mean, if you read it and don't see any projects on it you like, or
you have questions about what kind of help you could get from other
people, or whatever, please do ask. But the task list and other items
on the web site go into a fair bit of detail so I hope that they are
helpful for people who are considering various projects.
> To work on freeVMS, I really need an alpha at home and I would
> love it to boot VMS.
Well, it depends on what part of the system you want to work on. For
example, Kevin Handy's STR$* routines run on VMS, but they run just as
well on linux (or other systems running gcc) if you compile with
-fdollars-in-identifiers. Granted, for some subsystems it isn't quite
that easy. But I wouldn't go so far as to say that everyone who wants
to help MUST start out by getting a VMS system of their own. Having a
VMS system will probably make things easier, but each person needs to
balance the temporary workarounds involved in not having VMS versus
the difficulty of getting VMS.
Guest accounts on VMS systems are also probably available, although I
don't want to put any words in the mouths of the people who might be
willing to offer them--I'm sure they don't want to get inundated with
requests.
> It would have to be suitably different to not get in trouble with DEC.
I would recommend following the law, as best we are able, and not
getting too paranoid about this. Most of what we want to implement is
covered by published sources such as the VMS documentation anyway.
================================================================================
Archive-Date: Fri, 31 Jan 1997 04:21:49 +0100
Sender: owner-free-vms@lp.se
Date: Thu, 30 Jan 1997 22:22:25 -0500
Message-ID: <199701310322.WAA05846@harvey.cyclic.com>
From: Jim Kingdon <kingdonc@harvey.cyclic.com>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Subject: Re: VMSBACKUP 4.1 uploaded
I think it would be great if Acorn finds some use for VMSBACKUP, and I
would be glad to work with you towards integrating any enhancements
you might make back into the "official" release.
However, I do need to warn you: VMSBACKUP is *well* short of the
functionality of BACKUP. For example, it doesn't write savesets at
all, and there are many aspects of reading them that it doesn't do
either. If you want to enhance it, great. But I would be a little
bit surprised if it would be the "quality of tape archiving" you are
looking for, as it exists now.
================================================================================
Archive-Date: Fri, 31 Jan 1997 05:43:21 +0100
Sender: owner-free-vms@lp.se
Message-ID: <v03007810af17202a8aa7@[199.232.78.21]>
References: <97013010533285@spike.cs.mci.com> (tucker@spike.cs.mci.com)
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Thu, 30 Jan 1997 23:08:52 -0500
To: Free-VMS@lp.se
From: Dick Munroe <munroe@acornsw.com>
Reply-To: Free-VMS@lp.se
Subject: Re: How to help on Free-VMS
>Well, it depends on what part of the system you want to work on. For
>example, Kevin Handy's STR$* routines run on VMS, but they run just as
>well on linux (or other systems running gcc) if you compile with
>-fdollars-in-identifiers. Granted, for some subsystems it isn't quite
>that easy. But I wouldn't go so far as to say that everyone who wants
>to help MUST start out by getting a VMS system of their own. Having a
>VMS system will probably make things easier, but each person needs to
>balance the temporary workarounds involved in not having VMS versus
>the difficulty of getting VMS.
And its cheaper than you expect. I'm pretty able to find 3100s (systems,
including disk, keyboard, mouse, and monitor) for < $1500 which is STILL
the cheapest VMS system I'be been able to locate reliably. It's not the
Taj Mahal, but it WILL run OpenVMS and you can develope pretty much
everything you need.
For that matter, I've got a couple of them lying around now that are pretty
much unused. If someone wants to buy one of them, I'll make you a good
deal... Once I've made sure they still work.
Dick
--
Dick Munroe munroe@acornsw.com
Acorn Software (508) 568 1618 x1
267 Cox St. FAX: 562 1133
Hudson, Ma. 01749 http://www.acornsw.com/
Need a web site? a web server? other web service? Contact us...
"They told me to get Windows 3.1 or better so I bought a Macintosh!"
"Then they told me to get Windows 95 or better...so I bought another Mac!"
"Now they tell me to get Windows NT 4.0 Server or better...I've ordered
another Mac!"
================================================================================
Archive-Date: Fri, 31 Jan 1997 05:43:37 +0100
Sender: owner-free-vms@lp.se
Message-ID: <v03007811af1720f7bad7@[199.232.78.21]>
References: <v03007803af168305d51d@[140.186.88.14]> (message from Dick Munroe on Thu, 30 Jan 1997 11:59:19 -0500)
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Thu, 30 Jan 1997 23:15:05 -0500
To: Free-VMS@lp.se
From: Dick Munroe <munroe@acornsw.com>
Reply-To: Free-VMS@lp.se
Subject: Re: VMSBACKUP 4.1 uploaded
>I think it would be great if Acorn finds some use for VMSBACKUP, and I
>would be glad to work with you towards integrating any enhancements
>you might make back into the "official" release.
>
>However, I do need to warn you: VMSBACKUP is *well* short of the
>functionality of BACKUP. For example, it doesn't write savesets at
>all, and there are many aspects of reading them that it doesn't do
>either. If you want to enhance it, great. But I would be a little
>bit surprised if it would be the "quality of tape archiving" you are
>looking for, as it exists now.
Well, mostly what I'm looking for now is something that is basically a bit
stream converter, i.e., takes a bit stream in, applies the "appropriate"
blocking, checksumming, etc. on it and pumps out a bit stream that is a
"saveset". Done right (and a reader should be relatively easily turned
into a writer using this model) you should be able to pretty much
completely divorce the system specific characteristics of most of the
styles of backup from the code. Yes, physical and image backup would and
should be relatively system specific but I would think that everything else
could be relatively easily done... (VIGOROUS WAVING OF HANDS for those of
you who think I speak too quickly).
Low performance, implementation first to get the writer working. Higher
speed front and back ends to handle things like multiple I/O streams on the
input side, multiple buffering on the output side later. The saveset
engine SHOULD be relatively robust and independent from the front and back
ends given this sort of an architecture.
Dick
--
Dick Munroe munroe@acornsw.com
Acorn Software (508) 568 1618 x1
267 Cox St. FAX: 562 1133
Hudson, Ma. 01749 http://www.acornsw.com/
Need a web site? a web server? other web service? Contact us...
"They told me to get Windows 3.1 or better so I bought a Macintosh!"
"Then they told me to get Windows 95 or better...so I bought another Mac!"
"Now they tell me to get Windows NT 4.0 Server or better...I've ordered
another Mac!"
================================================================================
Archive-Date: Fri, 31 Jan 1997 08:02:52 +0100
Sender: owner-free-vms@lp.se
Date: Fri, 31 Jan 97 01:44:02 EST
Message-ID: <009AF2A0B3130BA0.0000251C@ais.com>
From: bruce@ais.com
Reply-To: Free-VMS@lp.se
Subject: Re: FW: Just a thought
To: Free-VMS@lp.se
Elliot Lee wrote:
On Thu, 30 Jan 1997, Chris Kaler wrote:
>> > I guess I just want to see a bootable kernel...I feel that I cannot
>> > contribute anything untill such a kernel has been constructed.
>>
>> Oh, there is a lot you can contribute. Take a closer look at the task
>> list. Just to pick one random example, enhance GNU diff (which
>> already runs on a variety of operating systems) to have the output
>> format of VMS DIFFERENCES. And/or look into the literature on
>
>And what are you going to run this vmsdiff on? UNIX?
>
>> CK> I couldn't agree more! If you look at the
>> CK> evolution of FreeBSD you will see that the
>> CK> first step is to replace all of the utilities with
>> CK> free versions... the infrastructure can come later.
>
>So let's just make a big house of cards and build it on air.
>
>If someone comes along with (a) a basically working kernel and (b) a
>basically working DCL(sp?) shell, I'll jump on their bandwagon, because
>from experience I know that they are much closer to having a working
>VMS-like system.
>
>Top down implementation won't work here, folks. You have to have something
>to work on. The main advantages that I see to VMS (clustering,
>scalability, reliability) take a kernel to implement, not a diff that
>works VMS-style ;-)
>
>My $0.02, coming from a Linux user's point of view.
I think this attitude is unrealistic. Even DEC's VMS didn't support
clustering, for example, for many years after its initial release; if
you require that a "working kernel" have clustering, scalability, and
reliability (but not necessarily any software that runs on it) before
you're interested in it, you'll wait quite a while. Ideally, yes, it
would be nice to have a kernel, but this will require a large effort
before there is *anything* useful at any level out of it. You would
also of course need an appropriate ACP for the file system, and to get
DCL to work you'd also want RMS and the RTL to work since DCL makes
heavy use of them. The RTL would probably also be useful for the ACPs
though not as necessary. In fact in the case of many utilities there
are DECUS and GNU programs floating around that could be used instead
of the standard DEC utilities; there may be minor differences in many
cases but I don't know how many people would find an *exact* look-alike
to be all that necessary. I for one would be quite happy with a Unix-
style diff rather than a VMS diff; the more important issue would be
the ability to port VMS software to the FreeVMS environment.
Alternatively a very simple kernel could be written withOUT provision
for anything like file systems, paging, DCL etc, but which had a VMS-like
API interface (SYS$*) -- sort of a Free-VAXeln, a real-time monitor for
things like embedded systems. Such a thing could eventually grow into a
more useable system, but don't kid yourself that it would look much like
a fully loaded DEC VMS box. Would anyone be interested in something
like this for Intel boxes, for example? Personlly I don't think I'd be
able to find much use for it except to try to extend it, but there might
be someone out there who would.
My opinion is that the highest priority is the API (both the SYS$ as
well as the various RTL components such as LIB$, MTH$, STR$, SMG$, etc),
not either the utilities or the kernel. You will need the API regardless
of what architecture the rest of the system has; it might even be useful
on Unix or Windows NT to make a VMS-like environment there. You will
need the API to implement both the utilities and some parts of the
system-level code such as the ACPs and DCL and possibly even some parts
of the kernel (particularly SYS$). You will also need the API to port
any other software (such as your own or DECUS software) that was written
for VMS. A version of the API that ran, for example, under Windows NT
and that allowed a significant number of DECUS utilities to compile and
run under NT would be extremely useful, and long before we could ever
put together a kernel that has all the attributes of the kernel that
DEC has been polishing for 20 years now with resources that we can only
dream of getting. Get some flavor of DCL to run on this as well and
you're well on your way to having a VMS-like environment on a non-DEC
box; then putting a new kernel under this becomes more attractive.
But you get some useful output far sooner than diving in and trying to
build the lowest level first. Frankly, my opinion is that by the time
you have a kernel and [full] DCL (which implies the RTL), even without
having any of the utilities, the problem is practically solved! Seems
rather late to be volunteering only after that point is reached ...
The analogy with Linux is only partly relevant. Linux has the distinct
advantage that there was all this GNU and other free Unix software that
was already floating around the net, so there was little need to put
together anything *other* than a kernel - you just had to make it run
the already existing utilities. But Unix has very little in the way
of an API: most of it is either kernel calls (which you can provide in
a new kernel) or C runtime calls (which are provided in the GNU C
libraries). There's nothing equivalent to the VMS RTL - something that
applications need that's not provided by either the kernel proper nor
by the language runtime libraries.
Anyway that's my two cents worth. I might be able to find the time to
do something on the API, especially considering that having such things
available on other platforms would often make some of what I do easier;
but there's no way I would be able to make the enormous commitment to
help build a kernel _first_ -- the payback is just too far off, the
amount of (unpaid!) work is too great before I or anyone else could get
much out of it, and I have other deadlines to meet.
Bruce C. Wright
================================================================================
Archive-Date: Fri, 31 Jan 1997 16:05:32 +0100
Sender: owner-free-vms@lp.se
MIME-Version: 1.0
Content-Type: Text/Plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Date: Fri, 31 Jan 1997 16:05:28 MET
Message-ID: <13042.6216.271759.feedmail.pl5.LEVITTE@devil.bofh.se>
From: Richard Levitte - VMS Whacker <levitte@lp.se>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Subject: Re: How to help on Free-VMS
From: Jim Kingdon <kingdonc@harvey.cyclic.com>
Guest accounts on VMS systems are also probably available, although I
don't want to put any words in the mouths of the people who might be
willing to offer them--I'm sure they don't want to get inundated with
requests.
Such a system is being prepared. As a matter of fact, I should have
finished the preparations before christmas, but other events came in
my way.
What is planned is really a double system: one computer you can log in
to, and that is "crash friendly", meaning that if someone screws up, he
or she will be forgiven, since this will be expected (after all, we'll
do quite a lot of experimenting, right? :-)). The other computer will
not be possible for anyone (well, almost) to log in to, but it will serve
one disk through VMS-to-VMS NFS (at least that's what's planned right now).
That machine will serve as FTP and HTTP server for the project, using
the data on the served disk. Or at least that's how I've understood the
plan (I'm currently just responsible of getting the "crash friendly"
system up and running). If I'm wrong, the owner of those computers will
probably correct me with a swift slapp :-).
--
R Levitte, Levitte Programming; Spannv. 38, I; S-161 43 Bromma; SWEDEN
Tel: +46-8-26 52 47; Cel: +46-10-222 64 05; No fax right now
PGP key fingerprint = A6 96 C0 34 3A 96 AA 6C B0 D5 9A DF D2 E9 9C 65
Check http://www.lp.se/~levitte for my public key. bastard@bofh.se
================================================================================
Archive-Date: Fri, 31 Jan 1997 16:23:04 +0100
Sender: owner-free-vms@lp.se
MIME-Version: 1.0
Content-Type: Text/Plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Date: Fri, 31 Jan 1997 16:23:01 MET
Message-ID: <13042.7269.271759.feedmail.pl5.LEVITTE@devil.bofh.se>
From: Richard Levitte - VMS Whacker <levitte@lp.se>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Subject: Re: FW: Just a thought
From: bruce@ais.com
Anyway that's my two cents worth. I might be able to find the time to
do something on the API, especially considering that having such things
available on other platforms would often make some of what I do easier;
That's a deal that could work out very well! If I understand what you're
saying correctly, this could mean that you could develop at least an embryo
to an API (be it SYS$, LIB$ or whatever$) for whatever project you need
them for, and let us have a piece of the cake.
I'm all for this kind of way to do it, BUT there's one thing to watch out
for, and that's what some companies call "intelectual property" (am I
correct?), and they are not necessarely willing to part from it. This
means that if anyone of you want to build something as part of their
job, and later release it to the Free-VMS project, you'd better talk to
your boss about it FIRST!
As a matter of fact, RMS usually asks contributors to the GNU project to
sign a paper, and to have their boss sign a disclaimer saying they will
not claim any rights to the program given to the FSF. I'm not exactly
sure I want to go that far, but this might be something to keep in mind.
I dearly prefer to have the Free-VMS project stay out of legal trouble.
If anyone is more informed in such matters, please speak up!
--
R Levitte, Levitte Programming; Spannv. 38, I; S-161 43 Bromma; SWEDEN
Tel: +46-8-26 52 47; Cel: +46-10-222 64 05; No fax right now
PGP key fingerprint = A6 96 C0 34 3A 96 AA 6C B0 D5 9A DF D2 E9 9C 65
Check http://www.lp.se/~levitte for my public key. bastard@bofh.se
================================================================================
Archive-Date: Fri, 31 Jan 1997 18:58:55 +0100
Sender: owner-free-vms@lp.se
Message-ID: <c=US%a=_%p=msft%l=RED-70-MSG-970131175718Z-6116@INET-04-IMC.microsoft.com>
From: Chris Kaler <ckaler@microsoft.com>
Reply-To: Free-VMS@lp.se
To: "'Free-VMS@lp.se'" <Free-VMS@lp.se>
Subject: RE: VMSBACKUP 4.1 uploaded
Date: Fri, 31 Jan 1997 09:57:18 -0800
> We are still trying to get the VMS-On-Mach and Intrepid
> sources released to the public domain for non-commercial use
Well, I think eventually we'll want something which can be used
commercially.
Of course, we could always start with that and replace it one piece at
a time, sort of the way that unix turned into BSD turned into
FreeBSD. So it would not be a bad thing if the sources are released
for non-commercial use; it just wouldn't be as good a thing as if
something shows up which is OK for commercial use too.
CK> Well you run into some interesting problems. First, using
CK> the Digital stuff would taint your product. That is not a
CK> pretty battle... ask the FreeBSD folks, they spent some
CK> time fighting that one.
CK> More to the point... if you are planning to build a commercial
CK> product then (a) the name shouldn't be FreeVMS, and (b) you
CK> will not get help from people like me and Bill. For myself,
CK> it would be improper because of my time at DEC and, more
CK> importantly, I fundamentally disagree. Digital makes a
CK> perfectly good VMS. FreeVMS goes after a different market.
CK> try to compete and you are just going to piss people off.
================================================================================
Archive-Date: Fri, 31 Jan 1997 19:01:25 +0100
Sender: owner-free-vms@lp.se
Message-ID: <c=US%a=_%p=msft%l=RED-70-MSG-970131180035Z-6144@INET-05-IMC.microsoft.com>
From: Chris Kaler <ckaler@microsoft.com>
Reply-To: Free-VMS@lp.se
To: "'Free-VMS@lp.se'" <Free-VMS@lp.se>
Subject: RE: VMSBACKUP 4.1 uploaded
Date: Fri, 31 Jan 1997 10:00:35 -0800
>I think it would be great if Acorn finds some use for VMSBACKUP, and I
>would be glad to work with you towards integrating any enhancements
>you might make back into the "official" release.
>
>However, I do need to warn you: VMSBACKUP is *well* short of the
>functionality of BACKUP. For example, it doesn't write savesets at
>all, and there are many aspects of reading them that it doesn't do
>either. If you want to enhance it, great. But I would be a little
>bit surprised if it would be the "quality of tape archiving" you are
>looking for, as it exists now.
Well, mostly what I'm looking for now is something that is basically a bit
stream converter, i.e., takes a bit stream in, applies the "appropriate"
blocking, checksumming, etc. on it and pumps out a bit stream that is a
"saveset". Done right (and a reader should be relatively easily turned
into a writer using this model) you should be able to pretty much
completely divorce the system specific characteristics of most of the
styles of backup from the code. Yes, physical and image backup would and
should be relatively system specific but I would think that everything else
could be relatively easily done... (VIGOROUS WAVING OF HANDS for those of
you who think I speak too quickly).
Low performance, implementation first to get the writer working. Higher
speed front and back ends to handle things like multiple I/O streams on the
input side, multiple buffering on the output side later. The saveset
engine SHOULD be relatively robust and independent from the front and back
ends given this sort of an architecture.
CK> This is very similar to an A/D project I did. The idea was to have
CK> a generic framework into which you plugged readers and writers.
CK> Readers converted format X to a canonical form, writers converted
CK> the canonical form to a specific form. It is part of the source set
CK> we are trying to get released.
================================================================================
Archive-Date: Fri, 31 Jan 1997 19:24:38 +0100
Sender: owner-free-vms@lp.se
Message-ID: <c=US%a=_%p=msft%l=RED-70-MSG-970131182358Z-6307@INET-05-IMC.microsoft.com>
From: Chris Kaler <ckaler@microsoft.com>
Reply-To: Free-VMS@lp.se
To: "'Free-VMS@lp.se'" <Free-VMS@lp.se>
Subject: RE: FW: Just a thought
Date: Fri, 31 Jan 1997 10:23:58 -0800
So this thread has been coming and going since this mailing list
started. To be honest, I'm tired of it, so I am turning on my
flame thrower ;-).
It is totally insane to even try to write a kernel from scratch.
First of all, there are very few people who can do it right, and
I can probably count them on my fingers (and yes, I have a
standard set of digits). Second, it takes a shitlaod of time.
Any idea how long the NT kernel took? The Mach kernel?
Third, building a kernel that supports VMS... well NT doesn't
event do it all yet!
I don't really want to be mean, but it seems really clear to
me that the people who are pushing to write a kernel either
(a) have never written a real O/S kernel before, or (b) really
like that sort of thing. For the people in (a), you need to
listen to people who have been there. For the people in (b),
go for it, but I don't think FreeVMS is the write venue for you.
That, at least, is my understanding of the FreeVMS charter.
So all I can do is offer some paternal advice. I have been
there. I have climbed the mountain... oops, wrong speech.
Seriously, we had some really smart people think about
this for a while and we decided to leverage an existing
kernel. It increases portability. And unless you have about
50 people to throw at full-time kernel development, you
probably shouldn't even try. Unless you really understand
memory management, paging, scheduling, and io subsystems,
you really shouldn't try. If you really think you understand it,
then we could most likely throw several curves at you that
you never expected (unless you have worked on a real
kernel before). People spend their lives writing
O/S kernels... it is not an easy thing to do as a side project.
I believe what FreeVMS is trying to achieve is the VMS user
(and developer) experience. That has almost nothing to
do with the kernel. Focus on the user experience - the tools
and the run-time environment. Kernels will come and go.
If your environment is layered on an abstracted API, then
you can port to any kernel. That is the theory of micro-kernels.
The shortcoming of this is that you don't always get the
performance. Mach had to pull video and network drivers
back into the kernel to get commercial speed. But let them
worry about that.
Just my two cents...
-----Original Message-----
From: bruce@ais.com [SMTP:bruce@ais.com]
Sent: Thursday, January 30, 1997 10:44 PM
To: Free-VMS@lp.se
Subject: Re: FW: Just a thought
Elliot Lee wrote:
On Thu, 30 Jan 1997, Chris Kaler wrote:
>> > I guess I just want to see a bootable kernel...I feel that I cannot
>> > contribute anything untill such a kernel has been constructed.
>>
>> Oh, there is a lot you can contribute. Take a closer look at the task
>> list. Just to pick one random example, enhance GNU diff (which
>> already runs on a variety of operating systems) to have the output
>> format of VMS DIFFERENCES. And/or look into the literature on
>
>And what are you going to run this vmsdiff on? UNIX?
>
>> CK> I couldn't agree more! If you look at the
>> CK> evolution of FreeBSD you will see that the
>> CK> first step is to replace all of the utilities with
>> CK> free versions... the infrastructure can come later.
>
>So let's just make a big house of cards and build it on air.
>
>If someone comes along with (a) a basically working kernel and (b) a
>basically working DCL(sp?) shell, I'll jump on their bandwagon, because
>from experience I know that they are much closer to having a working
>VMS-like system.
>
>Top down implementation won't work here, folks. You have to have something
>to work on. The main advantages that I see to VMS (clustering,
>scalability, reliability) take a kernel to implement, not a diff that
>works VMS-style ;-)
>
>My $0.02, coming from a Linux user's point of view.
I think this attitude is unrealistic. Even DEC's VMS didn't support
clustering, for example, for many years after its initial release; if
you require that a "working kernel" have clustering, scalability, and
reliability (but not necessarily any software that runs on it) before
you're interested in it, you'll wait quite a while. Ideally, yes, it
would be nice to have a kernel, but this will require a large effort
before there is *anything* useful at any level out of it. You would
also of course need an appropriate ACP for the file system, and to get
DCL to work you'd also want RMS and the RTL to work since DCL makes
heavy use of them. The RTL would probably also be useful for the ACPs
though not as necessary. In fact in the case of many utilities there
are DECUS and GNU programs floating around that could be used instead
of the standard DEC utilities; there may be minor differences in many
cases but I don't know how many people would find an *exact* look-alike
to be all that necessary. I for one would be quite happy with a Unix-
style diff rather than a VMS diff; the more important issue would be
the ability to port VMS software to the FreeVMS environment.
Alternatively a very simple kernel could be written withOUT provision
for anything like file systems, paging, DCL etc, but which had a VMS-like
API interface (SYS$*) -- sort of a Free-VAXeln, a real-time monitor for
things like embedded systems. Such a thing could eventually grow into a
more useable system, but don't kid yourself that it would look much like
a fully loaded DEC VMS box. Would anyone be interested in something
like this for Intel boxes, for example? Personlly I don't think I'd be
able to find much use for it except to try to extend it, but there might
be someone out there who would.
My opinion is that the highest priority is the API (both the SYS$ as
well as the various RTL components such as LIB$, MTH$, STR$, SMG$, etc),
not either the utilities or the kernel. You will need the API regardless
of what architecture the rest of the system has; it might even be useful
on Unix or Windows NT to make a VMS-like environment there. You will
need the API to implement both the utilities and some parts of the
system-level code such as the ACPs and DCL and possibly even some parts
of the kernel (particularly SYS$). You will also need the API to port
any other software (such as your own or DECUS software) that was written
for VMS. A version of the API that ran, for example, under Windows NT
and that allowed a significant number of DECUS utilities to compile and
run under NT would be extremely useful, and long before we could ever
put together a kernel that has all the attributes of the kernel that
DEC has been polishing for 20 years now with resources that we can only
dream of getting. Get some flavor of DCL to run on this as well and
you're well on your way to having a VMS-like environment on a non-DEC
box; then putting a new kernel under this becomes more attractive.
But you get some useful output far sooner than diving in and trying to
build the lowest level first. Frankly, my opinion is that by the time
you have a kernel and [full] DCL (which implies the RTL), even without
having any of the utilities, the problem is practically solved! Seems
rather late to be volunteering only after that point is reached ...
The analogy with Linux is only partly relevant. Linux has the distinct
advantage that there was all this GNU and other free Unix software that
was already floating around the net, so there was little need to put
together anything *other* than a kernel - you just had to make it run
the already existing utilities. But Unix has very little in the way
of an API: most of it is either kernel calls (which you can provide in
a new kernel) or C runtime calls (which are provided in the GNU C
libraries). There's nothing equivalent to the VMS RTL - something that
applications need that's not provided by either the kernel proper nor
by the language runtime libraries.
Anyway that's my two cents worth. I might be able to find the time to
do something on the API, especially considering that having such things
available on other platforms would often make some of what I do easier;
but there's no way I would be able to make the enormous commitment to
help build a kernel _first_ -- the payback is just too far off, the
amount of (unpaid!) work is too great before I or anyone else could get
much out of it, and I have other deadlines to meet.
Bruce C. Wright
================================================================================
Archive-Date: Fri, 31 Jan 1997 20:13:20 +0100
Sender: owner-free-vms@lp.se
Date: Fri, 31 Jan 1997 11:11:39 -0800
From: <jdg@rahul.net>
Reply-To: Free-VMS@lp.se
Message-ID: <199701311911.AA16764@foxtrot.rahul.net>
To: Free-VMS@lp.se
Subject: Re: How to help on Free-VMS
> For that matter, I've got a couple of them lying around now that are pretty
> much unused. If someone wants to buy one of them, I'll make you a good
> deal... Once I've made sure they still work.
I'm probably interested, but would like to know more. Thanks!
John David Galt
408 559-4615
jdg@rahul.net
================================================================================
Archive-Date: Fri, 31 Jan 1997 21:14:00 +0100
Sender: owner-free-vms@lp.se
Message-ID: <32F260F8.56ED@net-link.net>
Date: Fri, 31 Jan 1997 15:15:36 -0600
From: "Todd C. Campbell" <toddc@net-link.net>
Reply-To: Free-VMS@lp.se
MIME-Version: 1.0
To: Free-VMS@lp.se
Subject: Re: FW: Just a thought
References: <c=US%a=_%p=msft%l=RED-70-MSG-970131182358Z-6307@INET-05-IMC.microsoft.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Chris Kaler wrote:
>
> So this thread has been coming and going since this mailing list
> started. To be honest, I'm tired of it, so I am turning on my
> flame thrower ;-).
>
> It is totally insane to even try to write a kernel from scratch.
> First of all, there are very few people who can do it right, and
> I can probably count them on my fingers (and yes, I have a
> standard set of digits). Second, it takes a shitlaod of time.
> Any idea how long the NT kernel took? The Mach kernel?
> Third, building a kernel that supports VMS... well NT doesn't
> event do it all yet!
I do understand what you are trying to say. The problem, however, is that aside
from the VMSBACKUP that may or may not actually work on FreeVMS, I have seen little
work out of this group. There seems to be less and less orginization as the months go
on. If there is to be a kernel that is to be built upon, then which one is it, where
can it be downloaded from and what platforms will it run on? I have offered both
machines and T1 to DS3 access, I have offered newsgroups or what ever it will take to
get a little more orginized discussions, I have had no response.
Yes, I have heard about binary compatiblity with FreeBSD and VMS, but now that I have
been told there is no such animal, I really have nothing to say. It seems like many
people would like to contribute their time, but they do not have the experience that
some of the rest of the group may have. Don't slam these people in front of the list,
e-mail that person privately, help them. Otherwise this group will only become a place
where FreeVMS will exist in theory, and you will scare off those lurkers that
aren't quite sure what they would like or be able to help with, nothing more will come
of it if we continue to punce on those who speak up with an idea or opinion.(I sure hope
you don't think you have all the answers, I know I don't)
================================================================================
Archive-Date: Fri, 31 Jan 1997 21:53:01 +0100
Sender: owner-free-vms@lp.se
Date: Fri, 31 Jan 97 15:06:22 EST
Message-ID: <009AF310C8A31BC0.00002493@ais.com>
From: bruce@ais.com
Reply-To: Free-VMS@lp.se
Subject: RE: FW: Just a thought
To: Free-VMS@lp.se
Chris Kaler writes in reply to my message:
>So this thread has been coming and going since this mailing list
>started. To be honest, I'm tired of it, so I am turning on my
>flame thrower ;-).
>
>It is totally insane to even try to write a kernel from scratch.
>First of all, there are very few people who can do it right, and
>I can probably count them on my fingers (and yes, I have a
>standard set of digits). Second, it takes a shitlaod of time.
>Any idea how long the NT kernel took? The Mach kernel?
>Third, building a kernel that supports VMS... well NT doesn't
>event do it all yet!
>
>I don't really want to be mean, but it seems really clear to
>me that the people who are pushing to write a kernel either
>(a) have never written a real O/S kernel before, or (b) really
>like that sort of thing. For the people in (a), you need to
>listen to people who have been there. For the people in (b),
>go for it, but I don't think FreeVMS is the write venue for you.
>That, at least, is my understanding of the FreeVMS charter.
As far as I'm concerned you're preaching to the choir. I too have
"been there, done that," and have done my share of kernel hacking
on several systems going back to PDP-11 days. I'm quite aware of
how much effort goes into writing a kernel; that's one reason I'm
not very enthusiastic about getting into a project that will be a
time sink for years before there's much useful output at all. I
will admit that I do like to do kernel hacking but I like even more
to get something useful actually accomplished.
Bruce C. Wright
================================================================================
Archive-Date: Fri, 31 Jan 1997 21:53:12 +0100
Sender: owner-free-vms@lp.se
Date: Fri, 31 Jan 97 15:39:22 EST
Message-ID: <009AF31564F23A20.00002493@ais.com>
From: bruce@ais.com
Reply-To: Free-VMS@lp.se
Subject: Re: FW: Just a thought
To: Free-VMS@lp.se
Richard Levitte writes:
> From: bruce@ais.com
>
> Anyway that's my two cents worth. I might be able to find the time to
> do something on the API, especially considering that having such things
> available on other platforms would often make some of what I do easier;
>
>That's a deal that could work out very well! If I understand what you're
>saying correctly, this could mean that you could develop at least an embryo
>to an API (be it SYS$, LIB$ or whatever$) for whatever project you need
>them for, and let us have a piece of the cake.
>
>I'm all for this kind of way to do it, BUT there's one thing to watch out
>for, and that's what some companies call "intelectual property" (am I
>correct?), and they are not necessarely willing to part from it. This
>means that if anyone of you want to build something as part of their
>job, and later release it to the Free-VMS project, you'd better talk to
>your boss about it FIRST!
Yes, I'm quite aware of intellectual property rights - I am part owner
of a small software shop :-). I can ensure that nothing I contribute
will be encumbered in that way.
One problem in this whole area is the issue of calling standards: most
systems don't pass a parameter count, and often don't have anything
analogous to the VMS signal mechanism. There had been some discussion
about this some time ago but I haven't seen anything lately; this is
something that will need compiler mods to do correctly. Does anyone
know what the current state of affairs is on any possible compiler
mods to support varying numbers of arguments? Many routines don't use
argument counts and for these components the issue is moot, but there
are a lot of routines that do allow variable argument counts and in
order to do them correctly we need a compiler first. (The signal
mechanism will also be needed before too long but is less immediately
critical). If nobody's working on that yet I suppose that's one place
to start, as well as some of the RTL/SYS$ calls that don't allow
variable length arguments.
Bruce C. Wright
================================================================================
Archive-Date: Fri, 31 Jan 1997 21:54:04 +0100
Sender: owner-free-vms@lp.se
MIME-Version: 1.0
Content-Type: Text/Plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Date: Fri, 31 Jan 1997 21:53:59 MET
Message-ID: <13042.27127.271759.feedmail.pl5.LEVITTE@devil.bofh.se>
From: Richard Levitte - VMS Whacker <levitte@lp.se>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Subject: Re: How to help on Free-VMS
> For that matter, I've got a couple of them lying around now that are pretty
> much unused. If someone wants to buy one of them, I'll make you a good
> deal... Once I've made sure they still work.
I'm probably interested, but would like to know more. Thanks!
It's far from me to discourage someone wanting to sell his surplus VAXen
(that's one of the reasons I didn't protest against THAT post), but please,
keep negociations outside this list, OK?
Thank you.
--
R Levitte, Levitte Programming; Spannv. 38, I; S-161 43 Bromma; SWEDEN
Tel: +46-8-26 52 47; Cel: +46-10-222 64 05; No fax right now
PGP key fingerprint = A6 96 C0 34 3A 96 AA 6C B0 D5 9A DF D2 E9 9C 65
Check http://www.lp.se/~levitte for my public key. bastard@bofh.se
================================================================================
Archive-Date: Fri, 31 Jan 1997 21:55:35 +0100
Sender: owner-free-vms@lp.se
Date: Fri, 31 Jan 1997 15:55:53 -0500
From: leichter@smarts.com (Jerry Leichter)
Reply-To: Free-VMS@lp.se
Message-ID: <199701312055.PAA06124@just.smarts.com>
To: Free-VMS@lp.se
Subject: Kernels
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
I completely agree with Chris Kaler that writing a kernel from scratch would be a bad idea. (Well, he said "totally insane", which
may be going too far.) Totally new kernels are few and far between. And I
don't know of *any* new kernel developed from scratch in a distributed fashion.
(Remember that Linux started out as the work of one talented maniac; by the time
others got on board, they could at least start with a system that would boot!)
I've seen two suggestions for kernels to start with: Linux and Mach. The
problem with Linux is that it's all under the Gnu copyleft, and some people
(myself included) object to that. Mach has already been used as the basis for
one experiment, and is free; but it's a pretty minimal system.
I'd like to suggest an alternative: 4.4BSD. This is a completely free system
that's actually been ported to more hardware platforms than Linux - there are
ports available for PC's, Alphas, and VAXes. (The availability of a VAX port is
nice since it would allow someone who owns a cheap VAX to run either a BSD-based
FreeVMS or true VAX VMS - handy for development.) The system comes with both IP
and OSI stacks - handy. It has a supports multiple file systems. It's imported
from Mach some ideas on memory and process management. On the other hand, since
it was developed on the VAX, it to this days shows vestiges of VAX influences in
its design - which for FreeVMS would be a useful thing, since the pieces of the
VAX that influence the design are exactly those that were created in conjunction
with the VMS designers. Finally, there's a book describing the design and
implementation, so it's relatively easy to get started. (I'm talking, BTW, of
the new version of the book, describing 4.4BSD. There is an older one
describing 4.3BSD.)
Many things would have to be changed, at all levels, to turn a BSD kernel into a
VMS-like kernel. But they could be done gradually. As a first cut, the user-
visible VM architectures are pretty similar, and could be left alone. I/O
differences should probably wait for a second pass. Instead, I'd start with
things like event flags, lock management (actually, 4.4BSD *has* a file lock
manager), AST's (BSD's "reliable signals" are a good deal of the way to true
AST's already). It would be ideal if one could run within either a VMS or a BSD
"universe", so that at least at first you could use Unix the old tools.
-- Jerry
================================================================================
Archive-Date: Fri, 31 Jan 1997 22:12:01 +0100
Sender: owner-free-vms@lp.se
MIME-Version: 1.0
Content-Type: Text/Plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Date: Fri, 31 Jan 1997 22:11:55 MET
Message-ID: <13042.28203.271759.feedmail.pl5.LEVITTE@devil.bofh.se>
From: Richard Levitte - VMS Whacker <levitte@lp.se>
Reply-To: Free-VMS@lp.se
To: Free-VMS@lp.se
Subject: Re: FW: Just a thought
From: "Todd C. Campbell" <toddc@net-link.net>
can it be downloaded from and what platforms will it run on? I have offered
both machines and T1 to DS3 access,
There have been several offerings. I'm personally involved in one. This
also requires some organization, and I don't think it's a bad thing if
development is spread over several machines. Question is, how do YOU want
to organize your machine(s). Do you give a free account to everyone who
wants to develop some piece of software for Free-VMS and thinks his capable?
What kind of development tools will you have inside? Or is your offering
mostly meant to work as an FTP archive? I think such questions still remain
unanswered. Since it's your machine(s), it's your decision how you want them
to be used. I'm open to many suggestions.
I have offered newsgroups or what ever it will take to
get a little more orginized discussions, I have had no response.
I'm not sure, and I think I recall this was the concensus from those who
spoke out at the time, that it's too early to think about newsgroups.
Many experienced news managers will tell you that you will either have to
put up with a lot of spam (we see a lot of that in most open newsgroups),
or have it moderated (which requires a moderator, anyone willing?), or have
pretty good control over the paths that the articles will take (also called
"leak control"). The last of the three is not very desirable, since it takes
a toll on the news managers involved, and in any case doesn't give us much
more than a bunch of mailing lists.
I don't mind newsgroups, but I don't think that's something we should rush
into just yet.
--
R Levitte, Levitte Programming; Spannv. 38, I; S-161 43 Bromma; SWEDEN
Tel: +46-8-26 52 47; Cel: +46-10-222 64 05; No fax right now
PGP key fingerprint = A6 96 C0 34 3A 96 AA 6C B0 D5 9A DF D2 E9 9C 65
Check http://www.lp.se/~levitte for my public key. bastard@bofh.se
================================================================================
Archive-Date: Fri, 31 Jan 1997 22:22:20 +0100
Sender: owner-free-vms@lp.se
Message-ID: <1.5.4.32.19970131212521.006a61dc@snake.srv.net>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Fri, 31 Jan 1997 14:25:21 -0700
To: Free-VMS@lp.se
From: Kevin Handy <kth@srv.net>
Reply-To: Free-VMS@lp.se
Subject: Re: How to help on Free-VMS
At 11:08 PM 1/30/97 -0500, you wrote:
>>Well, it depends on what part of the system you want to work on. For
>>example, Kevin Handy's STR$* routines run on VMS, but they run just as
>>well on linux (or other systems running gcc) if you compile with
>>-fdollars-in-identifiers. Granted, for some subsystems it isn't quite
>>that easy. But I wouldn't go so far as to say that everyone who wants
>>to help MUST start out by getting a VMS system of their own. Having a
>>VMS system will probably make things easier, but each person needs to
>>balance the temporary workarounds involved in not having VMS versus
>>the difficulty of getting VMS.
>
>And its cheaper than you expect. I'm pretty able to find 3100s (systems,
>including disk, keyboard, mouse, and monitor) for < $1500 which is STILL
>the cheapest VMS system I'be been able to locate reliably. It's not the
>Taj Mahal, but it WILL run OpenVMS and you can develope pretty much
>everything you need.
>
One thing that those who are uncomfortable coding in a section of
FreeVMS could do, is to do testing of those routines that do exist.
Much of what I've released in the STR$ library is untested, but what
I've tried has worked for me. I'm limited in the time I can devote
to this project, and have concentrated more on getting it coded than
extensively testing it.
There are also a lot of system/library routines in VMS that are very
simple that could also be easily written, even if very roughly.
Release them, and allow others to evaluate/improve them.
Once enough are available, they can be used for developing
more complex funtions/programs.
I think that we can create a large subset of the library that
is Operating System/Machine independent, so that anyone with a
C compilerand a VMS manual could make a beneficial addition,
even without having a VMS system.
-------------------------------------------------------------
Kevin Handy kth@srv.ne Accounting Software for
Software Solutions. Inc. VAX/VMS Computer Systems
Idaho Falls, Idaho
================================================================================
Archive-Date: Fri, 31 Jan 1997 22:30:30 +0100
Sender: owner-free-vms@lp.se
Message-ID: <1.5.4.32.19970131213326.006a8574@snake.srv.net>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Fri, 31 Jan 1997 14:33:26 -0700
To: Free-VMS@lp.se
From: Kevin Handy <kth@srv.net>
Reply-To: Free-VMS@lp.se
Subject: Re: FW: Just a thought
At 03:39 PM 1/31/97 EST, you wrote:
>Richard Levitte writes:
>
>One problem in this whole area is the issue of calling standards: most
>systems don't pass a parameter count, and often don't have anything
>analogous to the VMS signal mechanism. There had been some discussion
>about this some time ago but I haven't seen anything lately; this is
>something that will need compiler mods to do correctly. Does anyone
>know what the current state of affairs is on any possible compiler
>mods to support varying numbers of arguments? Many routines don't use
>argument counts and for these components the issue is moot, but there
>are a lot of routines that do allow variable argument counts and in
>order to do them correctly we need a compiler first. (The signal
>mechanism will also be needed before too long but is less immediately
>critical). If nobody's working on that yet I suppose that's one place
>to start, as well as some of the RTL/SYS$ calls that don't allow
>variable length arguments.
>
For now, in the STR$ routines I've put up, I'm ignoring the 'optional
arguement' capability of the fields, and coding everything as 'required
arguements'. Many other routines could be writen like this for now,
and once this capability is designed, be fixed for compatibility.
Those passed variable number of parameters are not possible at the
moment.
-------------------------------------------------------------
Kevin Handy kth@srv.ne Accounting Software for
Software Solutions. Inc. VAX/VMS Computer Systems
Idaho Falls, Idaho
================================================================================
Archive-Date: Fri, 31 Jan 1997 22:34:59 +0100
Sender: owner-free-vms@lp.se
MIME-Version: 1.0
Content-Type: Text/Plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Date: Fri, 31 Jan 1997 22:34:53 MET
Message-ID: <13042.29581.271759.feedmail.pl5.LEVITTE@devil.bofh.se>
From: Richard Levitte - VMS Whacker <levitte@lp.se>
Reply-To: Free-VMS@lp.se
To: free-vms@lp.se
Subject: Kernels and stuff...
I feel like we're going over a discussion that we have already had, and
that we're back to personal preferences, and some flame throwing. Back
in the autumn, I think the final concensus seemed to be that we would
use Mach. Now, 4.4BSD is thrown into the game. I've never looked into
that kernel, so I don't know if it's any good. I have looked in other
kernels (Linux, most notably), and my guts has told me (yep, I listen to
my guts quite a lot. They're often right) that a Unix kernel didn't
seem exactly feasable to build a Free-VMS kernel upon.
I might be wrong in my assumption, but it seems to me that a micro-kernel
like Mach would be the easiest to build upon, among other reasons because
you don't need to wade through it too much to decide what to keep, what
to throw away and what to rebuild.
And (and I beg forgiveness from those who want things done fast) I don't
yet mind this discussion going on, but we WILL have to decide for good
sooner or later. And I'll honestly say that I tend to lend a larger ear
to those who have done this kind of work before (VMS-on-Mach). But not
very much, I assure you.
--
R Levitte, Levitte Programming; Spannv. 38, I; S-161 43 Bromma; SWEDEN
Tel: +46-8-26 52 47; Cel: +46-10-222 64 05; No fax right now
PGP key fingerprint = A6 96 C0 34 3A 96 AA 6C B0 D5 9A DF D2 E9 9C 65
Check http://www.lp.se/~levitte for my public key. bastard@bofh.se
================================================================================
Archive-Date: Fri, 31 Jan 1997 23:36:26 +0100
Sender: owner-free-vms@lp.se
From: Bill/Carolyn Pechter <pechter@shell.monmouth.com>
Reply-To: Free-VMS@lp.se
Message-ID: <199701312236.RAA19716@shell.monmouth.com>
Subject: Re: Kernels
To: Free-VMS@lp.se
Date: Fri, 31 Jan 1997 17:36:04 -0500 (EST)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
> I'd like to suggest an alternative: 4.4BSD. This is a completely free system
> that's actually been ported to more hardware platforms than Linux - there are
> ports available for PC's, Alphas, and VAXes. (The availability of a VAX port is
> nice since it would allow someone who owns a cheap VAX to run either a BSD-based
> FreeVMS or true VAX VMS - handy for development.) The system comes with both IP
> and OSI stacks - handy. It has a supports multiple file systems. It's imported
> from Mach some ideas on memory and process management. On the other hand, since
> it was developed on the VAX, it to this days shows vestiges of VAX influences in
> its design - which for FreeVMS would be a useful thing, since the pieces of the
> VAX that influence the design are exactly those that were created in conjunction
> with the VMS designers. Finally, there's a book describing the design and
> implementation, so it's relatively easy to get started. (I'm talking, BTW, of
> the new version of the book, describing 4.4BSD. There is an older one
> describing 4.3BSD.)
>
Sounds good -- except the unencumbered 4.4-lite is not bootable and is missing
pieces. NetBSD or OpenBSD might be reasonable. I prefer FreeBSD for
ease of installation and setup -- but it's still Intel - or - x86 specific.
Bill